25
HAL Id: hal-00982293 https://hal.archives-ouvertes.fr/hal-00982293 Submitted on 25 Apr 2014 HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés. Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya during 1999-2011 (vol 7, pg 1263, 2013) J. Gardelle, E. Berthier, Y. Arnaud, A. Kaab To cite this version: J. Gardelle, E. Berthier, Y. Arnaud, A. Kaab. Region-wide glacier mass balances over the Pamir- Karakoram-Himalaya during 1999-2011 (vol 7, pg 1263, 2013). The Cryosphere, Copernicus 2013, 7 (6), pp.1885-1886. 10.5194/tc-7-1885-2013. hal-00982293

Region-wide glacier mass balances over the Pamir-Karakoram

  • Upload
    others

  • View
    5

  • Download
    0

Embed Size (px)

Citation preview

HAL Id hal-00982293httpshalarchives-ouvertesfrhal-00982293

Submitted on 25 Apr 2014

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents whether they are pub-lished or not The documents may come fromteaching and research institutions in France orabroad or from public or private research centers

Lrsquoarchive ouverte pluridisciplinaire HAL estdestineacutee au deacutepocirct et agrave la diffusion de documentsscientifiques de niveau recherche publieacutes ou noneacutemanant des eacutetablissements drsquoenseignement et derecherche franccedilais ou eacutetrangers des laboratoirespublics ou priveacutes

Region-wide glacier mass balances over thePamir-Karakoram-Himalaya during 1999-2011 (vol 7 pg

1263 2013)J Gardelle E Berthier Y Arnaud A Kaab

To cite this versionJ Gardelle E Berthier Y Arnaud A Kaab Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya during 1999-2011 (vol 7 pg 1263 2013) The Cryosphere Copernicus 2013 7(6) pp1885-1886 105194tc-7-1885-2013 hal-00982293

The Cryosphere 7 1263ndash1286 2013wwwthe-cryospherenet712632013doi105194tc-7-1263-2013copy Author(s) 2013 CC Attribution 30 License

Geoscientiic Geoscientiic

Geoscientiic Geoscientiic

Op

en

Acce

ss

The Cryosphere

Op

en

Acce

ss

Region-wide glacier mass balances over thePamir-Karakoram-Himalaya during 1999ndash2011

J Gardelle1 E Berthier2 Y Arnaud3 and A Kaumlaumlb4

1CNRS ndash Universiteacute Grenoble 1 Laboratoire de Glaciologie et de Geacuteophysique de lrsquoEnvironnement (LGGE) UMR518354 rue Moliegravere BP 96 38402 Saint Martin drsquoHegraveres Cedex France2CNRS Universiteacute de Toulouse LEGOS 14 av Edouard Belin Toulouse 31400 France3IRD ndash Universiteacute Grenoble 1 LTHELGGE 54 rue Moliegravere BP 96 38402 Saint Martin drsquoHegraveres Cedex France4Department of Geosciences University of Oslo PO Box 1047 Blindern 0316 Oslo Norway

Correspondence to E Berthier (etienneberthierlegosobs-mipfr)

Received 29 January 2013 ndash Published in The Cryosphere Discuss 7 March 2013Revised 2 July 2013 ndash Accepted 5 July 2013 ndash Published 9 August 2013

Abstract The recent evolution of Pamir-Karakoram-Himalaya (PKH) glaciers widely acknowledged as valu-able high-altitude as well as mid-latitude climatic indi-cators remains poorly known To estimate the region-wide glacier mass balance for 9 study sites spread fromthe Pamir to the Hengduan Shan (eastern Himalaya) wecompared the 2000 Shuttle Radar Topography Mission(SRTM) digital elevation model (DEM) to recent (2008ndash2011) DEMs derived from SPOT5 stereo imagery Dur-ing the last decade the region-wide glacier mass bal-ances were contrasted with moderate mass losses in theeastern and central Himalaya (minus022plusmn 012 m we yrminus1 tominus033plusmn 014 m we yrminus1) and larger losses in the west-ern Himalaya (minus045plusmn 013 m we yrminus1) Recently reportedslight mass gain or balanced mass budget of glaciersin the central Karakoram is confirmed for a larger area(+010plusmn 016 m we yrminus1) and also observed for glaciersin the western Pamir (+014plusmn 013 m we yrminus1) Thus theldquoKarakoram anomalyrdquo should be renamed the ldquoPamir-Karakoram anomalyrdquo at least for the last decade The overallmass balance of PKH glaciersminus014plusmn 008 m we yrminus1 istwo to three times less negative than the global average forglaciers distinct from the Greenland and Antarctic ice sheetsTogether with recent studies using ICESat and GRACE dataDEM differencing confirms a contrasted pattern of glaciermass change in the PKH during the first decade of the21st century

1 Introduction

The Pamir-Karakoram-Himalaya (PKH) mountain ranges arecovered by more than 70 000 km2 of glaciers (Arendt et al2012) Assessing glacier evolution over such a large andremote region is challenging but nevertheless required tocharacterize the impacts of climate change in the region(eg Bolch et al 2012) to assess glacial contribution to wa-ter resources (eg Immerzeel et al 2010 Kaser et al 2010)and global sea level rise (eg Kaumlaumlb et al 2012 Gardner etal 2013) and ultimately to reliably project glacier responseto 21st century climate changes (eg Radic and Hock 2011Marzeion et al 2012)

Length changes have been measured for about 100 glaciersin the PKH and revealed predominant retreat of glacier frontssince the mid-19th century except in the Karakoram (Scher-ler et al 2011a Bhambri et al 2012 Bolch et al 2012)The majority of glaciers lost area over the past decades andin most cases the rates of area loss have been increasing inrecent years (eg Bolch et al 2012) But changes in glacierlength and area should be treated with care when used toevaluate the impact of climate change on glaciers especiallyin the presence of surge-type andor debris-covered glacierswhich are common in the PKH (Yde and Paasche 2010)

Rather mass balance is the most relevant variable to assessglacier responses to climate variability (Oerlemans 2001Vincent 2002) Mass balance estimates however remainscarce in the PKH and may not adequately sample the widerange of climates and glacier responses in the region Five

Published by Copernicus Publications on behalf of the European Geosciences Union

1264 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

different methods of mass balance measurements have beenused in the PKH

i Glaciological (or traditional) measurements are rela-tively short-termed in general less than 10 yr mainlybecause of difficult access to generally remote glaciersIn addition measurement series are often biased to-wards small-to-medium-sized and debris-free glaciersfor logistical reasons (eg Dobhal et al 2008 Fujitaand Nuimura 2011 Azam et al 2012 Bolch et al2012 Yao et al 2012) Vincent et al (2013) discussedin detail the inadequate spatial and temporal samplingof those glaciological measurements in the western Hi-malaya Gardner et al (2013) have shown that during2003ndash2009 the extrapolation of those few and short-term glaciological mass balances to entire High Moun-tain Asia lead to an overestimation of the mass loss byabout a factor of 3

ii The accumulation area ratio (AAR) method relies heav-ily on the availability of glaciological mass balancesduring various years in order to establish an empiri-cal relationship between annual mass balance and AAR(eg Rabatel et al 2005) The extrapolation to un-measured glaciers is not straightforward This methodhas been applied by Kulkarni et al (2011) to 19 glaciersin north-west India

iii The hydrological method has only been used once inthe PKH to our knowledge providing 5 yr of mass bal-ance for Siachen Glacier Karakoram (Bhutiyani 1999)and is very difficult to implement due to a lack of accu-rate precipitation and runoff measurements at high alti-tude over the PKH (Bolch et al 2012 Immerzeel et al2012)

iv The geodetic method based on the comparison of topo-graphic data (DEM or laser altimetry) can be applied atregional scales in remote areas using satellite data suchas from ICESat SPOT5 or SRTM This method enablesan increase in the number and diversity of glacier mea-surements (eg in terms of glacier size elevation as-pect percentage of debris cover) (Berthier et al 2007Bolch et al 2011 Kaumlaumlb et al 2012 Nuimura et al2012 Pieczonka et al 2013) An important remaininglimitation is the difficulty to convert volume changes tomass changes (eg Huss 2013)

v The gravimetric method uses the data from the GravityRecovery and Climate Experiment (GRACE) missionlaunched in 2002 Gravity fields need to be corrected forthe influence of the glacial isostatic adjustment (GIA)and the hydrological changes off-glacier (eg Zhang etal 2013a) before the glacier mass change can be ex-tracted This method initially led to contrasted glaciermass budgets in High Mountain Asia (Matsuo and Heki

2010 Jacob et al 2012) The two most recent GRACEmass budgets for High Mountain Asia agree within their2σ error bars (minus14plusmn 17 Gt yrminus1 andminus23plusmn 24 Gt yrminus1

in Gardner et al 2013)

The ICESat- and GRACE-based methods are appropriateto assess the regional mass budget for large glacierized re-gions such as the PKH but also have their specific limita-tions The ICESat sampling is restricted to satellite orbitsthat are separated by tens of kilometers at PKH latitudes andthus does not allow estimating the mass balance of individ-ual glaciers Glacier changes can only be investigated overthe lifetime of ICESat 2003ndash2009 The gravimetric methodhas a coarse spatial resolution (typicallysim 400 km) so thatindividual glaciers and small hydrological basins cannot beresolved

The aim of our study is to present a homogeneous surveyof regional glacier mass balances along the PKH between1999 and 2011 using the geodetic method This is achievedby differencing two digital elevation models (DEMs) over9 study sites selected to be representative of the PKH cli-matic and glaciological diversity This allows for an estima-tion of the spatial pattern of glacier elevation changes alongthe 3000 km-long group of mountain ranges as well as theinfluence of debris cover on thinning rates Finally we ex-trapolate these measurements to provide a new estimate ofthe mass budget of PKH glaciers and their contribution toregional hydrology and sea level rise during the first decadeof the 21st century

Compared to Kaumlaumlb et al (2012) the length of the studyperiod has been doubled and the study area has been signif-icantly extended towards the east (Hengduan Shan China)and the west (Pamir Tajikistan) We also provide a nearly ex-haustive coverage of glacier elevation changes for our 9 studysites as opposed to the sampling by ICESat along satellitetracks only However contrary to the ICESat studies (Kaumlaumlbet al 2012 Gardner et al 2013) our method is limited bythe availability of SPOT5 DEMs on selected sites and doesnot provide insight into the seasonal and annual evolution ofglacier mass balance

2 Study area

Mass balances are investigated for 9 study sites spread alongthe PKH mountain ranges to capture the climatic and glacio-logical variability of the region The location of each studysite is displayed in Fig 1 as well as the extent of the cor-responding sub-regions used to extrapolate the results tothe whole PKH range (see Sect 34) PKH glacier meltwa-ter contributes to five major rivers of Asia (BrahmaputraGanges Indus Tarim and Amu Darya) whose catchment ar-eas are also presented in Fig 1

Each study site corresponds to the extent of a SPOT5 DEM(see Sect 31) A complete coverage of all glaciers in the

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1265

Fig 1 Overview of the extent and location of the nine study sites (black polygons) and corresponding sub-regions (red polygons) along thePKH range Brownish-filled background polygons represent the major river basins of the area Triangles indicate the location of dischargemeasurements discussed in Sect 55

PKH range could not be achieved because of funding restric-tions (SPOT5 is a commercial satellite) the busy acquisitionschedule of this satellite and cloud coverage A further con-strain is that the optical stereo images used to compute theDEMs should be acquired as close as possible to the end ofthe glacier ablation season In practice a time window of2ndash3 months is allowed for the acquisitions which is shortgiven the 26-day orbital cycle of SPOT5 Thus a maximumof 3 to 4 acquisition attempts are possible each year and onlyif the satellite is not programmed for higher priority targets(eg crisis situations) Despite these difficulties our SPOT5DEMs sampled in total 22 200 km2 of glaciers about 30 of the total glacierized area in the PKH (Arendt et al 2012)

The eastern-most sites (Hengduan Shan Bhutan Everestand West Nepal) are strongly influenced by the Indian andsouth-east Asian summer monsoons Ablation and accumula-tion of monsoon-type glaciers occur during the summer sea-son (eg Fujita 2008) Towards the other end of the PKHin the north-west (Pamir Hindu Kush and Karakoram) theclimate is dominated by the westerlies and glaciers are ofwinter-accumulation type The Spiti Lahaul site lies in a tran-sition zone influenced both by the monsoon and the wester-lies The topography also plays a strong role in the moisturetransfer as it dries out the southerly air flow and can preventthe air mass from travelling further north which results ina northward decrease of precipitation (Bookhagen and Bur-bank 2010)

Monsoon-type glaciers are expected to be more sensitivethan winter-accumulation type glaciers to a change in therainsnow limit driven by an increase in temperature (Fujita2008) Indeed a summer warming would increase the alti-

tude of the rainsnow limit and reduce snow accumulation onglaciers as well as decrease their surface albedo (less freshsnow with a high albedo) and thus enhance ablation

The Karakoram and the Pamir are known to host numer-ous surge-type glaciers (Barrand and Murray 2006 Hewitt2007 Kotlyakov et al 2008 Gardelle et al 2012a) char-acterized by the alternation between short active phases in-volving rapid mass transfer from high to low elevationsand longer quiescent phases of low mass fluxes Coplandet al (2011) reported a doubling in the number of glaciersurges after 1990 in the Karakoram with a total of 90 surge-type glaciers observed in the region since the late 1960sIn the Pamir an inventory from 1991 revealed 215 glacierswith signs of repeated surging (looped moraines fast ad-vance of the front) and 51 additional actively surging glaciers(Kotlyakov et al 2008)

Many PKH glaciers are heavily debris-covered in their ab-lation areas because of steep rock walls that surround themand intense avalanche activity (eg Bolch et al 2012) Thesize of these debris zones is highly variable and debris thick-ness can range from a few centimeters (dust or sand) to sev-eral meters (Mihalcea et al 2008 Zhang et al 2013b) Themean debris cover is about 10 of the total glacier area inthe Karakoram and Himalaya (Bolch et al 2012) and culmi-nate at 36 in the Central Himalaya (Scherler et al 2011a)

In the eastern PKH glacier termini are often connected topro-glacial lakes storing meltwater behind frontal morainesor dead-ice dams whereas in the western PKH pro-glaciallakes are less numerous and glacier ablation areas are oftenscattered with supra-glacial lakes (Gardelle et al 2011) re-sulting from successive coalescence of small melting ponds

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1266 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Table 1 Characteristics of the nine study sites Numbers between parentheses indicate the percentage difference between our user-definedglacier inventory and the RGI v20 (Arendt et al 2012)

Site name Glacier area Debris ( of Glacier area Arendt Date Landsat imagesthis study (km2) glacier area) et al (2012) (km2) (SPOT5 DEM)

PathRow Date

Hengduan Shan 1303 11 2541 (+88 ) 24 Nov 2011 p135r39 23 Sep 1999Bhutan 1384 11 1429 (+3 ) 20 Dec 2010 p138r40 19 Dec 2000

p138r41Everest 1461 22 1400 (minus4 ) 4 Jan 2011 p140r41 30 Oct 2000West Nepal 908 17 802 (minus12 ) 3 Jan 2011 p143r39 1 Aug 2009

p143r40Spiti Lahaul 2110 13 2257 (+7 ) 20 Oct 2011 p147r37 15 Oct 2000

p147r38Hindu Kush 793 7 724 (minus11 ) 17ndash21 Oct 2008 p150r34 15 Aug amp

p150r35 16 Sep 1999Karakoram East 5328 10 5384 (+1 ) 31 Oct 2010 p148r35 7 Sep 1998Karakoram West 5434 12 5687 (+1 ) 3 Dec 2008 p149r35 29 Aug 1998Pamir 3178 11 3210 (minus1 ) 29 Nov 2011 p152r33 16 Sep 2000

3 Data and methods

31 Digital elevation models

For each study site we used the Shuttle Radar TopographicMission (SRTM) version 4 DEM as the reference topogra-phy (Rabus et al 2003) This data set acquired in Febru-ary 2000 comes along with a mask to identify the pixelswhich have been interpolated (both were downloaded fromhttpsrtmcsicgiarorg) On each site the SRTM DEM issubtracted from a more recent DEM built from a stereo pairacquired by the HRS sensor onboard the SPOT5 satellite(Korona et al 2009) except for the Hindu Kush study sitewhere the recent DEM has been created from a stereo pairacquired by the HRG sensor also onboard SPOT5 (Berthieret al 2007) The date of each SPOT5 DEM acquisition dif-fers depending on the study site (Table 1) but 7 out of 9 wereacquired between October 2010 and November 2011 EachSPOT5-HRS DEM is also provided with a correlation maskand an ortho-image generated from the rear HRS image

SPOT5-HRS DEMs were produced by the French Map-ping Agency (IGN) using two sets of correlation parametersdefined during the SPIRIT (SPOT 5 stereoscopic survey ofPolar Ice Reference Images and Topographies) internationalpolar year project (Korona et al 2009) Thus two versionsof the DEM are delivered with their respective mask one op-timized (v1) for a gentle topography and the second one (v2)for a more rugged relief Earlier work has shown that v2 hasa slightly larger percentage of interpolated pixels but for thenon-interpolated pixels smaller errors (Korona et al 2009)This version (v2) is preferred here Given that both DEMsare calculated from the same image pair little difference isexpected between v1 and v2 However locally we identifiedsome large bull eye elevation differences between the two

versions of the DEM both off and on glaciers Comparisonwith the SRTM DEM shows that both versions of the DEMare erroneous for those areas and thus we preferred to ex-clude from further analysis all pixels for which the absoluteelevation difference between the two versions of the DEMis larger than 5 m About 20 of the valid DEM pixels areexcluded through that procedure

The SRTM DEM and mask originally at a 3 arcsecondresolution (sim 90 m) are resampled bilinearly to 40 m (UTMprojection WGS84 ellipsoid) to match the posting and pro-jection of the SPOT5 DEM Altitudes are defined above theEGM96 geoid for both DEMs

32 Glacier mask

Where available glacier outlines were downloaded fromthe GLIMS database (Raup et al 2007 Supplement) andmanually corrected if needed Otherwise the glacier maskwas derived from Landsat-TM and Landsat-ETM+ imagesPathrow and dates of Landsat acquisition are given in Ta-ble 1 for each study site A threshold varying between 05and 07 depending on the study site was applied to the Nor-malized Difference Snow Index (TM2-TM5)(TM2+TM5)to detect clean ice and snow automatically whereas debris-covered parts were digitized manually by visual interpreta-tion (eg Racoviteanu et al 2009) The total glacier areaand debris-covered part can be found for each study site inTable 1

Most inventories are derived from Landsat images withminimal snow cover from around the year 2000 prior to the2003 failure of the Scan Line Corrector of the ETM+ sensoronboard Landsat 7 that resulted in image striping Glacierareas are expected to have experienced minor changes sincethen except for glacier fronts that retreated due to the rapid

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1267

Table 2 Elevation ranges equilibrium line altitudes (ELA) and accumulation area ratios (AAR) for the nine study sites The dates ofacquisition of Landsat images used for ELA determination are also given

Sub-region Glaciers elevation ELA (m) AAR Landsat imagesrange (m)

Min Max This study Scherler et Kaumlaumlb et This study Kaumlaumlb et PathRow Dateal (2011a) al (2012) al (2012)

Hengduan Shan 2630 6860 4970plusmn 320 NA NA 55 NA p135r39 23 Sep 1999Bhutan 4390 7480 5690plusmn 440

5700 555036

47p138r40 17 Nov 2000

Everest 4260 8380 5840plusmn 320 31 p140r41 31 oct 2009West Nepal 3950 6870 5590plusmn 138 37 p143r39 1 Aug 2009Spiti Lahaul 3780 6360 5390plusmn 140 5103 5500 34 35 p147r37 2 Aug 2002

p147r38Hindu Kush 2717 6385 5050plusmn 160 5138 30 p150r34 2 Sep 2000

p150r35Karakoram east 3230 7770

5030plusmn 280 4845 554076

47p148r35 7 Sep 1998

Karakoram west 2910 7920 66 p149r35 29 Aug 1998Pamir 2800 7090 4580plusmn 250 NA NA 52 NA p152r33 30 Jul 2000

growth of connected lakes and for glaciers that surged andthus advanced For the few glaciers that advanced their frontposition was updated manually based on the SPOT5 ortho-image

For each study site we favored our user-defined glacierinventories instead of the global Randolph Glacier Inventory(RGI v20 Arendt et al 2012) because the latter is hetero-geneous and for some study sites did not match our accu-racy requirements for the adjustments and corrections of theDEMs

We also estimated the altitude of the equilibrium line(ELA) as the snowline on thesim 2000 Landsat data with min-imal snow cover to separate ablation and accumulation ar-eas The ELA is assumed to be identical to the altitude ofthe snowline at the end of the ablation season (eg Raba-tel et al 2005) after the end of the monsoons or before thefirst snowfalls Therefore we manually digitized snowlinesfor more than 30 glaciers for each study site and computedtheir mean elevations Our ELAs are slightly different fromthose given by Scherler et al (2011a) and Kaumlaumlb et al (2012)probably due to the sensitivity of the ELA to the choice ofthe image (Table 2) Ideally the ELA should have been es-timated from images acquired at the end of the ablation sea-son during various years covering our whole study periodThis would be a highly time-consuming task that was not ac-complished here However we stress that in our study theELA is only used to compare elevation changes on the abla-tionaccumulation areas (Sect 41) and to estimate the sen-sitivity of our mass balances to the choice of the volume-to-mass conversion factor (Sect 35) Our preferred mass bal-ance estimate uses a unique volume-to-mass conversion fac-tor for the whole glacier and is thus independent from theELA

33 Adjustments and corrections of DEM biases

Different DEM processing steps are followed to extract un-biased glacier elevation changes for each study site (eg Nuthand Kaumlaumlb 2011 Gardelle et al 2012b)

331 Planimetric adjustment of the DEMs

A horizontal shift between two DEMs results in an aspect-dependent bias of elevation differences (Nuth and Kaumlaumlb2011) Here the horizontal shift is determined by minimizingthe root mean square error of elevation differences (SPOT5-SRTM) off glaciers where the terrain is assumed to be stableover the study period (Berthier et al 2007) The SRTM DEMis then resampled (cubic convolution) according to the shift

332 Alongacross track corrections (drift of thesatellite orbit)

For DEMs derived from stereo imagery the satellite acquisi-tion geometry can induce a bias along andor across the trackdirection (Berthier et al 2007) We estimate the azimuth ofeach SPOT5 ground track and use it to rotate the coordinatesystem accordingly (Nuth and Kaumlaumlb 2011) Then we com-pute the elevation differences along and across the satellitetrack on stable areas and when necessary correct the bias us-ing a 5th order polynomial fit

333 Curvature correction

As suggested by Paul (2008) and Gardelle et al (2012b) thedifference in original spatial resolutions of the two DEMs canlead to biases related to altitude in mountainous areas Whenthe curvature the first derivative of the slope is high (sharppeaks or ridges) the altitude tends to be underestimatedby the coarse DEM unable to reproduce high-frequency

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1268 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

slope variations Thus this apparent ldquoelevation biasrdquo is cor-rected using the relation between elevation differences andthe maximum terrain curvature estimated over stable areasoff glaciers (Gardelle et al 2012b) Note that the units inFig 1b in Gardelle et al (2012b) should be 10minus2 mminus1 in-stead of mminus1 (A Gardner personal communication 2012)However this does not impact the validity of the correction

334 Radar penetration correction

The SRTM DEM acquired in C-band potentially underesti-mates glacier elevations since at this wavelength (sim 56 cm)the penetration of the radar signal into snow and ice canreach several meters (eg Rignot et al 2001) Here we esti-mate this penetration for each study site by differencing theSRTM C-band (57 GHz) and X-band (97 GHz) DEMs Thelatter was acquired simultaneously with the SRTM X-bandat higher resolution (30 m) but with a narrower swath andthus has a coverage restricted to selected swaths only Sincethe penetration of the X-band is expected to be low com-pared to the C-band (an hypothesis that still needs to be con-firmed Ulaby et al 1986) we consider the elevation differ-ence (SRTMC-bandminus SRTMX-band) to be a first approxima-tion of C-band radar penetration over glaciers (Gardelle etal 2012b) The correction is calculated as a function of alti-tude and applied to each pixel separately

For the Hindu Kush Spiti Lahaul and West Nepal sitesdue to the lack of SRTMX-band data we were unable to de-termine the value of the penetration Thus we applied thecorrection computed over the nearest study site ie Karako-ram for Spiti Lahaul and Hindu Kush and Everest for WestNepal The mean values of the SRTMC-bandpenetrations overglaciers are given for each study site in Table 3

335 Seasonality correction

SPOT5 DEMs were acquired between late October and earlyJanuary depending on the study site Since the SRTM DEMis from mid-February we need to account for possible masschanges during this 1-to-5-month period in order to esti-mate glacier mass balance over an integer number of years(Gardelle et al 2012a) For the eastern sites (from WestNepal to Hengduan Shan) where glaciers are of summer-accumulation types (Fujita 2008) the correction was set to0 This choice is confirmed by recent field observations thatshow no winter accumulation on Mera and Pokalde glaciersin Nepal (Wagnon et al 2013) For the Karakoram studysites we used the winter accumulation rate+013 m we permonth measured on Biafo Glacier (Wake 1989) For theother western study sites (the Pamir Hindu Kush and SpitiLahaul sites) the correction is derived from the mean of win-ter mass balances of 35 glaciers all in the Northern Hemi-sphere+089 m we yrminus1 over 2000ndash2005 corresponding to+015 m we per winter month (Ohmura 2011) The lattervalue is slightly lower than the average winter mass balance

Table 3 Average SRTM C-band penetration estimates (in m)in February 2000 over glaciers in this study and from Kaumlaumlb etal (2012)

Sub-region This study Kaumlaumlb et al (2012)

Hengduan Shan 17 NABhutan 24

25plusmn 05Everest 14West Nepal NA

15plusmn 04Spiti Lahaul NAHindu Kush NA 24plusmn 04Karakoram East

34 24plusmn 03Karakoram WestPamir 18 NA

(+135plusmn 031 m we yrminus1) measured on Abramov Glacier inthe Pamir during 1968ndash1998 (WGMS 2012) and in reason-ably good agreement with the modeled mean winter massbalance (+108plusmn 034 m we yrminus1) of Chhota Shigri Glacierduring 1969ndash2012 in the western Himalaya (Azam et al2013)

336 Mean elevation changes and mass balancecalculation

Before averaging elevation changes we exclude all interpo-lated pixels in the SRTM or SPOT5 DEMs as well as unex-pected elevation changes exceedingplusmn150 m for study sites(Pamir and Karakoram) that include surge-type glaciers orplusmn80 m for other sites Those thresholds were chosen aftercareful inspection of the elevation difference maps and his-tograms in order to identify the largest realistic glacier eleva-tion changes Glaciers that are truncated at the edges of theDEM are also excluded from mass balance calculations asthey may bias the final results if for example only their accu-mulation or ablation area is sampled Although they are trun-cated Siachen (Karakoram East) and Fedtchenko (Pamir)glaciers were retained because of their wide coverage andbecause their accumulation and ablation areas were correctlysampled in the DEMs

Elevation changes are then analyzed for 100 m altitudebins Within each altitude band we average the elevationchanges after excluding pixels where absolute elevation dif-ferences differ by more than three standard deviations fromthe mean (Berthier et al 2004 Gardner et al 2012) Thisis an efficient way to exclude outliers based on the assump-tion that elevation changes should be similar at a given alti-tude within each study site This assumption does not holdfor regions containing actively surging glaciers Thus forthe Pamir and Karakoram study sites surge-type glaciersare treated separately ie elevation changes are averaged by100 m altitude bands for each surge-type glacier individually

Where no elevation change is available for a pixel weassign to it the value of the mean elevation change of the

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1269

altitude band it belongs to in order to assess the mass bal-ance over the whole glacier area

The conversion of elevation changes to mass balance re-quires knowledge of the density of the material that has beenlost or gained and its evolution during the study period Giventhe lack of repeat measurement of density profiles over theentire snowfirnice column in the PKH we applied a con-stant density conversion factor of 850 kg mminus3 for all ourstudy sites (Sapiano et al 1998 Huss 2013)

The mass balance of each surge-type glacier is computedseparately and added (area-weighted) to the mass balance ofthe rest of the glaciers to estimate the mass balance of thewhole study site

For study sites with a high concentration of growingpro-glacial lakes such as Bhutan Everest and West Nepal(Gardelle et al 2011) our mass losses are slightly underes-timated because they do not take into account the glacier icethat has been replaced by water during the expansion of thelake Subaqueous ice losses are only estimated for LhotseSharImja Glacier (Everest area Fujita et al 2009) for thesake of comparison to previous studious

The region-wide mass balance of PKH glaciers as definedin Sect 2 is computed by extrapolating the mass balancefrom each study site to a wider sub-region (Fig 1) that is as-sumed to experience the same glacier mass changes becauseof its similar climate (Bolch et al 2012 Kaumlaumlb et al 2012)The total glacier area for each sub-region is computed fromthe RGI v20 (Arendt et al 2012)

The periods over which we calculate mass balance areslightly different from one site to another (Table 1) depend-ing on the year of acquisition of the SPOT5 DEM (two in2008 two in 2010 and five in 2011) In the following we willassume that all mass balances are representative of the period1999ndash2011 in order to estimate a region-wide mass budgetof PKH glaciers neglecting thus part of the inter-annual vari-ability

34 Accuracy assessment

The errorE1hi to be expected for a pixel elevation change

1hi is assumed to equal the standard deviationσ1h of themean elevation change1h of the altitude band it belongs toThe value ofσ1h can range fromplusmn4 m toplusmn 20 m dependingon the study site and elevation This is rather conservative asthe value ofσ1h contains also the intrinsic natural variabilityof elevation changes within the altitude band (ie it containsboth noise and real geophysical signal)

The errorE1h of the mean elevation change1h in eachaltitude band is then calculated according to standard princi-ples of error propagation

E1h =E1hiradicNeff

(1)

Neff represents the number of independent measurements inthe altitude bandNeff will be lower thanNtot the total num-

ber of elevation change measurements1hi in the altitudeband since the latter are correlated spatially (Bretherton etal 1999)

Neff =Ntot middot PS

2d (2)

where PS is the pixel size (here 40 m)d is the distance ofspatial autocorrelation determined using Moranrsquos I autocor-relation index on elevation differences off glaciers On theaverage over the 9 study sitesd = 492plusmn 72 m

The error on the penetration correction was assumed to besystematic due to the not fully understood physics involvedand equal toplusmn 15 m This error was obtained by compar-ing the SRTM C-band penetration estimated using two inde-pendent methods (i) SRTMC-bandndash SRTMX-band and (ii) dif-ference between ICESat and SRTM when ICESat elevationchange trends during 2003ndash2008 are extrapolated to Febru-ary 2000 (Kaumlaumlb et al 2012) The maximum difference of thepenetration inferred from the two methods is 10 m (Table 3)A 50 additional error margin was added to account for thefact that this comparison could only be made on 2 of ourstudy sites

Given the slender observational support for the seasonal-ity correction (Section 33) we assumed its uncertainty tobe plusmn100 on the western sites ieplusmn015 m we per win-ter month The same absolute uncertainty (plusmn015 m we perwinter month) was used for the eastern sites where no sea-sonality correction is applied but ablation andor accumula-tion may occur in winter (Wagnon et al 2013)

All errors are summed quadratically within each altitudeband and then summed for all altitude bands assuming con-servatively that they are 100 dependent The error on theSRTM penetration correction clearly dominates our errorbudget for volume change

During the conversion from volume to mass we assumea plusmn60 kg mminus3 error on the density conversion factor (Huss2013) Thisplusmn7 error is similar to the one estimated byBolch et al (2011) In a sensitivity test the volume to massconversion is also performed using two alternative densityscenarios (Kaumlaumlb et al 2012) (i) a density of 900 kg mminus3 isassumed everywhere and (ii) 600 kg mminus3 in the accumula-tion area and 900 kg mminus3 in the ablation area These two sce-narios take among other things into account that elevationchanges could be due to changes in glacier dynamics or dueto surface mass balance We then calculate the maximum ab-solute difference between the mass balances calculated usingour preferred scenario (850plusmn 60 kg mminus3 everywhere) and themass balances calculated using the two others scenarios Forthe eastern sites (West Nepal to Hengduan Shan) the max-imum absolute difference is 003 m we yrminus1 For the west-ern sites (Pamir to Spiti Lahaul) it reaches 006 m we yrminus1Those uncertainties due to the choice of a given density sce-nario remain negligible compared to other sources of errors

We also include an error due to extrapolation of our massbalance estimate from our study site (the extent of each

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1270 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

SPOT5 DEM) to the unmeasured glaciers within each sub-region To estimate this regional extrapolation error we com-pare the ICESat elevation change trends (computed as inKaumlaumlb et al 2012) within 3 times 3 cells centered at the studysites (Table 5) with the ICESat-derived trends for the entiresub-regions The absolute difference between both trends isless than 002ndash003 m yrminus1 for Bhutan West Nepal Spiti La-haul and Karakoram ie negligible For Everest the eleva-tion trend for the entire sub-region is 012 m yrminus1 less nega-tive than for the 3 times 3 cells around the SPOT5 DEM cen-ter for Hindu Kush the trend for the entire sub-region is017 m yrminus1 more negative than for the 3 times 3 cell due to thevariability of elevation changes within this sub-region (Kaumlaumlbet al 2012) The regional extrapolation error is obtainedas the area-weighted mean absolute difference between themass balances derived from ICESat for those 3 times 3 cellsand the whole sub-region and equalsplusmn004 m we yrminus1

Finally for sub-regions and total PKH estimates of masschange we include an error on the glacier area computedfrom the RGI v20 (Arendt et al 2012) The inventory er-ror is specific to each study site and evaluated based on thecomparison of our user-defined inventory on each study sitewith the RGI v20 (Table 1) We note the remarkable accu-racy of the RGI for all but one of our study sites The rel-ative errors are generally of a few percents and up to 12 for West Nepal The differences between our and existing in-ventories are probably due to the difficulty of delimitatingdebris-covered glacier parts (eg Frey et al 2012 Paul etal 2013) and accumulation areas The Hengduan Shan studysite is an exception There the RGI is based on the ChineseGlacier Inventory (Shi et al 2009 Arendt et al 2012) andoverestimates the ice-covered area by 88 compared to ourLandsat-based inventory

Unless stated otherwise all error bars from this study andpreviously published studies correspond to one standard er-ror

4 Results

41 Overview of the elevation changes

The maps of glacier elevation changes are given in Figs 2ndash10for our nine study sites with as inset the distribution ofthe elevation differences off glaciers The full maps of ele-vation differences off glaciers are shown in the Supplement(Figs S2ndashS10) Those off glacier maps indicate a local ran-dom noise of aboutplusmn5ndash10 m However at length scales of afew kilometers the spatial variations in elevation differencesare small typically less than 1 m

For the four eastern study sites from the HengduanShan to the West Nepal (Fig 1) the elevation changesaveraged over the accumulation areas are small (between003plusmn 014 m yrminus1 to 013plusmn 015 m yrminus1) whereas clear sur-face lowering is observed for all four study sites in their

ablation areas ranging fromminus050plusmn 014 m yrminus1 (Bhutan)to minus081plusmn 013 m yrminus1 (Hengduan Shan) Over the BhutanEverest and West Nepal study sites glaciers in contact witha proglacial lake often experienced larger thinning rates

Further west the Spiti Lahaul is the only site wheresurface lowering is measured both in the accumula-tion area (minus034plusmn 016 m yrminus1) and the ablation area(minus070plusmn 014 m yrminus1) Clear surface lowering is also ob-served in the ablation areas of the Hindu Kush glaciers(minus050plusmn 018 m yrminus1)

In the Karakoram (east and west Figs 8 and 9) andin the Pamir (Fig 10) the pattern of elevation changesis highly heterogeneous due to the occurrence of numer-ous glacier surges or similar flow instabilities The eleva-tion changes in the accumulation areas of those three studysites are slightly positive (between+022plusmn 014 m yrminus1 and+030plusmn 018 m yrminus1) In the ablation areas small surfacelowering is observed for the two Karakoram study sites (av-erage of the two sitesminus033plusmn 016 m yrminus1) whereas no el-evation change is found in Pamir (minus002plusmn 013 m yrminus1)

Note that those mean rates of elevation changes are sensi-tive to the choice of the ELA assumed to equal the snowlinealtitude in Landsat images from around year 2000

42 Influence of a debris cover

To evaluate thinning rates over debris-covered and debris-free ice we perform a histogram adjustment in order tocompare data sets with similar altitude distributions This isneeded because debris-covered parts tend to be concentratedat lower elevations The adjustment consists in randomly ex-cluding pixels over debris-free ice from each elevation bandto match a scaled version of the debris-covered pixel distribu-tion (Kaumlaumlb et al 2012) Elevation changes with altitude canthen be compared for clean and debris-covered ice (Fig 11)

The thinning is higher for debris-covered ice in the Ever-est area and on the lowermost parts of glaciers in the PamirBut on average in the Pamir western Karakoram and SpitiLahaul the lowering of debris-free and debris-covered ice issimilar The thinning is stronger over clean ice in the Heng-duan Shan Bhutan West Nepal Hindu Kush and althoughto a lesser extent in the eastern Karakoram

43 Mass balance over the nine study sites

Glacier mass changes have been heterogeneous overthe PKH since 1999 Slight mass gains or balancedmass budgets are found in the north-west in thePamir (+014plusmn 013 m we yrminus1) and for the two studysites in the central Karakoram (+009plusmn 018 m we yrminus1

and +011plusmn 014 m we yrminus1) Mass loss is moderatein the adjacent Hindu Kush with a mass balance ofminus012plusmn 016 m we yrminus1 Glaciers in all our eastern studysites (two in Nepal Bhutan and Hengduan Shan) experi-enced homogeneous mass losses with mass balances ranging

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1271

Fig 2 Glacier elevation changes over the Hengduan Shan study site Inset distribution of the elevation differences off glacier The medianand the standard deviation of the elevation difference and the number of pixels off glaciers are given (by construction the mean elevationdifference is null)

Fig 3Same as Fig 2 but for the Bhutan study site

from minus033plusmn 014 m we yrminus1 to minus022plusmn 012 m we yrminus1The most negative mass balance is measured in Spiti Lahaul(western Himalaya) atminus045plusmn 013 m we yrminus1 (Table 4)

However within a study site mass balances can be highlyvariable from one glacier to another In the Everest area twolarge glaciers (71 km2 each) experienced distinctly differentmass balance withminus016plusmn 016 m we yrminus1 for the south-flowing Ngozumpa Glacier andminus059plusmn 016 m we yrminus1

for the north-flowing Rongbuk Glacier (Fig 4) In Bhutan(Fig 3) mass loss is higher south (minus025plusmn 013 m we yrminus1)

than north (minus014plusmn 012 m we yrminus1) of the main Hi-malayan ridge though not at a statistically significant level

The region-wide mass budget of PKH glaciers isnegative minus101plusmn 55 Gt yrminus1 (minus014plusmn 008 m we yrminus1)which corresponds to a sea level rise contribution of0028plusmn 0015 mm yrminus1 over 1999ndash2011

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1272 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 4Same as Fig 2 but for the Everest study site

Fig 5Same as Fig 2 but for the West Nepal study site

44 Glacier surges

In the Pamir and the Karakoram the spatial pattern of ele-vation changes of many glaciers is typical of surge events orflow instabilities They can be identified because of their veryhigh thinning and thickening rates that can both reach up to16 m yrminus1 (Figs 8ndash10)

Most of them have already been reported to be surge-type glaciers based on their velocity (Kotlyakov et al 2008Quincey et al 2011 Copland et al 2009 Heid and Kaumlaumlb2012) morphology (Copland et al 2011 Barrand and Mur-ray 2006 Hewitt 2007) or elevation changes (Gardelle etal 2012a) Here we only identify surge-type glaciers for thesake of mass balance calculation and we do not attempt toprovide an exhaustive up-to-date surge inventory This is why

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1273

Fig 6 Same as Fig 2 but for the Spiti Lahaul study site The map of elevation changes contains many voids due to the presence ofthin clouds during the acquisition of the SPOT5 DEM The mass balance of the Chhota Shigri and Bara Shigri glaciers are respectivelyminus039plusmn 018 m we yrminus1 andminus048plusmn 018 m we yrminus1

the published surge inventories in the Pamir and the Karako-ram do not necessarily match the surge-type glaciers labeledon the elevation change maps as these refer only to our ob-servation period (1999ndash2011)

The mass balance of surge-type glaciers (respectivelynon-surge-type glaciers) is +019plusmn 022 m we yrminus1

(+012plusmn 011 m we yrminus1) in the Pamir+009plusmn 030 m we yrminus1 (+009plusmn 015 m we yrminus1) inthe western Karakoram and+010plusmn 025 m we yrminus1

(+012plusmn 012 m we yrminus1) in the eastern Karakoram Theoverall similarity between the mass budgets for surge-typeand non-surge-type glaciers shows that those flow instabili-ties seem not to affect the glacier-wide mass balance at leastover short time periods here 10 yr

5 Discussion

51 SRTM penetration

Alternative estimates of SRTMC-bandpenetration for the PKHregion were obtained by Kaumlaumlb et al (2012) by comparing el-evations of the ICESat altimeter (Zwally et al 2002) with theSRTMC-bandDEM in PKH and extrapolating the 2003ndash2008trend of elevation changes back to the acquisition date ofSRTM The difference between east and west is more dis-tinct in our case with a mean penetration in the Karakoram of34 m (24plusmn 03 m in Kaumlaumlb et al 2012) 24 m in Bhutan and14 m around the Everest area (25plusmn 05 m for a wider areaincluding east Nepal and Bhutan in Kaumlaumlb et al 2012) Thiseastwest gradient is expected given that in February 2000when the SRTM mission was flown the snowfirn thicknesswas probably larger in the western PKH where glaciers areof winter-accumulation type than in the eastern PKH where

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1274 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 7Same as Fig 2 but for the Hindu Kush study site

glaciers are of summer-accumulation type Further studieswould be necessary to investigate the possible reasons be-hind the relatively low SRTMC-bandpenetration for the Pamir(18 m) that do not follow the eastwest gradient mentionedabove

Both our and Kaumlaumlb et alrsquos (2012) estimates agree withintheir error bars and seem thus to provide robust first-orderestimates for the SRTMC-bandradar penetration in the regionHowever it remains to be verified if a penetration depth com-puted at the regional scale is appropriate for a single smallglacier whose altitude distribution is different from the oneof whole study site (eg Abramov Glacier in the far north ofour Pamir study site)

52 Thinning over debris-covered ice

PKH glaciers are heavily debris-covered in their ablation ar-eas because of steep rock walls that surround them and in-tense avalanche activity Twelve percent (12 ) of the glacierarea in our study sites are covered with debris (up to 22 inthe Everest area Table 1) which is in agreement with pre-vious estimates for the Himalaya and the Karakoram 13 in Kaumlaumlb et al (2012) and 10 in Bolch et al (2012) Thedebris layer is expected to modify the ablation of the under-

lying ice It will increase ablation if its thickness is just a fewcentimeters (dominance of the albedo decrease) or decreaseablation it if the layer is thick enough to protect the ice fromsolar radiation (Mattson et al 1993 Mihalcea et al 2006Benn et al 2012 Lejeune et al 2013)

However recent studies have reported similar overall thin-ning rates over debris-covered and debris-free ice in thePKH (Kaumlaumlb et al 2012 Gardelle et al 2012a) or evenhigher lowering rates over debris-covered parts in the Ever-est area (Nuimura et al 2012) Our findings are consistentwith Nuimura et al (2012) for the Everest study site with astronger thinning (rate of elevation change ofminus097 m yrminus1)

in debris-covered areas than over clean ice (rate of elevationchange ofminus057 m yrminus1) In the Pamir Karakoram and SpitiLahaul study sites these rates are similar while in the HinduKush West Nepal Bhutan and Hengduan Shan study sitesthinning is greater over clean ice in agreement with the com-monly assumed protective effect of debris (Fig 4) Thosedifferences between our 9 study sites suggest a complex re-lationship between debris cover and glacier wastage

As local glacier thickness changes are a result of bothmass balance processes and a dynamic component theseunexpected high thinning rates over debris-covered parts

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1275

Fig 8Same as Fig 2 but for the Karakoram east study site Surge-type glaciers in an active surge phase (resp quiescent phase) are identifiedwith a solid triangle (resp solid circle)

are potentially the result of glacier-wide processes Over aglacier tongue the ablation can be enhanced by the pres-ence of steep exposed ice cliffs or supraglacial lakes andponds (Sakai et al 2000 2002) In addition it is also pos-sible that the glacier tongues are mostly covered by thindebris layers such that the albedo effect dominates the in-sulating effect resulting in an increased tongue-wide abla-tion The prevalence of thin debris was recently suggestedfor debris-covered glaciers around the Mount Gongga in thesouth-eastern Tibetan Plateau (Zhang et al 2013b) Finallyice-flow rates could be very low at debris-covered parts asshown by Quincey et al (2007) in the Everest area Kaumlaumlb(2005) for Bhutan and Scherler et al (2011b) in the HinduKush-Karakoram-Himalaya Thus all three factors are likelyto increase the overall thinning under debris despite the un-doubted protective effect of a continuous thick debris coverat a local scale Future investigations combining surface ve-locity fields and maps of elevation changes could allow eval-uating more closely the relative role of ablation and ice fluxesin thinning rates over PKH glacier tongues (eg Berthierand Vincent 2012 Nuth et al 2012) Measurements of the

spatial distribution of debris thickness over the PKH glaciertongues are also needed

Note that in the case of debris-covered tongues the ele-vation change measurement represents the glacier thicknesschange but also the possible debris thickness evolution Thelatter remains poorly known in the PKH or in any othermountain range But based on the debris discharges givenby Bishop et al (1995) for Batura Glacier (Pakistan) 48to 90times 103 m3 yrminus1 the thickening of the debris can be es-timated between 27 and 55 mm yrminus1 which is negligiblegiven the magnitude of the glacier elevation changes

53 Comparison to previous mass balance estimates

Throughout the PKH there is only a single peer-reviewedglaciological record (on Chhota Shigri Glacier) coveringmost of our study period (Wagnon et al 2007 Vincent etal 2013) Our geodetic mass balance estimate for the SpitiLahaul study site has been recently compared to those glacio-logical measurements on Chhota Shigri Glacier (Vincent etal 2013) Mass balance records reported in Yao et al (2012)for the Tibetan Plateau and the surroundings mountain ranges

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1276 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 9Same as Fig 8 but for the Karakoram west study site

only start during glaciological year 20052006 This is whyhere we do not compare our mass balance estimates toglaciological records and limit our comparison to others re-gional estimates obtained using the geodetic or gravimetricmethods

We stress that the comparison of our mass balance mea-surements to published estimates is complicated by thefact that generally they do not cover the same time pe-riod Little is known about the inter-annual variability inthe PKH during the 21st century due to the limited num-ber of long glaciological time series The 9 yr mass balance

record on Chhota Shigri Glacier reveals a relatively highinter-annual variability of the annual mass balance (stan-dard deviationplusmn 074 m we yrminus1 during 2003ndash2011 Vin-cent et al 2013) which is similar to the one obtained be-tween 1968 and 1998 on Abramov Glacier (standard devi-ation 069 m we yrminus1 WGMS 2012) Thus inter-annualvariability can explain part of the difference between two cu-mulative mass balance estimates over 5ndash10 yr even if they areoffset by only one or two years

Based on ICESat repeat track measurements and theSRTM DEM Kaumlaumlb et al (2012) measured a region-wide

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1277

Fig 10Same as Fig 8 but for the Pamir study site

mass balance ofminus021plusmn 005 m we yrminus1 between 2003and 2008 for Hindu Kush-Karakoram-Himalaya glaciersFor that same area (ie excluding the Pamir and theHengduan Shan study sites) we found a mass balance ofminus015plusmn 007 m we yrminus1 Thus our result agrees with Kaumlaumlbet al (2012) within the error bars although our study periodis longer (1999ndash2011) and our measurement principle andextrapolation approach are different In addition we havecalculated updated mass balances from Kaumlaumlb et al (2012)ie averages over 3times 3 degree cells centered over our studysites and find that they are consistent within error barswith our mass balance estimates (Table 5) Similar resultsare found when our mass balances are compared to a spa-tially extended ICESat analysis for 2003ndash2009 (Gardner etal 2013) Mass losses measured with ICESat (Kaumlaumlb et al2012 Gardner et al 2013) tend to be slightly larger thanours This could be due to the difference in time periodsAlso our smaller mass losses could be partly explained if wesystematically underestimated the poorly constrained pene-

tration of the C-band andor X-band radar signal into ice andsnow

At a more local scale Bolch et al (2011) reported a massbalance ofminus079plusmn 052 m yrminus1 between 2002 and 2007 byDEM differencing over a glacier area of 62 km2 includingten glaciers south and west of Mt Everest For these sameglaciers Nuimura et al (2012) found a mean mass balanceof minus045plusmn 060 m yrminus1 during 2000ndash2008 while they com-puted a mass balance ofminus040plusmn 025 m yrminus1 between 1992and 2008 over a glacier area of 183 km2 around Mt EverestIn the present study we found an average mass balance ofminus041plusmn 021 m yrminus1 over the ten glaciers mentioned above(Table 5 Fig S1) and a value ofminus026plusmn 013 m yrminus1 overthe whole Everest study site between 2000 and 2011 Ourvalues are comparable to those of Nuimura et al (2012)Our mass balance is less negative than the 2002ndash2007 massbalance of Bolch et al (2011) but not statistically differentwhen error bars are considered (Fig 12) Indeed Bolch etal (2011) acknowledged some high uncertainties for their

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1278 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Table 4Glacier mass budget of the nine sub-regions to which the results of the study sites (mass balances) have been extrapolated For eachof them the total glacierized area as well as the percentage of the glacier area over which the mass balance was computed (ie study sites)are given The total glacier area is derived from the RGI v20 (Arendt et al 2012) The error for the mass balance in each sub-region is theroot of sum of squares (RSS) of the mass balance error of each study site and the regional extrapolation error The error for the mass budgetin each sub-region includes additionally the error from the total glacier area in the RGI v20

Sub-region Total glacier Measured glacier Mass balance Mass budgetarea (km2) area () (m we yrminus1) (Gt yrminus1)

Hengduan Shan 11584 12 minus033plusmn 014 minus38plusmn 38Bhutan 4021 34 minus022plusmn 013 minus09plusmn 05Everest 6226 23 minus026plusmn 014 minus16plusmn 08West Nepal 6849 13 minus032plusmn 014 minus22plusmn 10Spiti Lahaul 9043 23 minus045plusmn 014 minus41plusmn 13Hindu Kush 6135 13 minus012plusmn 016 minus07plusmn 10Karakoram East

1902428 +011 plusmn 014 +19 plusmn 31

Karakoram West 30 +009 plusmn 018Pamir 9369 34 +014 plusmn 014 +13 plusmn 13

TotalArea- 72251 30 minus014plusmn 008 minus101plusmn 55weighted average

Table 5Comparison of geodetic mass balances between this study and previous published results with overlapping study periods and similargeographic locations Updated figures from Kaumlaumlb et al (2012) are averaged over 3 times 3 cells centered over the study sites of the presentstudy

GlacierSite name This study Bolch et al (2011) Nuimura et al (2012) Kaumlaumlb et al (2012 updated)2002ndash2007 2000ndash2008 2003ndash2008

Study Mass balance Area (km2)a Mass balance Area Mass balance Mass balanceperiod (m we yrminus1) (m we yrminus1) (km2) (m we yrminus1) (m we yrminus1)

Hengduan San 1999ndash2010 minus022 plusmn 014 1303 (55 )Bhutan 1999ndash2010 minus022 plusmn 012 1384 (64 ) minus052 plusmn 016b

Everest

1999ndash2011

minus026 plusmn 013 1461 (58 ) minus039 plusmn 011AX010 minus090plusmn 034 04Changri SharNup minus042plusmn 017 161 (79 ) minus029plusmn 052 130 minus055plusmn 038Khumbu minus051plusmn 019 203 (47 ) minus045plusmn 052 170 minus076plusmn 052Nuptse minus037plusmn 020 50 (74 ) minus040plusmn 053 40 minus034plusmn 027Lhotse Nup minus021plusmn 027 24 (70 ) minus103plusmn 051 19 minus022plusmn 047Lhotse minus043plusmn 018 85 (83 ) minus110plusmn 052 65 minus067plusmn 051Lhotse SharImja minus070plusmn 052 98 (44 ) minus145plusmn 052 107 minus093plusmn 060Amphu Laptse minus046plusmn 034 25 (38 ) minus077plusmn 052 15 minus018plusmn 094Chukhung +044 plusmn 024 42 (47 ) +004 plusmn 054 38 +043 plusmn 081Ama Dablam minus049plusmn 017 36 (54 ) minus056plusmn 052 22 minus056plusmn 073Duwo minus016plusmn 026 19 (18 ) minus196plusmn 053 10 minus068plusmn 074Total Khumbu minus041plusmn 021 744 minus079plusmn 052 617 minus045plusmn 060(10 Glaciers above)

West Nepal 1999ndash2011 minus032 plusmn 013 908 (40 ) minus032 plusmn 012Spiti Lahaul 1999ndash2011 minus045 plusmn 013 2110 (46 ) minus038 plusmn 006Karakoram East 1999ndash2010 +011 plusmn 014 5328 (42 ) minus004 plusmn 004Karakoram West 1999ndash2008 +009 plusmn 018 5434 (45 )Hindu Kush 1999ndash2008 minus012 plusmn 016 793 (80 ) minus020 plusmn 006Pamir 1999ndash2011 +014 plusmn 013 3178 (50 )

a The total glacier area is given in km2 and in parenthesis the of the glacier area actually covered with measurementsb For this cell the ICESat coverage is insufficient and does not sample all glacier elevations

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1279

Fig 11 Elevation changes (SPOT5-SRTM) in the ablation area for clean ice (blue circles) and debris-covered ice (black circles) for eachstudy site For the Karakoram (east and west) and the Pamir study sites only non-surge-type glaciers are considered Standard error of theelevation differences are typically less thanplusmn2 m with each 100 m elevation band (not shown for the sake of clarity) Note elevation changesover separate sections of a glacier cannot be treated as mass changes due to the disregard of glacier dynamics

estimate over this short time period The differences in sur-vey periods and in the glacier outlines may also explainpart of these discrepancies In particular the delimitation ofthe accumulation area is notoriously difficult and Bolch etal (2011) did not include some of the steep high altitudeslopes which we included due to their potential avalanchecontribution For the 10 glaciers mentioned above our massbalances would become 012 m we yrminus1 more negative ifwe used the exact same outlines as Bolch et al (2011)Our positive mass balance for the small Chukhung Glacier(+044plusmn 024 m we yrminus1) is in agreement with Nuimura etal (2012) but enigmatic in a context of regional glacier lossand deserves further investigation If the 2002ndash2007 massbalance estimate by Bolch et al (2011) is excluded all otherrecent mass balance estimates suggest no acceleration of themass loss in the Everest area when compared to the long-term(1970ndash2007) mass balance measured by Bolch et al (2011)in this regionminus032plusmn 008 m weyrminus1

In Bhutan our results indicate a stronger mass loss forsouthern glaciers than for northern ones Interestingly in thisarea Kaumlaumlb (2005) measured high speeds (up to 200 m yrminus1)

for glaciers north of the Himalayan main ridge and ratherlow velocities over southern glacier tongues by using repeat

ASTER data This is consistent with the more negative massbalance that we report for the southern glaciers in Bhutanie for the slow flowing presumably downwasting glaciers

Berthier et al (2007) reported a mass balance betweenminus069 andminus085 m we yrminus1 for an ice-covered area of915 km2 around Chhota Shigri Glacier between 1999 (SRTMDEM) and 2004 (SPOT5-HRG DEM) For the same areawe found a mass balance ofminus045plusmn 013 m we yrminus1 over1999ndash2011 The difference in the study period may explainpart of the differences Also the methodology by Berthieret al (2007) did not explicitly take into account the SRTMradar penetration over glaciers and also corrected an appar-ent elevation-dependent bias according to glacier elevationswhich is now known as a resolution issue and is best cor-rected according to the terrain maximum curvature (Gardelleet al 2012b) More details and discussions about these dif-ferences can be found in Vincent et al (2013)

Importantly the results of the eastern Karakoram studysite confirm the balanced or slightly positive mass budgetreported by Gardelle et al (2012a) in the western Karako-ram over the past decade The two Karakoram study sitesare adjacent with a small overlapping area (sim 1100 km2 ofglaciers) where elevation differences agree within 1 m Both

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1280 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 12 Comparison of geodetic mass balances for 10 glaciers inthe Mount Everest area 1999ndash2010 mass balances measured inthe present study are compared with published estimates during2002ndash2007 (Bolch et al 2011) and during 2000ndash2008 (Nuimuraet al 2012) The 1-to-1 line is drawn The mean difference (plusmn1standard deviation) between (i) our values and Bolch et al (2011)are (047plusmn 058 m we yrminus1) and (ii) our values and Nuimura etal (2012) are (012plusmn 021 m we yrminus1) Part of the differences be-tween estimates may be due to the different periods surveyed andthe different glacier outlines used

sites show similar mass balances over slightly different pe-riods +011plusmn 014 m yrminus1 over 1999ndash2010 in the east and+009plusmn 018 m yrminus1 over 1999ndash2008 in the west This con-firms a ldquoKarakoram anomalyrdquo that was first suggested basedon field observations (Hewitt 2005) We report a mass bal-ance of+010plusmn 013 m yrminus1 for Baltoro Glacier (see loca-tion in Fig 8 size= 304 km2) between 1999 and 2010which is consistent with its gradual speed-up reported since2003 (Quincey et al 2009) However given the completelydifferent methods used in our and the latter study we can-not conclude that this demonstrates a real shift from nega-tive to positive mass balance The mass budget of SiachenGlacier (1145 km2 sometimes referred to as the highestbattlefield in the world) is also balanced or slightly pos-itive (+014plusmn 014 m we yrminus1) Thus the alarming state-ment by Rasul (2012) that this glacier retreated by 59 kmand lost 17 of its mass during 1989ndash2009 is very likelyerroneous Again within error bars our regional mass bal-ance in the Karakoram agrees with the mass balance ofminus003plusmn 004 m yrminus1 found by Kaumlaumlb et al (2012) over 2003ndash2008 and with the mass balance ofminus010plusmn 018 m we yrminus1

(2-sigma error level) obtained by Gardner et al (2013) for aregion including both the Karakoram and the Hindu Kush

In the northernmost part of the Pamir in the Alay moun-tain range the mass balance of Abramov Glacier has beenmeasured in the field for 30 yr between 1968 and 1998(WGMS 2012) with a mean value ofminus046 m yrminus1 Herewe report a mass balance ofminus003plusmn 014 m yrminus1 for this

glacier over 1999ndash2011 Our near zero mass budget estimatefor this glacier disagrees with negative mass balances recon-structed for the same period with a model run using biascorrected NCEPNCAR re-analysis data and calibrated withfield data (Barundun et al 2013) An underestimate of ourSRTM C-Band penetration in the Pamir as a whole and inparticular for the small Abramov Glacier may contribute tothese differences

For the whole Pamir study site our mass budget is bal-anced or slightly positive (+014plusmn 013 m we yrminus1) In theeastern Pamir the mass balance of Muztag Ata Glacier was+025 m we yrminus1 between 2005 and 2010 (Yao et al 2012)Although Muztag Ata Glacier is located outside of our Pamirstudy site and his glaciological records only cover the sec-ond half of our study period this measurement is consistentwith our remotely sensed mass balance Thus the ldquoKarako-ram anomalyrdquo should probably be renamed the ldquoPamir-Karakoram anomalyrdquo

This slightly positive or balanced mass budget measuredin the Pamir seems to contradict previous findings by Heidand Kaumlaumlb (2012) They reported rapidly decreasing glacierspeeds in this region between two snapshots of annual ve-locities in 20002001 and 20092010 an observation thatwould fit with negative mass balances (Span and Kuhn 2003Azam et al 2012) We note though that the relationship be-tween glacier mass balance and ice fluxes is not straightfor-ward and might in particular involve a time delay betweena change in mass balance and the related dynamic response(Heid and Kaumlaumlb 2012) Thus the decrease in speed couldwell be due to negative mass balances before or partiallybefore 1999ndash2011 We acknowledge that this discrepancyneeds to be investigated further in particular by examiningcontinuous changes in annual glacier speed through the firstdecade of the 21st Century Indeed Heid and Kaumlaumlb (2012)measured differences between two annual periods which maynot represent mean decadal velocity changes

Jacob et al (2012) using satellite gravimetry (GRACE)measured a mass budget ofminus5plusmn 6 Gt yrminus1 over 2003ndash2010for the ldquoKarakoram-Himalayardquo their region 8c which cor-responds to our study area excluding the Pamir Karakoramand Hindu Kush sub-regions but including some glaciersnorth of the Himalayan ridge already on the Tibetan PlateauFor this subset of our PKH study area we measured amass budget ofminus126plusmn 42 Gt yrminus1 The two estimates over-lap within their error bars especially if our error is dou-bled to match the two standard error level used by Jacob etal (2012) The estimate of Jacob et al (2012) over our com-plete study region (PKH) ieminus6plusmn 8 Gt yrminus1 (sum of theirregions 8b and 8c) also includes mass changes for the KunlunShan sub-region for which we did not measure glacier massbalances Therefore the comparison with our PKH value(minus101plusmn 55 Gt yrminus1) is not straightforward but suggests areasonable agreement especially if we consider the slightmass gain (15plusmn 17 Gt yrminus1) measured with ICESat in theWest Kunlun (Gardner et al 2013)

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1281

Table 6 Annual average of the glacier seasonal and decadal mass loss contributions to Indus Ganges and Brahmaputra discharges Basinarea and seasonal contributions are given by Kaser et al (2010) glacier area in each basin is derived from the RGI v20 (Arendt et al 2012)Numbers in parentheses indicate the percentage of glacierized area within the river basin Seasonal contributions were modeled by Kaser etal (2010) based on the CRU 20 dataset over 1961ndash1990 assuming a balanced glacier mass budget Glacier decadal mass loss contributionsare given as annual average over 1999ndash2011

River Basin area Glacier area Glacier contribution Mean discharge(km2) (km2) (m3 sminus1) (m3 sminus1)

Seasonally delayed Decadal mass lossKaser et al (2010) (this study)

Indus 1 139 814 25 598 (2 ) 105 103plusmn 72 2146 (Chevallier personalcommunication 2012)

Ganges 1 023 609 11 168 (1 ) 47 103plusmn 49 11739 Papa et al (2012Brahmaputra 527 666 15 296 (3 ) 33 147plusmn 64 19710 updated)

54 Reasons for the heterogeneous pattern of massbalance

The PKH glacier mass balance is slightly negative(minus014plusmn 008 m we yrminus1) but changes are not homoge-neous from one sub-region to another and seem to coincidewith the climatic settings of the mountain range For the east-ern sites (from the Hengduan Shan to West Nepal) which areunder the influence of the Indian and south-east Asian sum-mer monsoons the mass balances are moderately negative(sim minus026 m we yrminus1) In the northwest of the PKH at thePamir and Karakoram sites where glacier climate is char-acterized by the westerlies in winter mass balances are bal-anced or slightly positive (sim +010 m we yrminus1) In betweenthese two influences the glaciers of the Spiti Lahaul sitein a monsoon-arid transition zone experienced the strongestmass loss (minus045plusmn 013 m we yrminus1)

This heterogeneous pattern of mass balances seems to berelated to trends in precipitation throughout the PKH Archerand Fowler (2004) reported an increase in winter precipi-tation over the Upper Indus Basin since 1961 a trend con-firmed by Yao et al (2012) over the Pamir and the Karako-ram based on the analysis of the Global Precipitation Cli-matology Project data set (GPCP Adler et al 2003) Thisprecipitation increase is a source of greater accumulation forglaciers and thus a possible explanation for their slightly pos-itive mass budgets In addition in the Upper Indus Basin themean summer temperature has been decreasing since 1961(Fowler and Archer 2006) which is consistent with the find-ings of Shekhar et al (2010) in the Karakoram who reporteda decrease in both maximum and minimum temperature since1988 These increases in winter precipitation and summertemperature likely translate in greater accumulation and re-duced ablation changes which are consistent with the bal-anced or slightly positive glacier mass budget

On the other hand in the east of the PKH the Indiansummer monsoon has been weakening since the 1950s (Bol-lasina et al 2011) which could have reduced the amount

of snowfall over the eastern study sites and thus resultedin negative mass balances In addition maximum and min-imum temperatures increased since 1988 in the western Hi-malaya (Shekhar et al 2010) and maximum temperaturesincreased in the central Himalaya (Nepal) since the mid-1970s (Shrestha et al 1999) Thus both precipitation andtemperature trends in the Himalaya likely contributed to thenegative mass balances measured over the correspondingstudy sites (from Spiti Lahaul to Hengduan Shan Fig 1)

However gridded data sets such as GPCP are not neces-sarily representative of the climate at high altitude IndeedHewitt (2005) reported a 5-to-10-fold increase in precipi-tation between the glacier front and the accumulation areain the Karakoram that is not captured by reanalysis dataas recently confirmed by Immerzeel et al (2012) Thus di-rect measurements of accumulation on High Mountain Asiaglaciers (eg Azam et al 2013) and high-altitude weatherstations are eagerly needed to validate the large scale grid-ded data set (eg reanalysis) test their ability to describe thespecific climate near to PKH glaciers and assess the relativerole played by temperature and precipitation changes

55 Contribution to water resources

Our glacier mass balance can be averaged over large riverbasins to estimate the mean contribution to river runoffdue to a decade of glacier mass loss We assume therebyonly direct river runoff without loss terms For the threerivers for which we cover the entire glacierized area of theircatchments within our sub-regions (Fig 1) these contribu-tions to the river discharge are equal to 103 m3 sminus1 (Indus)103 m3 sminus1 (Ganges) and 147 m3 sminus1 (Brahmaputra) for theperiod 1999ndash2011 (Table 6)

Rivers of PKH being key water resources measures oftheir discharges are highly sensitive and difficult to access forpolitical reasons Papa et al (2012) built recent time series ofdischarges for Ganges and Brahmaputra by using altimetrymeasurements The locations of the corresponding dischargeestimates in Bangladesh are given in Fig 1 We can therefore

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1282 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

evaluate the relative contribution of glacier decadal mass loss(Table 6) to annual river run-off at 09 for the Gangesat Hardinge (basin area ofsim 1 020 000 km2) and 07 forBrahmaputra at Bahadurabad (basin area ofsim 530 000 km2)between 1999 and 2011 Given the discharge measured at theGuddu station by the Surface Water Hydrology Project of theWater and Power Development Authority (Pakistan see lo-cation on Fig 1) between 1999 and 2003 we can estimatethe contribution of glacier mass loss at 48 for the IndusRiver at this station

The seasonal contributions of glaciers to river run-off (iethe part of the annual precipitation that experiences season-ally delayed release from the glaciers) directly results fromseasonal variations of meteorological quantities in contrastto the glacier mass loss contribution which results from thelong-term climatic change Both runoff components directlyaffect the amount of water available in a river at a given lo-cation and at a given time so that their comparison could beof interest Even if the seasonal timing of the glacier massloss contribution is unclear it could in general peak at asimilar time of the year than the seasonal glacier contribu-tion so that both components rather enlarge each other thancancel Seasonal glacier contributions have been modeled byKaser et al (2010) The latter study assumed glaciers to bein equilibrium with the present climate and thus prescribedglacier mass balances equal to 0 Thus their seasonally de-layed glacier contributions do not include the part due todecadal glacier mass loss and is complementary to our es-timates (Table 6) For the eastern basins under the influenceof the Indian summer monsoon (Ganges and Brahmaputra)the contribution of glacier mass loss is higher than the sea-sonal contribution In other words for the first decade of the21st century the mass loss of glaciers is on average moreimportant for water availability in those two rivers than theirseasonal water storage effect The situation is different for theIndus Basin where the glacier melt season is also the dry sea-son which results in a relatively higher seasonally delayedglacier contribution to the river discharge There the season-ally delayed water release from the glaciers and the decadalglacier mass loss contribute on average equally to the riverdischarge Both contributions are strongly (and equally) de-pendent on the location of the discharge measurement andwill be significantly larger (in relative term) in more heav-ily glacierized and smaller mountain catchments located inthe upper part of these major river basins (Bookhagen andBurbank 2010)

6 Conclusion

In this study we assessed the spatial pattern of glaciermass balance over the PKH by comparing the SRTM andSPOT5 DEMs over 9 well-distributed study sites between1999 and 2011 We found slightly positive or balanced massbudgets in the western part (Pamir Karakoram) moder-ate mass loss in the east (Nepal Bhutan Hengduan Shan)

and the most negative mass balance in the monsoon-aridtransition zone (Spiti Lahaul in the western Himalaya)By extrapolating these values to climatically homogeneoussub-regions we estimate the PKH overall mass balanceto have beenminus014plusmn 008 m we yrminus1 between 1999 and2011 which corresponds to a contribution to sea level riseof 0028plusmn 0015 mm yrminus1 This is about 4 of the globalglacier mass loss estimated by Gardner et al (2013) thoughover a slightly shorter time period (2003ndash2009) The largestsource of uncertainties in those geodetic mass balance esti-mates is the poorly constrained penetration of the C-BandSRTM radar signal into snow and ice

Our technique of DEM differencing allows mapping in de-tail the spatial pattern of glacier elevation changes Thosemaps reveal many surge-type behaviors both in the Pamirand the Karakoram It also appears that the glacier-wide massbalance does not seem to be considerably affected by themass transfer from surges over thesim 10 yr time period stud-ied Similar lowering rates over debris-covered and clean iceare observed for four study sites despite the commonly as-sumed insulating effect of debris covers This suggests forthe scale of entire glacier tongues a spatial mixture of re-duced ablation under intact thick debris covers and enhancedablation by supraglacial lakes or ice cliffs or due to thin de-bris layer and very low glacier velocities in ablation areasthat reduce the ice advection downstream

The regionally heterogeneous pattern of ice wastage is inbroad agreement with contrasted trends in large-scale pre-cipitation and temperature However glaciological (eg ac-cumulation) and meteorological records are needed at highaltitude in the vicinity of glaciers to evaluate more closelythe local influence of atmospheric variables on glacier massbalance and fully understand the reasons behind those con-trasted mass budgets in the PKH

Supplementary material related to this article isavailable online at httpwwwthe-cryospherenet712632013tc-7-1263-2013-supplementpdf

Acknowledgements We thank G Cogley A Gardner T NuimuraT Bolch P Wagnon M Barandun M Hoelzle M Huss and ananonymous referee for their constructive comments that led to aconsiderably improved manuscript We thank the Water and PowerDevelopment Authority (WAPDA) for their hydrological data(processed by P Chevallier) and F Papa for sharing his updatedrun-off data ahead of publication We thank T Bolch for sharinghis glacier outlines from the Everest area We thank the USGSfor allowing free access to their Landsat archive CGIAR-SCI forSRTM C-band data DLR for SRTM X-band data and GLIMS forglacier inventories J Gardelle acknowledges a PhD fellowshipfrom the French Space Agency (CNES) and the French NationalResearch Center (CNRS) E Berthier acknowledges support fromCNES through the TOSCA and ISIS proposal no 397 and fromthe Programme National de Teacuteleacutedeacutetection Spatiale E Berthieracknowledges M Masiokas and R Villalba for welcoming him

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1283

at the IANIGLA (Mendoza Argentina) Y Arnaud acknowledgessupport from French National Research Agency through ANR-09-CEP-005-01PAPRIKA A Kaumlaumlb acknowledges support fromESA through Glaciers_cci (400010177810I-AM) and from theEuropean Research Council under the European Unionrsquos SeventhFramework Programme (FP2007-2013)ERC Grant Agreementn 320816

Edited by I M Howat

The publication of this articleis financed by CNRS-INSU

References

Adler R Huffman G Chang A Ferraro R Xie P JanowiakJ Rudolf B Schneider U Curtis S Bolvin D Gruber ASusskind J Arkin P and Nelkin E The Version-2 GlobalPrecipitation Climatology Project (GPCP) Monthly Precipita-tion Analysis (1979ndashPresent) J Hydrometeorol 4 1147ndash11672003

Archer D R and Fowler H J Spatial and temporal variationsin precipitation in the Upper Indus Basin global teleconnectionsand hydrological implications Hydrol Earth Syst Sci 8 47ndash61doi105194hess-8-47-2004 2004

Arendt A Bolch T Cogley J G Gardner A Hagen J-OHock R Kaser G Pfeffer W T Moholdt G Paul F RadicV Andreassen L Bajracharya S Beedle M Berthier EBhambri R Bliss A Brown I Burgess E Burgess DCawkwell F Chinn T Copland L Davies B de AngelisH Dolgova E Filbert K Forester R Fountain A Frey HGiffen B Glasser N Gurney S Hagg W Hall D Hari-tashya U K Hartmann G Helm C Herreid S Howat IKapustin G Khromova T Kienholz C Koenig M KohlerJ Kriegel D Kutuzov S Lavrentiev I LeBris R Lund JManley W Mayer C Miles E Li X Menounos B Mer-cer A Moelg N Mool P Nosenko G Negrete A NuthC Pettersson R Racoviteanu A Ranzi R Rastner P RauF Rich J Rott H Schneider C Seliverstov Y Sharp MSigurethsson O Stokes C Wheate R Winsvold S WolkenG Wyatt F and Zheltyhina N Randolph Glacier Inventory[v20] A Dataset of Global Glacier Outlines Global Land IceMeasurements from Space Boulder Colorado USA Digital Me-dia httpwwwglimsorgRGIrandolphhtml 2012

Azam M Wagnon P Ramanathan A Vincent C Sharma PArnaud Y Linda A Pottakkal J Chevallier P Singh V andBerthier E From balance to imbalance a shift in the dynamicbehaviour of Chhota Shigri glacier western Himalaya India JGlaciol 58 315ndash324 doi1031892012JoG11J123 2012

Azam M F Wagnon P Vincent C Ramanathan A Linda Aand Singh V B Reconstruction of the annual mass balanceof Chhota Shigri Glacier (Western Himalaya India) since 1969Ann Glaciol in revision 2013

Barrand N and Murray T Multivariate Controls on the In-cidence of Glacier Surging in the Karakoram Himalaya

Arct Antarct Alp Res 38 489ndash498 doi1016571523-0430(2006)38[489MCOTIO]20CO2 2006

Barundun M Huss M Azisov E Gafurov A HoelzleM Merkushkin A Salzmann N and Usubaliev R Re-establishing seasonal mass balance observation at AbramovGlacier Kyrgyzstan from 1968ndash2012 Abstract EGU2013-5668European Geophysical Union Vienna 2013

Benn D Bolch T Hands K Gulley J Luckman ANicholson L Quincey D Thompson S Toumi R andWiseman S Response of debris-covered glaciers in theMount Everest region to recent warming and implicationsfor outburst flood hazards Earth-Sci Rev 114 156ndash174doi101016jearscirev201203008 2012

Berthier E and Vincent C Relative contribution of surface mass-balance and ice-flux changes to the accelerated thinning of Merde Glace French Alps over 1979ndash2008 J Glaciol 58 501ndash512doi1031892012JoG11J083 2012

Berthier E Arnaud Y Baratoux D Vincent C and Reacutemy FRecent rapid thinning of the ldquo Mer de Glace rdquo glacier derivedfrom satellite optical images Geophys Res Lett 31 L17401doi1010292004GL020706 2004

Berthier E Arnaud Y Kumar R Ahmad S Wagnon P andChevallier P Remote sensing estimates of glacier mass balancesin the Himachal Pradesh (Western Himalaya India) RemoteSens Environ 108 327ndash338 doi101016jrse2006110172007

Bhambri R Bolch T Kawishwar P Dobhal D P SrivastavaD and Pratap B Heterogeneity in Glacier response from 1973to 2011 in the Shyok valley Karakoram India The CryosphereDiscuss 6 3049ndash3078 doi105194tcd-6-3049-2012 2012

Bhutiyani M Mass-balance studies on Siachen Glacier in theNubra valley Karakoram Himalaya India J Glaciol 45 112ndash118 1999

Bishop M P Schroder J F and Ward J L SPOT multispec-tral analysis for producing supraglacial debris-load estimates forBatura glacier Pakistan Geocarto Int 10 81ndash90 1995

Bolch T Pieczonka T and Benn D I Multi-decadal mass lossof glaciers in the Everest area (Nepal Himalaya) derived fromstereo imagery The Cryosphere 5 349ndash358 doi105194tc-5-349-2011 2011

Bolch T Kulkarni A Kaumlaumlb A Huggel C Paul F Cogley JFrey H Kargel J Fujita K Scheel M Bajracharya S andStoffel M The state and fate of Himalayan Glaciers Science336 310ndash314 doi101126science1215828 2012

Bollasina M Ming Y and Ramaswamy V Anthropogenicaerosols and the weakening of the South Asian summer mon-soon Science 334 502ndash505 doi101126science12049942011

Bookhagen B and Burbank D Toward a complete Himalayan hy-drological budget spatiotemporal distribution of snowmelt andrainfall and their impact on river discharge J Geophys Res115 F03019 doi1010292009JF001426 2010

Bretherton C Widmann M Dymnikov V Wallace J and BladeacuteI The effective number of spatial degrees of freedom of a time-varying field J Climate 12 1990ndash2009 1999

Copland L Pope S Bishop M Shroder J Clendon PBush A Kamp U Seong Y and Owen L Glacier veloc-ities across the central Karakoram Ann Glaciol 50 41ndash49doi103189172756409789624229 2009

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

The Cryosphere 7 1263ndash1286 2013wwwthe-cryospherenet712632013doi105194tc-7-1263-2013copy Author(s) 2013 CC Attribution 30 License

Geoscientiic Geoscientiic

Geoscientiic Geoscientiic

Op

en

Acce

ss

The Cryosphere

Op

en

Acce

ss

Region-wide glacier mass balances over thePamir-Karakoram-Himalaya during 1999ndash2011

J Gardelle1 E Berthier2 Y Arnaud3 and A Kaumlaumlb4

1CNRS ndash Universiteacute Grenoble 1 Laboratoire de Glaciologie et de Geacuteophysique de lrsquoEnvironnement (LGGE) UMR518354 rue Moliegravere BP 96 38402 Saint Martin drsquoHegraveres Cedex France2CNRS Universiteacute de Toulouse LEGOS 14 av Edouard Belin Toulouse 31400 France3IRD ndash Universiteacute Grenoble 1 LTHELGGE 54 rue Moliegravere BP 96 38402 Saint Martin drsquoHegraveres Cedex France4Department of Geosciences University of Oslo PO Box 1047 Blindern 0316 Oslo Norway

Correspondence to E Berthier (etienneberthierlegosobs-mipfr)

Received 29 January 2013 ndash Published in The Cryosphere Discuss 7 March 2013Revised 2 July 2013 ndash Accepted 5 July 2013 ndash Published 9 August 2013

Abstract The recent evolution of Pamir-Karakoram-Himalaya (PKH) glaciers widely acknowledged as valu-able high-altitude as well as mid-latitude climatic indi-cators remains poorly known To estimate the region-wide glacier mass balance for 9 study sites spread fromthe Pamir to the Hengduan Shan (eastern Himalaya) wecompared the 2000 Shuttle Radar Topography Mission(SRTM) digital elevation model (DEM) to recent (2008ndash2011) DEMs derived from SPOT5 stereo imagery Dur-ing the last decade the region-wide glacier mass bal-ances were contrasted with moderate mass losses in theeastern and central Himalaya (minus022plusmn 012 m we yrminus1 tominus033plusmn 014 m we yrminus1) and larger losses in the west-ern Himalaya (minus045plusmn 013 m we yrminus1) Recently reportedslight mass gain or balanced mass budget of glaciersin the central Karakoram is confirmed for a larger area(+010plusmn 016 m we yrminus1) and also observed for glaciersin the western Pamir (+014plusmn 013 m we yrminus1) Thus theldquoKarakoram anomalyrdquo should be renamed the ldquoPamir-Karakoram anomalyrdquo at least for the last decade The overallmass balance of PKH glaciersminus014plusmn 008 m we yrminus1 istwo to three times less negative than the global average forglaciers distinct from the Greenland and Antarctic ice sheetsTogether with recent studies using ICESat and GRACE dataDEM differencing confirms a contrasted pattern of glaciermass change in the PKH during the first decade of the21st century

1 Introduction

The Pamir-Karakoram-Himalaya (PKH) mountain ranges arecovered by more than 70 000 km2 of glaciers (Arendt et al2012) Assessing glacier evolution over such a large andremote region is challenging but nevertheless required tocharacterize the impacts of climate change in the region(eg Bolch et al 2012) to assess glacial contribution to wa-ter resources (eg Immerzeel et al 2010 Kaser et al 2010)and global sea level rise (eg Kaumlaumlb et al 2012 Gardner etal 2013) and ultimately to reliably project glacier responseto 21st century climate changes (eg Radic and Hock 2011Marzeion et al 2012)

Length changes have been measured for about 100 glaciersin the PKH and revealed predominant retreat of glacier frontssince the mid-19th century except in the Karakoram (Scher-ler et al 2011a Bhambri et al 2012 Bolch et al 2012)The majority of glaciers lost area over the past decades andin most cases the rates of area loss have been increasing inrecent years (eg Bolch et al 2012) But changes in glacierlength and area should be treated with care when used toevaluate the impact of climate change on glaciers especiallyin the presence of surge-type andor debris-covered glacierswhich are common in the PKH (Yde and Paasche 2010)

Rather mass balance is the most relevant variable to assessglacier responses to climate variability (Oerlemans 2001Vincent 2002) Mass balance estimates however remainscarce in the PKH and may not adequately sample the widerange of climates and glacier responses in the region Five

Published by Copernicus Publications on behalf of the European Geosciences Union

1264 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

different methods of mass balance measurements have beenused in the PKH

i Glaciological (or traditional) measurements are rela-tively short-termed in general less than 10 yr mainlybecause of difficult access to generally remote glaciersIn addition measurement series are often biased to-wards small-to-medium-sized and debris-free glaciersfor logistical reasons (eg Dobhal et al 2008 Fujitaand Nuimura 2011 Azam et al 2012 Bolch et al2012 Yao et al 2012) Vincent et al (2013) discussedin detail the inadequate spatial and temporal samplingof those glaciological measurements in the western Hi-malaya Gardner et al (2013) have shown that during2003ndash2009 the extrapolation of those few and short-term glaciological mass balances to entire High Moun-tain Asia lead to an overestimation of the mass loss byabout a factor of 3

ii The accumulation area ratio (AAR) method relies heav-ily on the availability of glaciological mass balancesduring various years in order to establish an empiri-cal relationship between annual mass balance and AAR(eg Rabatel et al 2005) The extrapolation to un-measured glaciers is not straightforward This methodhas been applied by Kulkarni et al (2011) to 19 glaciersin north-west India

iii The hydrological method has only been used once inthe PKH to our knowledge providing 5 yr of mass bal-ance for Siachen Glacier Karakoram (Bhutiyani 1999)and is very difficult to implement due to a lack of accu-rate precipitation and runoff measurements at high alti-tude over the PKH (Bolch et al 2012 Immerzeel et al2012)

iv The geodetic method based on the comparison of topo-graphic data (DEM or laser altimetry) can be applied atregional scales in remote areas using satellite data suchas from ICESat SPOT5 or SRTM This method enablesan increase in the number and diversity of glacier mea-surements (eg in terms of glacier size elevation as-pect percentage of debris cover) (Berthier et al 2007Bolch et al 2011 Kaumlaumlb et al 2012 Nuimura et al2012 Pieczonka et al 2013) An important remaininglimitation is the difficulty to convert volume changes tomass changes (eg Huss 2013)

v The gravimetric method uses the data from the GravityRecovery and Climate Experiment (GRACE) missionlaunched in 2002 Gravity fields need to be corrected forthe influence of the glacial isostatic adjustment (GIA)and the hydrological changes off-glacier (eg Zhang etal 2013a) before the glacier mass change can be ex-tracted This method initially led to contrasted glaciermass budgets in High Mountain Asia (Matsuo and Heki

2010 Jacob et al 2012) The two most recent GRACEmass budgets for High Mountain Asia agree within their2σ error bars (minus14plusmn 17 Gt yrminus1 andminus23plusmn 24 Gt yrminus1

in Gardner et al 2013)

The ICESat- and GRACE-based methods are appropriateto assess the regional mass budget for large glacierized re-gions such as the PKH but also have their specific limita-tions The ICESat sampling is restricted to satellite orbitsthat are separated by tens of kilometers at PKH latitudes andthus does not allow estimating the mass balance of individ-ual glaciers Glacier changes can only be investigated overthe lifetime of ICESat 2003ndash2009 The gravimetric methodhas a coarse spatial resolution (typicallysim 400 km) so thatindividual glaciers and small hydrological basins cannot beresolved

The aim of our study is to present a homogeneous surveyof regional glacier mass balances along the PKH between1999 and 2011 using the geodetic method This is achievedby differencing two digital elevation models (DEMs) over9 study sites selected to be representative of the PKH cli-matic and glaciological diversity This allows for an estima-tion of the spatial pattern of glacier elevation changes alongthe 3000 km-long group of mountain ranges as well as theinfluence of debris cover on thinning rates Finally we ex-trapolate these measurements to provide a new estimate ofthe mass budget of PKH glaciers and their contribution toregional hydrology and sea level rise during the first decadeof the 21st century

Compared to Kaumlaumlb et al (2012) the length of the studyperiod has been doubled and the study area has been signif-icantly extended towards the east (Hengduan Shan China)and the west (Pamir Tajikistan) We also provide a nearly ex-haustive coverage of glacier elevation changes for our 9 studysites as opposed to the sampling by ICESat along satellitetracks only However contrary to the ICESat studies (Kaumlaumlbet al 2012 Gardner et al 2013) our method is limited bythe availability of SPOT5 DEMs on selected sites and doesnot provide insight into the seasonal and annual evolution ofglacier mass balance

2 Study area

Mass balances are investigated for 9 study sites spread alongthe PKH mountain ranges to capture the climatic and glacio-logical variability of the region The location of each studysite is displayed in Fig 1 as well as the extent of the cor-responding sub-regions used to extrapolate the results tothe whole PKH range (see Sect 34) PKH glacier meltwa-ter contributes to five major rivers of Asia (BrahmaputraGanges Indus Tarim and Amu Darya) whose catchment ar-eas are also presented in Fig 1

Each study site corresponds to the extent of a SPOT5 DEM(see Sect 31) A complete coverage of all glaciers in the

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1265

Fig 1 Overview of the extent and location of the nine study sites (black polygons) and corresponding sub-regions (red polygons) along thePKH range Brownish-filled background polygons represent the major river basins of the area Triangles indicate the location of dischargemeasurements discussed in Sect 55

PKH range could not be achieved because of funding restric-tions (SPOT5 is a commercial satellite) the busy acquisitionschedule of this satellite and cloud coverage A further con-strain is that the optical stereo images used to compute theDEMs should be acquired as close as possible to the end ofthe glacier ablation season In practice a time window of2ndash3 months is allowed for the acquisitions which is shortgiven the 26-day orbital cycle of SPOT5 Thus a maximumof 3 to 4 acquisition attempts are possible each year and onlyif the satellite is not programmed for higher priority targets(eg crisis situations) Despite these difficulties our SPOT5DEMs sampled in total 22 200 km2 of glaciers about 30 of the total glacierized area in the PKH (Arendt et al 2012)

The eastern-most sites (Hengduan Shan Bhutan Everestand West Nepal) are strongly influenced by the Indian andsouth-east Asian summer monsoons Ablation and accumula-tion of monsoon-type glaciers occur during the summer sea-son (eg Fujita 2008) Towards the other end of the PKHin the north-west (Pamir Hindu Kush and Karakoram) theclimate is dominated by the westerlies and glaciers are ofwinter-accumulation type The Spiti Lahaul site lies in a tran-sition zone influenced both by the monsoon and the wester-lies The topography also plays a strong role in the moisturetransfer as it dries out the southerly air flow and can preventthe air mass from travelling further north which results ina northward decrease of precipitation (Bookhagen and Bur-bank 2010)

Monsoon-type glaciers are expected to be more sensitivethan winter-accumulation type glaciers to a change in therainsnow limit driven by an increase in temperature (Fujita2008) Indeed a summer warming would increase the alti-

tude of the rainsnow limit and reduce snow accumulation onglaciers as well as decrease their surface albedo (less freshsnow with a high albedo) and thus enhance ablation

The Karakoram and the Pamir are known to host numer-ous surge-type glaciers (Barrand and Murray 2006 Hewitt2007 Kotlyakov et al 2008 Gardelle et al 2012a) char-acterized by the alternation between short active phases in-volving rapid mass transfer from high to low elevationsand longer quiescent phases of low mass fluxes Coplandet al (2011) reported a doubling in the number of glaciersurges after 1990 in the Karakoram with a total of 90 surge-type glaciers observed in the region since the late 1960sIn the Pamir an inventory from 1991 revealed 215 glacierswith signs of repeated surging (looped moraines fast ad-vance of the front) and 51 additional actively surging glaciers(Kotlyakov et al 2008)

Many PKH glaciers are heavily debris-covered in their ab-lation areas because of steep rock walls that surround themand intense avalanche activity (eg Bolch et al 2012) Thesize of these debris zones is highly variable and debris thick-ness can range from a few centimeters (dust or sand) to sev-eral meters (Mihalcea et al 2008 Zhang et al 2013b) Themean debris cover is about 10 of the total glacier area inthe Karakoram and Himalaya (Bolch et al 2012) and culmi-nate at 36 in the Central Himalaya (Scherler et al 2011a)

In the eastern PKH glacier termini are often connected topro-glacial lakes storing meltwater behind frontal morainesor dead-ice dams whereas in the western PKH pro-glaciallakes are less numerous and glacier ablation areas are oftenscattered with supra-glacial lakes (Gardelle et al 2011) re-sulting from successive coalescence of small melting ponds

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1266 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Table 1 Characteristics of the nine study sites Numbers between parentheses indicate the percentage difference between our user-definedglacier inventory and the RGI v20 (Arendt et al 2012)

Site name Glacier area Debris ( of Glacier area Arendt Date Landsat imagesthis study (km2) glacier area) et al (2012) (km2) (SPOT5 DEM)

PathRow Date

Hengduan Shan 1303 11 2541 (+88 ) 24 Nov 2011 p135r39 23 Sep 1999Bhutan 1384 11 1429 (+3 ) 20 Dec 2010 p138r40 19 Dec 2000

p138r41Everest 1461 22 1400 (minus4 ) 4 Jan 2011 p140r41 30 Oct 2000West Nepal 908 17 802 (minus12 ) 3 Jan 2011 p143r39 1 Aug 2009

p143r40Spiti Lahaul 2110 13 2257 (+7 ) 20 Oct 2011 p147r37 15 Oct 2000

p147r38Hindu Kush 793 7 724 (minus11 ) 17ndash21 Oct 2008 p150r34 15 Aug amp

p150r35 16 Sep 1999Karakoram East 5328 10 5384 (+1 ) 31 Oct 2010 p148r35 7 Sep 1998Karakoram West 5434 12 5687 (+1 ) 3 Dec 2008 p149r35 29 Aug 1998Pamir 3178 11 3210 (minus1 ) 29 Nov 2011 p152r33 16 Sep 2000

3 Data and methods

31 Digital elevation models

For each study site we used the Shuttle Radar TopographicMission (SRTM) version 4 DEM as the reference topogra-phy (Rabus et al 2003) This data set acquired in Febru-ary 2000 comes along with a mask to identify the pixelswhich have been interpolated (both were downloaded fromhttpsrtmcsicgiarorg) On each site the SRTM DEM issubtracted from a more recent DEM built from a stereo pairacquired by the HRS sensor onboard the SPOT5 satellite(Korona et al 2009) except for the Hindu Kush study sitewhere the recent DEM has been created from a stereo pairacquired by the HRG sensor also onboard SPOT5 (Berthieret al 2007) The date of each SPOT5 DEM acquisition dif-fers depending on the study site (Table 1) but 7 out of 9 wereacquired between October 2010 and November 2011 EachSPOT5-HRS DEM is also provided with a correlation maskand an ortho-image generated from the rear HRS image

SPOT5-HRS DEMs were produced by the French Map-ping Agency (IGN) using two sets of correlation parametersdefined during the SPIRIT (SPOT 5 stereoscopic survey ofPolar Ice Reference Images and Topographies) internationalpolar year project (Korona et al 2009) Thus two versionsof the DEM are delivered with their respective mask one op-timized (v1) for a gentle topography and the second one (v2)for a more rugged relief Earlier work has shown that v2 hasa slightly larger percentage of interpolated pixels but for thenon-interpolated pixels smaller errors (Korona et al 2009)This version (v2) is preferred here Given that both DEMsare calculated from the same image pair little difference isexpected between v1 and v2 However locally we identifiedsome large bull eye elevation differences between the two

versions of the DEM both off and on glaciers Comparisonwith the SRTM DEM shows that both versions of the DEMare erroneous for those areas and thus we preferred to ex-clude from further analysis all pixels for which the absoluteelevation difference between the two versions of the DEMis larger than 5 m About 20 of the valid DEM pixels areexcluded through that procedure

The SRTM DEM and mask originally at a 3 arcsecondresolution (sim 90 m) are resampled bilinearly to 40 m (UTMprojection WGS84 ellipsoid) to match the posting and pro-jection of the SPOT5 DEM Altitudes are defined above theEGM96 geoid for both DEMs

32 Glacier mask

Where available glacier outlines were downloaded fromthe GLIMS database (Raup et al 2007 Supplement) andmanually corrected if needed Otherwise the glacier maskwas derived from Landsat-TM and Landsat-ETM+ imagesPathrow and dates of Landsat acquisition are given in Ta-ble 1 for each study site A threshold varying between 05and 07 depending on the study site was applied to the Nor-malized Difference Snow Index (TM2-TM5)(TM2+TM5)to detect clean ice and snow automatically whereas debris-covered parts were digitized manually by visual interpreta-tion (eg Racoviteanu et al 2009) The total glacier areaand debris-covered part can be found for each study site inTable 1

Most inventories are derived from Landsat images withminimal snow cover from around the year 2000 prior to the2003 failure of the Scan Line Corrector of the ETM+ sensoronboard Landsat 7 that resulted in image striping Glacierareas are expected to have experienced minor changes sincethen except for glacier fronts that retreated due to the rapid

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1267

Table 2 Elevation ranges equilibrium line altitudes (ELA) and accumulation area ratios (AAR) for the nine study sites The dates ofacquisition of Landsat images used for ELA determination are also given

Sub-region Glaciers elevation ELA (m) AAR Landsat imagesrange (m)

Min Max This study Scherler et Kaumlaumlb et This study Kaumlaumlb et PathRow Dateal (2011a) al (2012) al (2012)

Hengduan Shan 2630 6860 4970plusmn 320 NA NA 55 NA p135r39 23 Sep 1999Bhutan 4390 7480 5690plusmn 440

5700 555036

47p138r40 17 Nov 2000

Everest 4260 8380 5840plusmn 320 31 p140r41 31 oct 2009West Nepal 3950 6870 5590plusmn 138 37 p143r39 1 Aug 2009Spiti Lahaul 3780 6360 5390plusmn 140 5103 5500 34 35 p147r37 2 Aug 2002

p147r38Hindu Kush 2717 6385 5050plusmn 160 5138 30 p150r34 2 Sep 2000

p150r35Karakoram east 3230 7770

5030plusmn 280 4845 554076

47p148r35 7 Sep 1998

Karakoram west 2910 7920 66 p149r35 29 Aug 1998Pamir 2800 7090 4580plusmn 250 NA NA 52 NA p152r33 30 Jul 2000

growth of connected lakes and for glaciers that surged andthus advanced For the few glaciers that advanced their frontposition was updated manually based on the SPOT5 ortho-image

For each study site we favored our user-defined glacierinventories instead of the global Randolph Glacier Inventory(RGI v20 Arendt et al 2012) because the latter is hetero-geneous and for some study sites did not match our accu-racy requirements for the adjustments and corrections of theDEMs

We also estimated the altitude of the equilibrium line(ELA) as the snowline on thesim 2000 Landsat data with min-imal snow cover to separate ablation and accumulation ar-eas The ELA is assumed to be identical to the altitude ofthe snowline at the end of the ablation season (eg Raba-tel et al 2005) after the end of the monsoons or before thefirst snowfalls Therefore we manually digitized snowlinesfor more than 30 glaciers for each study site and computedtheir mean elevations Our ELAs are slightly different fromthose given by Scherler et al (2011a) and Kaumlaumlb et al (2012)probably due to the sensitivity of the ELA to the choice ofthe image (Table 2) Ideally the ELA should have been es-timated from images acquired at the end of the ablation sea-son during various years covering our whole study periodThis would be a highly time-consuming task that was not ac-complished here However we stress that in our study theELA is only used to compare elevation changes on the abla-tionaccumulation areas (Sect 41) and to estimate the sen-sitivity of our mass balances to the choice of the volume-to-mass conversion factor (Sect 35) Our preferred mass bal-ance estimate uses a unique volume-to-mass conversion fac-tor for the whole glacier and is thus independent from theELA

33 Adjustments and corrections of DEM biases

Different DEM processing steps are followed to extract un-biased glacier elevation changes for each study site (eg Nuthand Kaumlaumlb 2011 Gardelle et al 2012b)

331 Planimetric adjustment of the DEMs

A horizontal shift between two DEMs results in an aspect-dependent bias of elevation differences (Nuth and Kaumlaumlb2011) Here the horizontal shift is determined by minimizingthe root mean square error of elevation differences (SPOT5-SRTM) off glaciers where the terrain is assumed to be stableover the study period (Berthier et al 2007) The SRTM DEMis then resampled (cubic convolution) according to the shift

332 Alongacross track corrections (drift of thesatellite orbit)

For DEMs derived from stereo imagery the satellite acquisi-tion geometry can induce a bias along andor across the trackdirection (Berthier et al 2007) We estimate the azimuth ofeach SPOT5 ground track and use it to rotate the coordinatesystem accordingly (Nuth and Kaumlaumlb 2011) Then we com-pute the elevation differences along and across the satellitetrack on stable areas and when necessary correct the bias us-ing a 5th order polynomial fit

333 Curvature correction

As suggested by Paul (2008) and Gardelle et al (2012b) thedifference in original spatial resolutions of the two DEMs canlead to biases related to altitude in mountainous areas Whenthe curvature the first derivative of the slope is high (sharppeaks or ridges) the altitude tends to be underestimatedby the coarse DEM unable to reproduce high-frequency

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1268 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

slope variations Thus this apparent ldquoelevation biasrdquo is cor-rected using the relation between elevation differences andthe maximum terrain curvature estimated over stable areasoff glaciers (Gardelle et al 2012b) Note that the units inFig 1b in Gardelle et al (2012b) should be 10minus2 mminus1 in-stead of mminus1 (A Gardner personal communication 2012)However this does not impact the validity of the correction

334 Radar penetration correction

The SRTM DEM acquired in C-band potentially underesti-mates glacier elevations since at this wavelength (sim 56 cm)the penetration of the radar signal into snow and ice canreach several meters (eg Rignot et al 2001) Here we esti-mate this penetration for each study site by differencing theSRTM C-band (57 GHz) and X-band (97 GHz) DEMs Thelatter was acquired simultaneously with the SRTM X-bandat higher resolution (30 m) but with a narrower swath andthus has a coverage restricted to selected swaths only Sincethe penetration of the X-band is expected to be low com-pared to the C-band (an hypothesis that still needs to be con-firmed Ulaby et al 1986) we consider the elevation differ-ence (SRTMC-bandminus SRTMX-band) to be a first approxima-tion of C-band radar penetration over glaciers (Gardelle etal 2012b) The correction is calculated as a function of alti-tude and applied to each pixel separately

For the Hindu Kush Spiti Lahaul and West Nepal sitesdue to the lack of SRTMX-band data we were unable to de-termine the value of the penetration Thus we applied thecorrection computed over the nearest study site ie Karako-ram for Spiti Lahaul and Hindu Kush and Everest for WestNepal The mean values of the SRTMC-bandpenetrations overglaciers are given for each study site in Table 3

335 Seasonality correction

SPOT5 DEMs were acquired between late October and earlyJanuary depending on the study site Since the SRTM DEMis from mid-February we need to account for possible masschanges during this 1-to-5-month period in order to esti-mate glacier mass balance over an integer number of years(Gardelle et al 2012a) For the eastern sites (from WestNepal to Hengduan Shan) where glaciers are of summer-accumulation types (Fujita 2008) the correction was set to0 This choice is confirmed by recent field observations thatshow no winter accumulation on Mera and Pokalde glaciersin Nepal (Wagnon et al 2013) For the Karakoram studysites we used the winter accumulation rate+013 m we permonth measured on Biafo Glacier (Wake 1989) For theother western study sites (the Pamir Hindu Kush and SpitiLahaul sites) the correction is derived from the mean of win-ter mass balances of 35 glaciers all in the Northern Hemi-sphere+089 m we yrminus1 over 2000ndash2005 corresponding to+015 m we per winter month (Ohmura 2011) The lattervalue is slightly lower than the average winter mass balance

Table 3 Average SRTM C-band penetration estimates (in m)in February 2000 over glaciers in this study and from Kaumlaumlb etal (2012)

Sub-region This study Kaumlaumlb et al (2012)

Hengduan Shan 17 NABhutan 24

25plusmn 05Everest 14West Nepal NA

15plusmn 04Spiti Lahaul NAHindu Kush NA 24plusmn 04Karakoram East

34 24plusmn 03Karakoram WestPamir 18 NA

(+135plusmn 031 m we yrminus1) measured on Abramov Glacier inthe Pamir during 1968ndash1998 (WGMS 2012) and in reason-ably good agreement with the modeled mean winter massbalance (+108plusmn 034 m we yrminus1) of Chhota Shigri Glacierduring 1969ndash2012 in the western Himalaya (Azam et al2013)

336 Mean elevation changes and mass balancecalculation

Before averaging elevation changes we exclude all interpo-lated pixels in the SRTM or SPOT5 DEMs as well as unex-pected elevation changes exceedingplusmn150 m for study sites(Pamir and Karakoram) that include surge-type glaciers orplusmn80 m for other sites Those thresholds were chosen aftercareful inspection of the elevation difference maps and his-tograms in order to identify the largest realistic glacier eleva-tion changes Glaciers that are truncated at the edges of theDEM are also excluded from mass balance calculations asthey may bias the final results if for example only their accu-mulation or ablation area is sampled Although they are trun-cated Siachen (Karakoram East) and Fedtchenko (Pamir)glaciers were retained because of their wide coverage andbecause their accumulation and ablation areas were correctlysampled in the DEMs

Elevation changes are then analyzed for 100 m altitudebins Within each altitude band we average the elevationchanges after excluding pixels where absolute elevation dif-ferences differ by more than three standard deviations fromthe mean (Berthier et al 2004 Gardner et al 2012) Thisis an efficient way to exclude outliers based on the assump-tion that elevation changes should be similar at a given alti-tude within each study site This assumption does not holdfor regions containing actively surging glaciers Thus forthe Pamir and Karakoram study sites surge-type glaciersare treated separately ie elevation changes are averaged by100 m altitude bands for each surge-type glacier individually

Where no elevation change is available for a pixel weassign to it the value of the mean elevation change of the

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1269

altitude band it belongs to in order to assess the mass bal-ance over the whole glacier area

The conversion of elevation changes to mass balance re-quires knowledge of the density of the material that has beenlost or gained and its evolution during the study period Giventhe lack of repeat measurement of density profiles over theentire snowfirnice column in the PKH we applied a con-stant density conversion factor of 850 kg mminus3 for all ourstudy sites (Sapiano et al 1998 Huss 2013)

The mass balance of each surge-type glacier is computedseparately and added (area-weighted) to the mass balance ofthe rest of the glaciers to estimate the mass balance of thewhole study site

For study sites with a high concentration of growingpro-glacial lakes such as Bhutan Everest and West Nepal(Gardelle et al 2011) our mass losses are slightly underes-timated because they do not take into account the glacier icethat has been replaced by water during the expansion of thelake Subaqueous ice losses are only estimated for LhotseSharImja Glacier (Everest area Fujita et al 2009) for thesake of comparison to previous studious

The region-wide mass balance of PKH glaciers as definedin Sect 2 is computed by extrapolating the mass balancefrom each study site to a wider sub-region (Fig 1) that is as-sumed to experience the same glacier mass changes becauseof its similar climate (Bolch et al 2012 Kaumlaumlb et al 2012)The total glacier area for each sub-region is computed fromthe RGI v20 (Arendt et al 2012)

The periods over which we calculate mass balance areslightly different from one site to another (Table 1) depend-ing on the year of acquisition of the SPOT5 DEM (two in2008 two in 2010 and five in 2011) In the following we willassume that all mass balances are representative of the period1999ndash2011 in order to estimate a region-wide mass budgetof PKH glaciers neglecting thus part of the inter-annual vari-ability

34 Accuracy assessment

The errorE1hi to be expected for a pixel elevation change

1hi is assumed to equal the standard deviationσ1h of themean elevation change1h of the altitude band it belongs toThe value ofσ1h can range fromplusmn4 m toplusmn 20 m dependingon the study site and elevation This is rather conservative asthe value ofσ1h contains also the intrinsic natural variabilityof elevation changes within the altitude band (ie it containsboth noise and real geophysical signal)

The errorE1h of the mean elevation change1h in eachaltitude band is then calculated according to standard princi-ples of error propagation

E1h =E1hiradicNeff

(1)

Neff represents the number of independent measurements inthe altitude bandNeff will be lower thanNtot the total num-

ber of elevation change measurements1hi in the altitudeband since the latter are correlated spatially (Bretherton etal 1999)

Neff =Ntot middot PS

2d (2)

where PS is the pixel size (here 40 m)d is the distance ofspatial autocorrelation determined using Moranrsquos I autocor-relation index on elevation differences off glaciers On theaverage over the 9 study sitesd = 492plusmn 72 m

The error on the penetration correction was assumed to besystematic due to the not fully understood physics involvedand equal toplusmn 15 m This error was obtained by compar-ing the SRTM C-band penetration estimated using two inde-pendent methods (i) SRTMC-bandndash SRTMX-band and (ii) dif-ference between ICESat and SRTM when ICESat elevationchange trends during 2003ndash2008 are extrapolated to Febru-ary 2000 (Kaumlaumlb et al 2012) The maximum difference of thepenetration inferred from the two methods is 10 m (Table 3)A 50 additional error margin was added to account for thefact that this comparison could only be made on 2 of ourstudy sites

Given the slender observational support for the seasonal-ity correction (Section 33) we assumed its uncertainty tobe plusmn100 on the western sites ieplusmn015 m we per win-ter month The same absolute uncertainty (plusmn015 m we perwinter month) was used for the eastern sites where no sea-sonality correction is applied but ablation andor accumula-tion may occur in winter (Wagnon et al 2013)

All errors are summed quadratically within each altitudeband and then summed for all altitude bands assuming con-servatively that they are 100 dependent The error on theSRTM penetration correction clearly dominates our errorbudget for volume change

During the conversion from volume to mass we assumea plusmn60 kg mminus3 error on the density conversion factor (Huss2013) Thisplusmn7 error is similar to the one estimated byBolch et al (2011) In a sensitivity test the volume to massconversion is also performed using two alternative densityscenarios (Kaumlaumlb et al 2012) (i) a density of 900 kg mminus3 isassumed everywhere and (ii) 600 kg mminus3 in the accumula-tion area and 900 kg mminus3 in the ablation area These two sce-narios take among other things into account that elevationchanges could be due to changes in glacier dynamics or dueto surface mass balance We then calculate the maximum ab-solute difference between the mass balances calculated usingour preferred scenario (850plusmn 60 kg mminus3 everywhere) and themass balances calculated using the two others scenarios Forthe eastern sites (West Nepal to Hengduan Shan) the max-imum absolute difference is 003 m we yrminus1 For the west-ern sites (Pamir to Spiti Lahaul) it reaches 006 m we yrminus1Those uncertainties due to the choice of a given density sce-nario remain negligible compared to other sources of errors

We also include an error due to extrapolation of our massbalance estimate from our study site (the extent of each

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1270 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

SPOT5 DEM) to the unmeasured glaciers within each sub-region To estimate this regional extrapolation error we com-pare the ICESat elevation change trends (computed as inKaumlaumlb et al 2012) within 3 times 3 cells centered at the studysites (Table 5) with the ICESat-derived trends for the entiresub-regions The absolute difference between both trends isless than 002ndash003 m yrminus1 for Bhutan West Nepal Spiti La-haul and Karakoram ie negligible For Everest the eleva-tion trend for the entire sub-region is 012 m yrminus1 less nega-tive than for the 3 times 3 cells around the SPOT5 DEM cen-ter for Hindu Kush the trend for the entire sub-region is017 m yrminus1 more negative than for the 3 times 3 cell due to thevariability of elevation changes within this sub-region (Kaumlaumlbet al 2012) The regional extrapolation error is obtainedas the area-weighted mean absolute difference between themass balances derived from ICESat for those 3 times 3 cellsand the whole sub-region and equalsplusmn004 m we yrminus1

Finally for sub-regions and total PKH estimates of masschange we include an error on the glacier area computedfrom the RGI v20 (Arendt et al 2012) The inventory er-ror is specific to each study site and evaluated based on thecomparison of our user-defined inventory on each study sitewith the RGI v20 (Table 1) We note the remarkable accu-racy of the RGI for all but one of our study sites The rel-ative errors are generally of a few percents and up to 12 for West Nepal The differences between our and existing in-ventories are probably due to the difficulty of delimitatingdebris-covered glacier parts (eg Frey et al 2012 Paul etal 2013) and accumulation areas The Hengduan Shan studysite is an exception There the RGI is based on the ChineseGlacier Inventory (Shi et al 2009 Arendt et al 2012) andoverestimates the ice-covered area by 88 compared to ourLandsat-based inventory

Unless stated otherwise all error bars from this study andpreviously published studies correspond to one standard er-ror

4 Results

41 Overview of the elevation changes

The maps of glacier elevation changes are given in Figs 2ndash10for our nine study sites with as inset the distribution ofthe elevation differences off glaciers The full maps of ele-vation differences off glaciers are shown in the Supplement(Figs S2ndashS10) Those off glacier maps indicate a local ran-dom noise of aboutplusmn5ndash10 m However at length scales of afew kilometers the spatial variations in elevation differencesare small typically less than 1 m

For the four eastern study sites from the HengduanShan to the West Nepal (Fig 1) the elevation changesaveraged over the accumulation areas are small (between003plusmn 014 m yrminus1 to 013plusmn 015 m yrminus1) whereas clear sur-face lowering is observed for all four study sites in their

ablation areas ranging fromminus050plusmn 014 m yrminus1 (Bhutan)to minus081plusmn 013 m yrminus1 (Hengduan Shan) Over the BhutanEverest and West Nepal study sites glaciers in contact witha proglacial lake often experienced larger thinning rates

Further west the Spiti Lahaul is the only site wheresurface lowering is measured both in the accumula-tion area (minus034plusmn 016 m yrminus1) and the ablation area(minus070plusmn 014 m yrminus1) Clear surface lowering is also ob-served in the ablation areas of the Hindu Kush glaciers(minus050plusmn 018 m yrminus1)

In the Karakoram (east and west Figs 8 and 9) andin the Pamir (Fig 10) the pattern of elevation changesis highly heterogeneous due to the occurrence of numer-ous glacier surges or similar flow instabilities The eleva-tion changes in the accumulation areas of those three studysites are slightly positive (between+022plusmn 014 m yrminus1 and+030plusmn 018 m yrminus1) In the ablation areas small surfacelowering is observed for the two Karakoram study sites (av-erage of the two sitesminus033plusmn 016 m yrminus1) whereas no el-evation change is found in Pamir (minus002plusmn 013 m yrminus1)

Note that those mean rates of elevation changes are sensi-tive to the choice of the ELA assumed to equal the snowlinealtitude in Landsat images from around year 2000

42 Influence of a debris cover

To evaluate thinning rates over debris-covered and debris-free ice we perform a histogram adjustment in order tocompare data sets with similar altitude distributions This isneeded because debris-covered parts tend to be concentratedat lower elevations The adjustment consists in randomly ex-cluding pixels over debris-free ice from each elevation bandto match a scaled version of the debris-covered pixel distribu-tion (Kaumlaumlb et al 2012) Elevation changes with altitude canthen be compared for clean and debris-covered ice (Fig 11)

The thinning is higher for debris-covered ice in the Ever-est area and on the lowermost parts of glaciers in the PamirBut on average in the Pamir western Karakoram and SpitiLahaul the lowering of debris-free and debris-covered ice issimilar The thinning is stronger over clean ice in the Heng-duan Shan Bhutan West Nepal Hindu Kush and althoughto a lesser extent in the eastern Karakoram

43 Mass balance over the nine study sites

Glacier mass changes have been heterogeneous overthe PKH since 1999 Slight mass gains or balancedmass budgets are found in the north-west in thePamir (+014plusmn 013 m we yrminus1) and for the two studysites in the central Karakoram (+009plusmn 018 m we yrminus1

and +011plusmn 014 m we yrminus1) Mass loss is moderatein the adjacent Hindu Kush with a mass balance ofminus012plusmn 016 m we yrminus1 Glaciers in all our eastern studysites (two in Nepal Bhutan and Hengduan Shan) experi-enced homogeneous mass losses with mass balances ranging

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1271

Fig 2 Glacier elevation changes over the Hengduan Shan study site Inset distribution of the elevation differences off glacier The medianand the standard deviation of the elevation difference and the number of pixels off glaciers are given (by construction the mean elevationdifference is null)

Fig 3Same as Fig 2 but for the Bhutan study site

from minus033plusmn 014 m we yrminus1 to minus022plusmn 012 m we yrminus1The most negative mass balance is measured in Spiti Lahaul(western Himalaya) atminus045plusmn 013 m we yrminus1 (Table 4)

However within a study site mass balances can be highlyvariable from one glacier to another In the Everest area twolarge glaciers (71 km2 each) experienced distinctly differentmass balance withminus016plusmn 016 m we yrminus1 for the south-flowing Ngozumpa Glacier andminus059plusmn 016 m we yrminus1

for the north-flowing Rongbuk Glacier (Fig 4) In Bhutan(Fig 3) mass loss is higher south (minus025plusmn 013 m we yrminus1)

than north (minus014plusmn 012 m we yrminus1) of the main Hi-malayan ridge though not at a statistically significant level

The region-wide mass budget of PKH glaciers isnegative minus101plusmn 55 Gt yrminus1 (minus014plusmn 008 m we yrminus1)which corresponds to a sea level rise contribution of0028plusmn 0015 mm yrminus1 over 1999ndash2011

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1272 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 4Same as Fig 2 but for the Everest study site

Fig 5Same as Fig 2 but for the West Nepal study site

44 Glacier surges

In the Pamir and the Karakoram the spatial pattern of ele-vation changes of many glaciers is typical of surge events orflow instabilities They can be identified because of their veryhigh thinning and thickening rates that can both reach up to16 m yrminus1 (Figs 8ndash10)

Most of them have already been reported to be surge-type glaciers based on their velocity (Kotlyakov et al 2008Quincey et al 2011 Copland et al 2009 Heid and Kaumlaumlb2012) morphology (Copland et al 2011 Barrand and Mur-ray 2006 Hewitt 2007) or elevation changes (Gardelle etal 2012a) Here we only identify surge-type glaciers for thesake of mass balance calculation and we do not attempt toprovide an exhaustive up-to-date surge inventory This is why

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1273

Fig 6 Same as Fig 2 but for the Spiti Lahaul study site The map of elevation changes contains many voids due to the presence ofthin clouds during the acquisition of the SPOT5 DEM The mass balance of the Chhota Shigri and Bara Shigri glaciers are respectivelyminus039plusmn 018 m we yrminus1 andminus048plusmn 018 m we yrminus1

the published surge inventories in the Pamir and the Karako-ram do not necessarily match the surge-type glaciers labeledon the elevation change maps as these refer only to our ob-servation period (1999ndash2011)

The mass balance of surge-type glaciers (respectivelynon-surge-type glaciers) is +019plusmn 022 m we yrminus1

(+012plusmn 011 m we yrminus1) in the Pamir+009plusmn 030 m we yrminus1 (+009plusmn 015 m we yrminus1) inthe western Karakoram and+010plusmn 025 m we yrminus1

(+012plusmn 012 m we yrminus1) in the eastern Karakoram Theoverall similarity between the mass budgets for surge-typeand non-surge-type glaciers shows that those flow instabili-ties seem not to affect the glacier-wide mass balance at leastover short time periods here 10 yr

5 Discussion

51 SRTM penetration

Alternative estimates of SRTMC-bandpenetration for the PKHregion were obtained by Kaumlaumlb et al (2012) by comparing el-evations of the ICESat altimeter (Zwally et al 2002) with theSRTMC-bandDEM in PKH and extrapolating the 2003ndash2008trend of elevation changes back to the acquisition date ofSRTM The difference between east and west is more dis-tinct in our case with a mean penetration in the Karakoram of34 m (24plusmn 03 m in Kaumlaumlb et al 2012) 24 m in Bhutan and14 m around the Everest area (25plusmn 05 m for a wider areaincluding east Nepal and Bhutan in Kaumlaumlb et al 2012) Thiseastwest gradient is expected given that in February 2000when the SRTM mission was flown the snowfirn thicknesswas probably larger in the western PKH where glaciers areof winter-accumulation type than in the eastern PKH where

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1274 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 7Same as Fig 2 but for the Hindu Kush study site

glaciers are of summer-accumulation type Further studieswould be necessary to investigate the possible reasons be-hind the relatively low SRTMC-bandpenetration for the Pamir(18 m) that do not follow the eastwest gradient mentionedabove

Both our and Kaumlaumlb et alrsquos (2012) estimates agree withintheir error bars and seem thus to provide robust first-orderestimates for the SRTMC-bandradar penetration in the regionHowever it remains to be verified if a penetration depth com-puted at the regional scale is appropriate for a single smallglacier whose altitude distribution is different from the oneof whole study site (eg Abramov Glacier in the far north ofour Pamir study site)

52 Thinning over debris-covered ice

PKH glaciers are heavily debris-covered in their ablation ar-eas because of steep rock walls that surround them and in-tense avalanche activity Twelve percent (12 ) of the glacierarea in our study sites are covered with debris (up to 22 inthe Everest area Table 1) which is in agreement with pre-vious estimates for the Himalaya and the Karakoram 13 in Kaumlaumlb et al (2012) and 10 in Bolch et al (2012) Thedebris layer is expected to modify the ablation of the under-

lying ice It will increase ablation if its thickness is just a fewcentimeters (dominance of the albedo decrease) or decreaseablation it if the layer is thick enough to protect the ice fromsolar radiation (Mattson et al 1993 Mihalcea et al 2006Benn et al 2012 Lejeune et al 2013)

However recent studies have reported similar overall thin-ning rates over debris-covered and debris-free ice in thePKH (Kaumlaumlb et al 2012 Gardelle et al 2012a) or evenhigher lowering rates over debris-covered parts in the Ever-est area (Nuimura et al 2012) Our findings are consistentwith Nuimura et al (2012) for the Everest study site with astronger thinning (rate of elevation change ofminus097 m yrminus1)

in debris-covered areas than over clean ice (rate of elevationchange ofminus057 m yrminus1) In the Pamir Karakoram and SpitiLahaul study sites these rates are similar while in the HinduKush West Nepal Bhutan and Hengduan Shan study sitesthinning is greater over clean ice in agreement with the com-monly assumed protective effect of debris (Fig 4) Thosedifferences between our 9 study sites suggest a complex re-lationship between debris cover and glacier wastage

As local glacier thickness changes are a result of bothmass balance processes and a dynamic component theseunexpected high thinning rates over debris-covered parts

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1275

Fig 8Same as Fig 2 but for the Karakoram east study site Surge-type glaciers in an active surge phase (resp quiescent phase) are identifiedwith a solid triangle (resp solid circle)

are potentially the result of glacier-wide processes Over aglacier tongue the ablation can be enhanced by the pres-ence of steep exposed ice cliffs or supraglacial lakes andponds (Sakai et al 2000 2002) In addition it is also pos-sible that the glacier tongues are mostly covered by thindebris layers such that the albedo effect dominates the in-sulating effect resulting in an increased tongue-wide abla-tion The prevalence of thin debris was recently suggestedfor debris-covered glaciers around the Mount Gongga in thesouth-eastern Tibetan Plateau (Zhang et al 2013b) Finallyice-flow rates could be very low at debris-covered parts asshown by Quincey et al (2007) in the Everest area Kaumlaumlb(2005) for Bhutan and Scherler et al (2011b) in the HinduKush-Karakoram-Himalaya Thus all three factors are likelyto increase the overall thinning under debris despite the un-doubted protective effect of a continuous thick debris coverat a local scale Future investigations combining surface ve-locity fields and maps of elevation changes could allow eval-uating more closely the relative role of ablation and ice fluxesin thinning rates over PKH glacier tongues (eg Berthierand Vincent 2012 Nuth et al 2012) Measurements of the

spatial distribution of debris thickness over the PKH glaciertongues are also needed

Note that in the case of debris-covered tongues the ele-vation change measurement represents the glacier thicknesschange but also the possible debris thickness evolution Thelatter remains poorly known in the PKH or in any othermountain range But based on the debris discharges givenby Bishop et al (1995) for Batura Glacier (Pakistan) 48to 90times 103 m3 yrminus1 the thickening of the debris can be es-timated between 27 and 55 mm yrminus1 which is negligiblegiven the magnitude of the glacier elevation changes

53 Comparison to previous mass balance estimates

Throughout the PKH there is only a single peer-reviewedglaciological record (on Chhota Shigri Glacier) coveringmost of our study period (Wagnon et al 2007 Vincent etal 2013) Our geodetic mass balance estimate for the SpitiLahaul study site has been recently compared to those glacio-logical measurements on Chhota Shigri Glacier (Vincent etal 2013) Mass balance records reported in Yao et al (2012)for the Tibetan Plateau and the surroundings mountain ranges

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1276 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 9Same as Fig 8 but for the Karakoram west study site

only start during glaciological year 20052006 This is whyhere we do not compare our mass balance estimates toglaciological records and limit our comparison to others re-gional estimates obtained using the geodetic or gravimetricmethods

We stress that the comparison of our mass balance mea-surements to published estimates is complicated by thefact that generally they do not cover the same time pe-riod Little is known about the inter-annual variability inthe PKH during the 21st century due to the limited num-ber of long glaciological time series The 9 yr mass balance

record on Chhota Shigri Glacier reveals a relatively highinter-annual variability of the annual mass balance (stan-dard deviationplusmn 074 m we yrminus1 during 2003ndash2011 Vin-cent et al 2013) which is similar to the one obtained be-tween 1968 and 1998 on Abramov Glacier (standard devi-ation 069 m we yrminus1 WGMS 2012) Thus inter-annualvariability can explain part of the difference between two cu-mulative mass balance estimates over 5ndash10 yr even if they areoffset by only one or two years

Based on ICESat repeat track measurements and theSRTM DEM Kaumlaumlb et al (2012) measured a region-wide

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1277

Fig 10Same as Fig 8 but for the Pamir study site

mass balance ofminus021plusmn 005 m we yrminus1 between 2003and 2008 for Hindu Kush-Karakoram-Himalaya glaciersFor that same area (ie excluding the Pamir and theHengduan Shan study sites) we found a mass balance ofminus015plusmn 007 m we yrminus1 Thus our result agrees with Kaumlaumlbet al (2012) within the error bars although our study periodis longer (1999ndash2011) and our measurement principle andextrapolation approach are different In addition we havecalculated updated mass balances from Kaumlaumlb et al (2012)ie averages over 3times 3 degree cells centered over our studysites and find that they are consistent within error barswith our mass balance estimates (Table 5) Similar resultsare found when our mass balances are compared to a spa-tially extended ICESat analysis for 2003ndash2009 (Gardner etal 2013) Mass losses measured with ICESat (Kaumlaumlb et al2012 Gardner et al 2013) tend to be slightly larger thanours This could be due to the difference in time periodsAlso our smaller mass losses could be partly explained if wesystematically underestimated the poorly constrained pene-

tration of the C-band andor X-band radar signal into ice andsnow

At a more local scale Bolch et al (2011) reported a massbalance ofminus079plusmn 052 m yrminus1 between 2002 and 2007 byDEM differencing over a glacier area of 62 km2 includingten glaciers south and west of Mt Everest For these sameglaciers Nuimura et al (2012) found a mean mass balanceof minus045plusmn 060 m yrminus1 during 2000ndash2008 while they com-puted a mass balance ofminus040plusmn 025 m yrminus1 between 1992and 2008 over a glacier area of 183 km2 around Mt EverestIn the present study we found an average mass balance ofminus041plusmn 021 m yrminus1 over the ten glaciers mentioned above(Table 5 Fig S1) and a value ofminus026plusmn 013 m yrminus1 overthe whole Everest study site between 2000 and 2011 Ourvalues are comparable to those of Nuimura et al (2012)Our mass balance is less negative than the 2002ndash2007 massbalance of Bolch et al (2011) but not statistically differentwhen error bars are considered (Fig 12) Indeed Bolch etal (2011) acknowledged some high uncertainties for their

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1278 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Table 4Glacier mass budget of the nine sub-regions to which the results of the study sites (mass balances) have been extrapolated For eachof them the total glacierized area as well as the percentage of the glacier area over which the mass balance was computed (ie study sites)are given The total glacier area is derived from the RGI v20 (Arendt et al 2012) The error for the mass balance in each sub-region is theroot of sum of squares (RSS) of the mass balance error of each study site and the regional extrapolation error The error for the mass budgetin each sub-region includes additionally the error from the total glacier area in the RGI v20

Sub-region Total glacier Measured glacier Mass balance Mass budgetarea (km2) area () (m we yrminus1) (Gt yrminus1)

Hengduan Shan 11584 12 minus033plusmn 014 minus38plusmn 38Bhutan 4021 34 minus022plusmn 013 minus09plusmn 05Everest 6226 23 minus026plusmn 014 minus16plusmn 08West Nepal 6849 13 minus032plusmn 014 minus22plusmn 10Spiti Lahaul 9043 23 minus045plusmn 014 minus41plusmn 13Hindu Kush 6135 13 minus012plusmn 016 minus07plusmn 10Karakoram East

1902428 +011 plusmn 014 +19 plusmn 31

Karakoram West 30 +009 plusmn 018Pamir 9369 34 +014 plusmn 014 +13 plusmn 13

TotalArea- 72251 30 minus014plusmn 008 minus101plusmn 55weighted average

Table 5Comparison of geodetic mass balances between this study and previous published results with overlapping study periods and similargeographic locations Updated figures from Kaumlaumlb et al (2012) are averaged over 3 times 3 cells centered over the study sites of the presentstudy

GlacierSite name This study Bolch et al (2011) Nuimura et al (2012) Kaumlaumlb et al (2012 updated)2002ndash2007 2000ndash2008 2003ndash2008

Study Mass balance Area (km2)a Mass balance Area Mass balance Mass balanceperiod (m we yrminus1) (m we yrminus1) (km2) (m we yrminus1) (m we yrminus1)

Hengduan San 1999ndash2010 minus022 plusmn 014 1303 (55 )Bhutan 1999ndash2010 minus022 plusmn 012 1384 (64 ) minus052 plusmn 016b

Everest

1999ndash2011

minus026 plusmn 013 1461 (58 ) minus039 plusmn 011AX010 minus090plusmn 034 04Changri SharNup minus042plusmn 017 161 (79 ) minus029plusmn 052 130 minus055plusmn 038Khumbu minus051plusmn 019 203 (47 ) minus045plusmn 052 170 minus076plusmn 052Nuptse minus037plusmn 020 50 (74 ) minus040plusmn 053 40 minus034plusmn 027Lhotse Nup minus021plusmn 027 24 (70 ) minus103plusmn 051 19 minus022plusmn 047Lhotse minus043plusmn 018 85 (83 ) minus110plusmn 052 65 minus067plusmn 051Lhotse SharImja minus070plusmn 052 98 (44 ) minus145plusmn 052 107 minus093plusmn 060Amphu Laptse minus046plusmn 034 25 (38 ) minus077plusmn 052 15 minus018plusmn 094Chukhung +044 plusmn 024 42 (47 ) +004 plusmn 054 38 +043 plusmn 081Ama Dablam minus049plusmn 017 36 (54 ) minus056plusmn 052 22 minus056plusmn 073Duwo minus016plusmn 026 19 (18 ) minus196plusmn 053 10 minus068plusmn 074Total Khumbu minus041plusmn 021 744 minus079plusmn 052 617 minus045plusmn 060(10 Glaciers above)

West Nepal 1999ndash2011 minus032 plusmn 013 908 (40 ) minus032 plusmn 012Spiti Lahaul 1999ndash2011 minus045 plusmn 013 2110 (46 ) minus038 plusmn 006Karakoram East 1999ndash2010 +011 plusmn 014 5328 (42 ) minus004 plusmn 004Karakoram West 1999ndash2008 +009 plusmn 018 5434 (45 )Hindu Kush 1999ndash2008 minus012 plusmn 016 793 (80 ) minus020 plusmn 006Pamir 1999ndash2011 +014 plusmn 013 3178 (50 )

a The total glacier area is given in km2 and in parenthesis the of the glacier area actually covered with measurementsb For this cell the ICESat coverage is insufficient and does not sample all glacier elevations

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1279

Fig 11 Elevation changes (SPOT5-SRTM) in the ablation area for clean ice (blue circles) and debris-covered ice (black circles) for eachstudy site For the Karakoram (east and west) and the Pamir study sites only non-surge-type glaciers are considered Standard error of theelevation differences are typically less thanplusmn2 m with each 100 m elevation band (not shown for the sake of clarity) Note elevation changesover separate sections of a glacier cannot be treated as mass changes due to the disregard of glacier dynamics

estimate over this short time period The differences in sur-vey periods and in the glacier outlines may also explainpart of these discrepancies In particular the delimitation ofthe accumulation area is notoriously difficult and Bolch etal (2011) did not include some of the steep high altitudeslopes which we included due to their potential avalanchecontribution For the 10 glaciers mentioned above our massbalances would become 012 m we yrminus1 more negative ifwe used the exact same outlines as Bolch et al (2011)Our positive mass balance for the small Chukhung Glacier(+044plusmn 024 m we yrminus1) is in agreement with Nuimura etal (2012) but enigmatic in a context of regional glacier lossand deserves further investigation If the 2002ndash2007 massbalance estimate by Bolch et al (2011) is excluded all otherrecent mass balance estimates suggest no acceleration of themass loss in the Everest area when compared to the long-term(1970ndash2007) mass balance measured by Bolch et al (2011)in this regionminus032plusmn 008 m weyrminus1

In Bhutan our results indicate a stronger mass loss forsouthern glaciers than for northern ones Interestingly in thisarea Kaumlaumlb (2005) measured high speeds (up to 200 m yrminus1)

for glaciers north of the Himalayan main ridge and ratherlow velocities over southern glacier tongues by using repeat

ASTER data This is consistent with the more negative massbalance that we report for the southern glaciers in Bhutanie for the slow flowing presumably downwasting glaciers

Berthier et al (2007) reported a mass balance betweenminus069 andminus085 m we yrminus1 for an ice-covered area of915 km2 around Chhota Shigri Glacier between 1999 (SRTMDEM) and 2004 (SPOT5-HRG DEM) For the same areawe found a mass balance ofminus045plusmn 013 m we yrminus1 over1999ndash2011 The difference in the study period may explainpart of the differences Also the methodology by Berthieret al (2007) did not explicitly take into account the SRTMradar penetration over glaciers and also corrected an appar-ent elevation-dependent bias according to glacier elevationswhich is now known as a resolution issue and is best cor-rected according to the terrain maximum curvature (Gardelleet al 2012b) More details and discussions about these dif-ferences can be found in Vincent et al (2013)

Importantly the results of the eastern Karakoram studysite confirm the balanced or slightly positive mass budgetreported by Gardelle et al (2012a) in the western Karako-ram over the past decade The two Karakoram study sitesare adjacent with a small overlapping area (sim 1100 km2 ofglaciers) where elevation differences agree within 1 m Both

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1280 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 12 Comparison of geodetic mass balances for 10 glaciers inthe Mount Everest area 1999ndash2010 mass balances measured inthe present study are compared with published estimates during2002ndash2007 (Bolch et al 2011) and during 2000ndash2008 (Nuimuraet al 2012) The 1-to-1 line is drawn The mean difference (plusmn1standard deviation) between (i) our values and Bolch et al (2011)are (047plusmn 058 m we yrminus1) and (ii) our values and Nuimura etal (2012) are (012plusmn 021 m we yrminus1) Part of the differences be-tween estimates may be due to the different periods surveyed andthe different glacier outlines used

sites show similar mass balances over slightly different pe-riods +011plusmn 014 m yrminus1 over 1999ndash2010 in the east and+009plusmn 018 m yrminus1 over 1999ndash2008 in the west This con-firms a ldquoKarakoram anomalyrdquo that was first suggested basedon field observations (Hewitt 2005) We report a mass bal-ance of+010plusmn 013 m yrminus1 for Baltoro Glacier (see loca-tion in Fig 8 size= 304 km2) between 1999 and 2010which is consistent with its gradual speed-up reported since2003 (Quincey et al 2009) However given the completelydifferent methods used in our and the latter study we can-not conclude that this demonstrates a real shift from nega-tive to positive mass balance The mass budget of SiachenGlacier (1145 km2 sometimes referred to as the highestbattlefield in the world) is also balanced or slightly pos-itive (+014plusmn 014 m we yrminus1) Thus the alarming state-ment by Rasul (2012) that this glacier retreated by 59 kmand lost 17 of its mass during 1989ndash2009 is very likelyerroneous Again within error bars our regional mass bal-ance in the Karakoram agrees with the mass balance ofminus003plusmn 004 m yrminus1 found by Kaumlaumlb et al (2012) over 2003ndash2008 and with the mass balance ofminus010plusmn 018 m we yrminus1

(2-sigma error level) obtained by Gardner et al (2013) for aregion including both the Karakoram and the Hindu Kush

In the northernmost part of the Pamir in the Alay moun-tain range the mass balance of Abramov Glacier has beenmeasured in the field for 30 yr between 1968 and 1998(WGMS 2012) with a mean value ofminus046 m yrminus1 Herewe report a mass balance ofminus003plusmn 014 m yrminus1 for this

glacier over 1999ndash2011 Our near zero mass budget estimatefor this glacier disagrees with negative mass balances recon-structed for the same period with a model run using biascorrected NCEPNCAR re-analysis data and calibrated withfield data (Barundun et al 2013) An underestimate of ourSRTM C-Band penetration in the Pamir as a whole and inparticular for the small Abramov Glacier may contribute tothese differences

For the whole Pamir study site our mass budget is bal-anced or slightly positive (+014plusmn 013 m we yrminus1) In theeastern Pamir the mass balance of Muztag Ata Glacier was+025 m we yrminus1 between 2005 and 2010 (Yao et al 2012)Although Muztag Ata Glacier is located outside of our Pamirstudy site and his glaciological records only cover the sec-ond half of our study period this measurement is consistentwith our remotely sensed mass balance Thus the ldquoKarako-ram anomalyrdquo should probably be renamed the ldquoPamir-Karakoram anomalyrdquo

This slightly positive or balanced mass budget measuredin the Pamir seems to contradict previous findings by Heidand Kaumlaumlb (2012) They reported rapidly decreasing glacierspeeds in this region between two snapshots of annual ve-locities in 20002001 and 20092010 an observation thatwould fit with negative mass balances (Span and Kuhn 2003Azam et al 2012) We note though that the relationship be-tween glacier mass balance and ice fluxes is not straightfor-ward and might in particular involve a time delay betweena change in mass balance and the related dynamic response(Heid and Kaumlaumlb 2012) Thus the decrease in speed couldwell be due to negative mass balances before or partiallybefore 1999ndash2011 We acknowledge that this discrepancyneeds to be investigated further in particular by examiningcontinuous changes in annual glacier speed through the firstdecade of the 21st Century Indeed Heid and Kaumlaumlb (2012)measured differences between two annual periods which maynot represent mean decadal velocity changes

Jacob et al (2012) using satellite gravimetry (GRACE)measured a mass budget ofminus5plusmn 6 Gt yrminus1 over 2003ndash2010for the ldquoKarakoram-Himalayardquo their region 8c which cor-responds to our study area excluding the Pamir Karakoramand Hindu Kush sub-regions but including some glaciersnorth of the Himalayan ridge already on the Tibetan PlateauFor this subset of our PKH study area we measured amass budget ofminus126plusmn 42 Gt yrminus1 The two estimates over-lap within their error bars especially if our error is dou-bled to match the two standard error level used by Jacob etal (2012) The estimate of Jacob et al (2012) over our com-plete study region (PKH) ieminus6plusmn 8 Gt yrminus1 (sum of theirregions 8b and 8c) also includes mass changes for the KunlunShan sub-region for which we did not measure glacier massbalances Therefore the comparison with our PKH value(minus101plusmn 55 Gt yrminus1) is not straightforward but suggests areasonable agreement especially if we consider the slightmass gain (15plusmn 17 Gt yrminus1) measured with ICESat in theWest Kunlun (Gardner et al 2013)

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1281

Table 6 Annual average of the glacier seasonal and decadal mass loss contributions to Indus Ganges and Brahmaputra discharges Basinarea and seasonal contributions are given by Kaser et al (2010) glacier area in each basin is derived from the RGI v20 (Arendt et al 2012)Numbers in parentheses indicate the percentage of glacierized area within the river basin Seasonal contributions were modeled by Kaser etal (2010) based on the CRU 20 dataset over 1961ndash1990 assuming a balanced glacier mass budget Glacier decadal mass loss contributionsare given as annual average over 1999ndash2011

River Basin area Glacier area Glacier contribution Mean discharge(km2) (km2) (m3 sminus1) (m3 sminus1)

Seasonally delayed Decadal mass lossKaser et al (2010) (this study)

Indus 1 139 814 25 598 (2 ) 105 103plusmn 72 2146 (Chevallier personalcommunication 2012)

Ganges 1 023 609 11 168 (1 ) 47 103plusmn 49 11739 Papa et al (2012Brahmaputra 527 666 15 296 (3 ) 33 147plusmn 64 19710 updated)

54 Reasons for the heterogeneous pattern of massbalance

The PKH glacier mass balance is slightly negative(minus014plusmn 008 m we yrminus1) but changes are not homoge-neous from one sub-region to another and seem to coincidewith the climatic settings of the mountain range For the east-ern sites (from the Hengduan Shan to West Nepal) which areunder the influence of the Indian and south-east Asian sum-mer monsoons the mass balances are moderately negative(sim minus026 m we yrminus1) In the northwest of the PKH at thePamir and Karakoram sites where glacier climate is char-acterized by the westerlies in winter mass balances are bal-anced or slightly positive (sim +010 m we yrminus1) In betweenthese two influences the glaciers of the Spiti Lahaul sitein a monsoon-arid transition zone experienced the strongestmass loss (minus045plusmn 013 m we yrminus1)

This heterogeneous pattern of mass balances seems to berelated to trends in precipitation throughout the PKH Archerand Fowler (2004) reported an increase in winter precipi-tation over the Upper Indus Basin since 1961 a trend con-firmed by Yao et al (2012) over the Pamir and the Karako-ram based on the analysis of the Global Precipitation Cli-matology Project data set (GPCP Adler et al 2003) Thisprecipitation increase is a source of greater accumulation forglaciers and thus a possible explanation for their slightly pos-itive mass budgets In addition in the Upper Indus Basin themean summer temperature has been decreasing since 1961(Fowler and Archer 2006) which is consistent with the find-ings of Shekhar et al (2010) in the Karakoram who reporteda decrease in both maximum and minimum temperature since1988 These increases in winter precipitation and summertemperature likely translate in greater accumulation and re-duced ablation changes which are consistent with the bal-anced or slightly positive glacier mass budget

On the other hand in the east of the PKH the Indiansummer monsoon has been weakening since the 1950s (Bol-lasina et al 2011) which could have reduced the amount

of snowfall over the eastern study sites and thus resultedin negative mass balances In addition maximum and min-imum temperatures increased since 1988 in the western Hi-malaya (Shekhar et al 2010) and maximum temperaturesincreased in the central Himalaya (Nepal) since the mid-1970s (Shrestha et al 1999) Thus both precipitation andtemperature trends in the Himalaya likely contributed to thenegative mass balances measured over the correspondingstudy sites (from Spiti Lahaul to Hengduan Shan Fig 1)

However gridded data sets such as GPCP are not neces-sarily representative of the climate at high altitude IndeedHewitt (2005) reported a 5-to-10-fold increase in precipi-tation between the glacier front and the accumulation areain the Karakoram that is not captured by reanalysis dataas recently confirmed by Immerzeel et al (2012) Thus di-rect measurements of accumulation on High Mountain Asiaglaciers (eg Azam et al 2013) and high-altitude weatherstations are eagerly needed to validate the large scale grid-ded data set (eg reanalysis) test their ability to describe thespecific climate near to PKH glaciers and assess the relativerole played by temperature and precipitation changes

55 Contribution to water resources

Our glacier mass balance can be averaged over large riverbasins to estimate the mean contribution to river runoffdue to a decade of glacier mass loss We assume therebyonly direct river runoff without loss terms For the threerivers for which we cover the entire glacierized area of theircatchments within our sub-regions (Fig 1) these contribu-tions to the river discharge are equal to 103 m3 sminus1 (Indus)103 m3 sminus1 (Ganges) and 147 m3 sminus1 (Brahmaputra) for theperiod 1999ndash2011 (Table 6)

Rivers of PKH being key water resources measures oftheir discharges are highly sensitive and difficult to access forpolitical reasons Papa et al (2012) built recent time series ofdischarges for Ganges and Brahmaputra by using altimetrymeasurements The locations of the corresponding dischargeestimates in Bangladesh are given in Fig 1 We can therefore

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1282 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

evaluate the relative contribution of glacier decadal mass loss(Table 6) to annual river run-off at 09 for the Gangesat Hardinge (basin area ofsim 1 020 000 km2) and 07 forBrahmaputra at Bahadurabad (basin area ofsim 530 000 km2)between 1999 and 2011 Given the discharge measured at theGuddu station by the Surface Water Hydrology Project of theWater and Power Development Authority (Pakistan see lo-cation on Fig 1) between 1999 and 2003 we can estimatethe contribution of glacier mass loss at 48 for the IndusRiver at this station

The seasonal contributions of glaciers to river run-off (iethe part of the annual precipitation that experiences season-ally delayed release from the glaciers) directly results fromseasonal variations of meteorological quantities in contrastto the glacier mass loss contribution which results from thelong-term climatic change Both runoff components directlyaffect the amount of water available in a river at a given lo-cation and at a given time so that their comparison could beof interest Even if the seasonal timing of the glacier massloss contribution is unclear it could in general peak at asimilar time of the year than the seasonal glacier contribu-tion so that both components rather enlarge each other thancancel Seasonal glacier contributions have been modeled byKaser et al (2010) The latter study assumed glaciers to bein equilibrium with the present climate and thus prescribedglacier mass balances equal to 0 Thus their seasonally de-layed glacier contributions do not include the part due todecadal glacier mass loss and is complementary to our es-timates (Table 6) For the eastern basins under the influenceof the Indian summer monsoon (Ganges and Brahmaputra)the contribution of glacier mass loss is higher than the sea-sonal contribution In other words for the first decade of the21st century the mass loss of glaciers is on average moreimportant for water availability in those two rivers than theirseasonal water storage effect The situation is different for theIndus Basin where the glacier melt season is also the dry sea-son which results in a relatively higher seasonally delayedglacier contribution to the river discharge There the season-ally delayed water release from the glaciers and the decadalglacier mass loss contribute on average equally to the riverdischarge Both contributions are strongly (and equally) de-pendent on the location of the discharge measurement andwill be significantly larger (in relative term) in more heav-ily glacierized and smaller mountain catchments located inthe upper part of these major river basins (Bookhagen andBurbank 2010)

6 Conclusion

In this study we assessed the spatial pattern of glaciermass balance over the PKH by comparing the SRTM andSPOT5 DEMs over 9 well-distributed study sites between1999 and 2011 We found slightly positive or balanced massbudgets in the western part (Pamir Karakoram) moder-ate mass loss in the east (Nepal Bhutan Hengduan Shan)

and the most negative mass balance in the monsoon-aridtransition zone (Spiti Lahaul in the western Himalaya)By extrapolating these values to climatically homogeneoussub-regions we estimate the PKH overall mass balanceto have beenminus014plusmn 008 m we yrminus1 between 1999 and2011 which corresponds to a contribution to sea level riseof 0028plusmn 0015 mm yrminus1 This is about 4 of the globalglacier mass loss estimated by Gardner et al (2013) thoughover a slightly shorter time period (2003ndash2009) The largestsource of uncertainties in those geodetic mass balance esti-mates is the poorly constrained penetration of the C-BandSRTM radar signal into snow and ice

Our technique of DEM differencing allows mapping in de-tail the spatial pattern of glacier elevation changes Thosemaps reveal many surge-type behaviors both in the Pamirand the Karakoram It also appears that the glacier-wide massbalance does not seem to be considerably affected by themass transfer from surges over thesim 10 yr time period stud-ied Similar lowering rates over debris-covered and clean iceare observed for four study sites despite the commonly as-sumed insulating effect of debris covers This suggests forthe scale of entire glacier tongues a spatial mixture of re-duced ablation under intact thick debris covers and enhancedablation by supraglacial lakes or ice cliffs or due to thin de-bris layer and very low glacier velocities in ablation areasthat reduce the ice advection downstream

The regionally heterogeneous pattern of ice wastage is inbroad agreement with contrasted trends in large-scale pre-cipitation and temperature However glaciological (eg ac-cumulation) and meteorological records are needed at highaltitude in the vicinity of glaciers to evaluate more closelythe local influence of atmospheric variables on glacier massbalance and fully understand the reasons behind those con-trasted mass budgets in the PKH

Supplementary material related to this article isavailable online at httpwwwthe-cryospherenet712632013tc-7-1263-2013-supplementpdf

Acknowledgements We thank G Cogley A Gardner T NuimuraT Bolch P Wagnon M Barandun M Hoelzle M Huss and ananonymous referee for their constructive comments that led to aconsiderably improved manuscript We thank the Water and PowerDevelopment Authority (WAPDA) for their hydrological data(processed by P Chevallier) and F Papa for sharing his updatedrun-off data ahead of publication We thank T Bolch for sharinghis glacier outlines from the Everest area We thank the USGSfor allowing free access to their Landsat archive CGIAR-SCI forSRTM C-band data DLR for SRTM X-band data and GLIMS forglacier inventories J Gardelle acknowledges a PhD fellowshipfrom the French Space Agency (CNES) and the French NationalResearch Center (CNRS) E Berthier acknowledges support fromCNES through the TOSCA and ISIS proposal no 397 and fromthe Programme National de Teacuteleacutedeacutetection Spatiale E Berthieracknowledges M Masiokas and R Villalba for welcoming him

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1283

at the IANIGLA (Mendoza Argentina) Y Arnaud acknowledgessupport from French National Research Agency through ANR-09-CEP-005-01PAPRIKA A Kaumlaumlb acknowledges support fromESA through Glaciers_cci (400010177810I-AM) and from theEuropean Research Council under the European Unionrsquos SeventhFramework Programme (FP2007-2013)ERC Grant Agreementn 320816

Edited by I M Howat

The publication of this articleis financed by CNRS-INSU

References

Adler R Huffman G Chang A Ferraro R Xie P JanowiakJ Rudolf B Schneider U Curtis S Bolvin D Gruber ASusskind J Arkin P and Nelkin E The Version-2 GlobalPrecipitation Climatology Project (GPCP) Monthly Precipita-tion Analysis (1979ndashPresent) J Hydrometeorol 4 1147ndash11672003

Archer D R and Fowler H J Spatial and temporal variationsin precipitation in the Upper Indus Basin global teleconnectionsand hydrological implications Hydrol Earth Syst Sci 8 47ndash61doi105194hess-8-47-2004 2004

Arendt A Bolch T Cogley J G Gardner A Hagen J-OHock R Kaser G Pfeffer W T Moholdt G Paul F RadicV Andreassen L Bajracharya S Beedle M Berthier EBhambri R Bliss A Brown I Burgess E Burgess DCawkwell F Chinn T Copland L Davies B de AngelisH Dolgova E Filbert K Forester R Fountain A Frey HGiffen B Glasser N Gurney S Hagg W Hall D Hari-tashya U K Hartmann G Helm C Herreid S Howat IKapustin G Khromova T Kienholz C Koenig M KohlerJ Kriegel D Kutuzov S Lavrentiev I LeBris R Lund JManley W Mayer C Miles E Li X Menounos B Mer-cer A Moelg N Mool P Nosenko G Negrete A NuthC Pettersson R Racoviteanu A Ranzi R Rastner P RauF Rich J Rott H Schneider C Seliverstov Y Sharp MSigurethsson O Stokes C Wheate R Winsvold S WolkenG Wyatt F and Zheltyhina N Randolph Glacier Inventory[v20] A Dataset of Global Glacier Outlines Global Land IceMeasurements from Space Boulder Colorado USA Digital Me-dia httpwwwglimsorgRGIrandolphhtml 2012

Azam M Wagnon P Ramanathan A Vincent C Sharma PArnaud Y Linda A Pottakkal J Chevallier P Singh V andBerthier E From balance to imbalance a shift in the dynamicbehaviour of Chhota Shigri glacier western Himalaya India JGlaciol 58 315ndash324 doi1031892012JoG11J123 2012

Azam M F Wagnon P Vincent C Ramanathan A Linda Aand Singh V B Reconstruction of the annual mass balanceof Chhota Shigri Glacier (Western Himalaya India) since 1969Ann Glaciol in revision 2013

Barrand N and Murray T Multivariate Controls on the In-cidence of Glacier Surging in the Karakoram Himalaya

Arct Antarct Alp Res 38 489ndash498 doi1016571523-0430(2006)38[489MCOTIO]20CO2 2006

Barundun M Huss M Azisov E Gafurov A HoelzleM Merkushkin A Salzmann N and Usubaliev R Re-establishing seasonal mass balance observation at AbramovGlacier Kyrgyzstan from 1968ndash2012 Abstract EGU2013-5668European Geophysical Union Vienna 2013

Benn D Bolch T Hands K Gulley J Luckman ANicholson L Quincey D Thompson S Toumi R andWiseman S Response of debris-covered glaciers in theMount Everest region to recent warming and implicationsfor outburst flood hazards Earth-Sci Rev 114 156ndash174doi101016jearscirev201203008 2012

Berthier E and Vincent C Relative contribution of surface mass-balance and ice-flux changes to the accelerated thinning of Merde Glace French Alps over 1979ndash2008 J Glaciol 58 501ndash512doi1031892012JoG11J083 2012

Berthier E Arnaud Y Baratoux D Vincent C and Reacutemy FRecent rapid thinning of the ldquo Mer de Glace rdquo glacier derivedfrom satellite optical images Geophys Res Lett 31 L17401doi1010292004GL020706 2004

Berthier E Arnaud Y Kumar R Ahmad S Wagnon P andChevallier P Remote sensing estimates of glacier mass balancesin the Himachal Pradesh (Western Himalaya India) RemoteSens Environ 108 327ndash338 doi101016jrse2006110172007

Bhambri R Bolch T Kawishwar P Dobhal D P SrivastavaD and Pratap B Heterogeneity in Glacier response from 1973to 2011 in the Shyok valley Karakoram India The CryosphereDiscuss 6 3049ndash3078 doi105194tcd-6-3049-2012 2012

Bhutiyani M Mass-balance studies on Siachen Glacier in theNubra valley Karakoram Himalaya India J Glaciol 45 112ndash118 1999

Bishop M P Schroder J F and Ward J L SPOT multispec-tral analysis for producing supraglacial debris-load estimates forBatura glacier Pakistan Geocarto Int 10 81ndash90 1995

Bolch T Pieczonka T and Benn D I Multi-decadal mass lossof glaciers in the Everest area (Nepal Himalaya) derived fromstereo imagery The Cryosphere 5 349ndash358 doi105194tc-5-349-2011 2011

Bolch T Kulkarni A Kaumlaumlb A Huggel C Paul F Cogley JFrey H Kargel J Fujita K Scheel M Bajracharya S andStoffel M The state and fate of Himalayan Glaciers Science336 310ndash314 doi101126science1215828 2012

Bollasina M Ming Y and Ramaswamy V Anthropogenicaerosols and the weakening of the South Asian summer mon-soon Science 334 502ndash505 doi101126science12049942011

Bookhagen B and Burbank D Toward a complete Himalayan hy-drological budget spatiotemporal distribution of snowmelt andrainfall and their impact on river discharge J Geophys Res115 F03019 doi1010292009JF001426 2010

Bretherton C Widmann M Dymnikov V Wallace J and BladeacuteI The effective number of spatial degrees of freedom of a time-varying field J Climate 12 1990ndash2009 1999

Copland L Pope S Bishop M Shroder J Clendon PBush A Kamp U Seong Y and Owen L Glacier veloc-ities across the central Karakoram Ann Glaciol 50 41ndash49doi103189172756409789624229 2009

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

1264 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

different methods of mass balance measurements have beenused in the PKH

i Glaciological (or traditional) measurements are rela-tively short-termed in general less than 10 yr mainlybecause of difficult access to generally remote glaciersIn addition measurement series are often biased to-wards small-to-medium-sized and debris-free glaciersfor logistical reasons (eg Dobhal et al 2008 Fujitaand Nuimura 2011 Azam et al 2012 Bolch et al2012 Yao et al 2012) Vincent et al (2013) discussedin detail the inadequate spatial and temporal samplingof those glaciological measurements in the western Hi-malaya Gardner et al (2013) have shown that during2003ndash2009 the extrapolation of those few and short-term glaciological mass balances to entire High Moun-tain Asia lead to an overestimation of the mass loss byabout a factor of 3

ii The accumulation area ratio (AAR) method relies heav-ily on the availability of glaciological mass balancesduring various years in order to establish an empiri-cal relationship between annual mass balance and AAR(eg Rabatel et al 2005) The extrapolation to un-measured glaciers is not straightforward This methodhas been applied by Kulkarni et al (2011) to 19 glaciersin north-west India

iii The hydrological method has only been used once inthe PKH to our knowledge providing 5 yr of mass bal-ance for Siachen Glacier Karakoram (Bhutiyani 1999)and is very difficult to implement due to a lack of accu-rate precipitation and runoff measurements at high alti-tude over the PKH (Bolch et al 2012 Immerzeel et al2012)

iv The geodetic method based on the comparison of topo-graphic data (DEM or laser altimetry) can be applied atregional scales in remote areas using satellite data suchas from ICESat SPOT5 or SRTM This method enablesan increase in the number and diversity of glacier mea-surements (eg in terms of glacier size elevation as-pect percentage of debris cover) (Berthier et al 2007Bolch et al 2011 Kaumlaumlb et al 2012 Nuimura et al2012 Pieczonka et al 2013) An important remaininglimitation is the difficulty to convert volume changes tomass changes (eg Huss 2013)

v The gravimetric method uses the data from the GravityRecovery and Climate Experiment (GRACE) missionlaunched in 2002 Gravity fields need to be corrected forthe influence of the glacial isostatic adjustment (GIA)and the hydrological changes off-glacier (eg Zhang etal 2013a) before the glacier mass change can be ex-tracted This method initially led to contrasted glaciermass budgets in High Mountain Asia (Matsuo and Heki

2010 Jacob et al 2012) The two most recent GRACEmass budgets for High Mountain Asia agree within their2σ error bars (minus14plusmn 17 Gt yrminus1 andminus23plusmn 24 Gt yrminus1

in Gardner et al 2013)

The ICESat- and GRACE-based methods are appropriateto assess the regional mass budget for large glacierized re-gions such as the PKH but also have their specific limita-tions The ICESat sampling is restricted to satellite orbitsthat are separated by tens of kilometers at PKH latitudes andthus does not allow estimating the mass balance of individ-ual glaciers Glacier changes can only be investigated overthe lifetime of ICESat 2003ndash2009 The gravimetric methodhas a coarse spatial resolution (typicallysim 400 km) so thatindividual glaciers and small hydrological basins cannot beresolved

The aim of our study is to present a homogeneous surveyof regional glacier mass balances along the PKH between1999 and 2011 using the geodetic method This is achievedby differencing two digital elevation models (DEMs) over9 study sites selected to be representative of the PKH cli-matic and glaciological diversity This allows for an estima-tion of the spatial pattern of glacier elevation changes alongthe 3000 km-long group of mountain ranges as well as theinfluence of debris cover on thinning rates Finally we ex-trapolate these measurements to provide a new estimate ofthe mass budget of PKH glaciers and their contribution toregional hydrology and sea level rise during the first decadeof the 21st century

Compared to Kaumlaumlb et al (2012) the length of the studyperiod has been doubled and the study area has been signif-icantly extended towards the east (Hengduan Shan China)and the west (Pamir Tajikistan) We also provide a nearly ex-haustive coverage of glacier elevation changes for our 9 studysites as opposed to the sampling by ICESat along satellitetracks only However contrary to the ICESat studies (Kaumlaumlbet al 2012 Gardner et al 2013) our method is limited bythe availability of SPOT5 DEMs on selected sites and doesnot provide insight into the seasonal and annual evolution ofglacier mass balance

2 Study area

Mass balances are investigated for 9 study sites spread alongthe PKH mountain ranges to capture the climatic and glacio-logical variability of the region The location of each studysite is displayed in Fig 1 as well as the extent of the cor-responding sub-regions used to extrapolate the results tothe whole PKH range (see Sect 34) PKH glacier meltwa-ter contributes to five major rivers of Asia (BrahmaputraGanges Indus Tarim and Amu Darya) whose catchment ar-eas are also presented in Fig 1

Each study site corresponds to the extent of a SPOT5 DEM(see Sect 31) A complete coverage of all glaciers in the

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1265

Fig 1 Overview of the extent and location of the nine study sites (black polygons) and corresponding sub-regions (red polygons) along thePKH range Brownish-filled background polygons represent the major river basins of the area Triangles indicate the location of dischargemeasurements discussed in Sect 55

PKH range could not be achieved because of funding restric-tions (SPOT5 is a commercial satellite) the busy acquisitionschedule of this satellite and cloud coverage A further con-strain is that the optical stereo images used to compute theDEMs should be acquired as close as possible to the end ofthe glacier ablation season In practice a time window of2ndash3 months is allowed for the acquisitions which is shortgiven the 26-day orbital cycle of SPOT5 Thus a maximumof 3 to 4 acquisition attempts are possible each year and onlyif the satellite is not programmed for higher priority targets(eg crisis situations) Despite these difficulties our SPOT5DEMs sampled in total 22 200 km2 of glaciers about 30 of the total glacierized area in the PKH (Arendt et al 2012)

The eastern-most sites (Hengduan Shan Bhutan Everestand West Nepal) are strongly influenced by the Indian andsouth-east Asian summer monsoons Ablation and accumula-tion of monsoon-type glaciers occur during the summer sea-son (eg Fujita 2008) Towards the other end of the PKHin the north-west (Pamir Hindu Kush and Karakoram) theclimate is dominated by the westerlies and glaciers are ofwinter-accumulation type The Spiti Lahaul site lies in a tran-sition zone influenced both by the monsoon and the wester-lies The topography also plays a strong role in the moisturetransfer as it dries out the southerly air flow and can preventthe air mass from travelling further north which results ina northward decrease of precipitation (Bookhagen and Bur-bank 2010)

Monsoon-type glaciers are expected to be more sensitivethan winter-accumulation type glaciers to a change in therainsnow limit driven by an increase in temperature (Fujita2008) Indeed a summer warming would increase the alti-

tude of the rainsnow limit and reduce snow accumulation onglaciers as well as decrease their surface albedo (less freshsnow with a high albedo) and thus enhance ablation

The Karakoram and the Pamir are known to host numer-ous surge-type glaciers (Barrand and Murray 2006 Hewitt2007 Kotlyakov et al 2008 Gardelle et al 2012a) char-acterized by the alternation between short active phases in-volving rapid mass transfer from high to low elevationsand longer quiescent phases of low mass fluxes Coplandet al (2011) reported a doubling in the number of glaciersurges after 1990 in the Karakoram with a total of 90 surge-type glaciers observed in the region since the late 1960sIn the Pamir an inventory from 1991 revealed 215 glacierswith signs of repeated surging (looped moraines fast ad-vance of the front) and 51 additional actively surging glaciers(Kotlyakov et al 2008)

Many PKH glaciers are heavily debris-covered in their ab-lation areas because of steep rock walls that surround themand intense avalanche activity (eg Bolch et al 2012) Thesize of these debris zones is highly variable and debris thick-ness can range from a few centimeters (dust or sand) to sev-eral meters (Mihalcea et al 2008 Zhang et al 2013b) Themean debris cover is about 10 of the total glacier area inthe Karakoram and Himalaya (Bolch et al 2012) and culmi-nate at 36 in the Central Himalaya (Scherler et al 2011a)

In the eastern PKH glacier termini are often connected topro-glacial lakes storing meltwater behind frontal morainesor dead-ice dams whereas in the western PKH pro-glaciallakes are less numerous and glacier ablation areas are oftenscattered with supra-glacial lakes (Gardelle et al 2011) re-sulting from successive coalescence of small melting ponds

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1266 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Table 1 Characteristics of the nine study sites Numbers between parentheses indicate the percentage difference between our user-definedglacier inventory and the RGI v20 (Arendt et al 2012)

Site name Glacier area Debris ( of Glacier area Arendt Date Landsat imagesthis study (km2) glacier area) et al (2012) (km2) (SPOT5 DEM)

PathRow Date

Hengduan Shan 1303 11 2541 (+88 ) 24 Nov 2011 p135r39 23 Sep 1999Bhutan 1384 11 1429 (+3 ) 20 Dec 2010 p138r40 19 Dec 2000

p138r41Everest 1461 22 1400 (minus4 ) 4 Jan 2011 p140r41 30 Oct 2000West Nepal 908 17 802 (minus12 ) 3 Jan 2011 p143r39 1 Aug 2009

p143r40Spiti Lahaul 2110 13 2257 (+7 ) 20 Oct 2011 p147r37 15 Oct 2000

p147r38Hindu Kush 793 7 724 (minus11 ) 17ndash21 Oct 2008 p150r34 15 Aug amp

p150r35 16 Sep 1999Karakoram East 5328 10 5384 (+1 ) 31 Oct 2010 p148r35 7 Sep 1998Karakoram West 5434 12 5687 (+1 ) 3 Dec 2008 p149r35 29 Aug 1998Pamir 3178 11 3210 (minus1 ) 29 Nov 2011 p152r33 16 Sep 2000

3 Data and methods

31 Digital elevation models

For each study site we used the Shuttle Radar TopographicMission (SRTM) version 4 DEM as the reference topogra-phy (Rabus et al 2003) This data set acquired in Febru-ary 2000 comes along with a mask to identify the pixelswhich have been interpolated (both were downloaded fromhttpsrtmcsicgiarorg) On each site the SRTM DEM issubtracted from a more recent DEM built from a stereo pairacquired by the HRS sensor onboard the SPOT5 satellite(Korona et al 2009) except for the Hindu Kush study sitewhere the recent DEM has been created from a stereo pairacquired by the HRG sensor also onboard SPOT5 (Berthieret al 2007) The date of each SPOT5 DEM acquisition dif-fers depending on the study site (Table 1) but 7 out of 9 wereacquired between October 2010 and November 2011 EachSPOT5-HRS DEM is also provided with a correlation maskand an ortho-image generated from the rear HRS image

SPOT5-HRS DEMs were produced by the French Map-ping Agency (IGN) using two sets of correlation parametersdefined during the SPIRIT (SPOT 5 stereoscopic survey ofPolar Ice Reference Images and Topographies) internationalpolar year project (Korona et al 2009) Thus two versionsof the DEM are delivered with their respective mask one op-timized (v1) for a gentle topography and the second one (v2)for a more rugged relief Earlier work has shown that v2 hasa slightly larger percentage of interpolated pixels but for thenon-interpolated pixels smaller errors (Korona et al 2009)This version (v2) is preferred here Given that both DEMsare calculated from the same image pair little difference isexpected between v1 and v2 However locally we identifiedsome large bull eye elevation differences between the two

versions of the DEM both off and on glaciers Comparisonwith the SRTM DEM shows that both versions of the DEMare erroneous for those areas and thus we preferred to ex-clude from further analysis all pixels for which the absoluteelevation difference between the two versions of the DEMis larger than 5 m About 20 of the valid DEM pixels areexcluded through that procedure

The SRTM DEM and mask originally at a 3 arcsecondresolution (sim 90 m) are resampled bilinearly to 40 m (UTMprojection WGS84 ellipsoid) to match the posting and pro-jection of the SPOT5 DEM Altitudes are defined above theEGM96 geoid for both DEMs

32 Glacier mask

Where available glacier outlines were downloaded fromthe GLIMS database (Raup et al 2007 Supplement) andmanually corrected if needed Otherwise the glacier maskwas derived from Landsat-TM and Landsat-ETM+ imagesPathrow and dates of Landsat acquisition are given in Ta-ble 1 for each study site A threshold varying between 05and 07 depending on the study site was applied to the Nor-malized Difference Snow Index (TM2-TM5)(TM2+TM5)to detect clean ice and snow automatically whereas debris-covered parts were digitized manually by visual interpreta-tion (eg Racoviteanu et al 2009) The total glacier areaand debris-covered part can be found for each study site inTable 1

Most inventories are derived from Landsat images withminimal snow cover from around the year 2000 prior to the2003 failure of the Scan Line Corrector of the ETM+ sensoronboard Landsat 7 that resulted in image striping Glacierareas are expected to have experienced minor changes sincethen except for glacier fronts that retreated due to the rapid

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1267

Table 2 Elevation ranges equilibrium line altitudes (ELA) and accumulation area ratios (AAR) for the nine study sites The dates ofacquisition of Landsat images used for ELA determination are also given

Sub-region Glaciers elevation ELA (m) AAR Landsat imagesrange (m)

Min Max This study Scherler et Kaumlaumlb et This study Kaumlaumlb et PathRow Dateal (2011a) al (2012) al (2012)

Hengduan Shan 2630 6860 4970plusmn 320 NA NA 55 NA p135r39 23 Sep 1999Bhutan 4390 7480 5690plusmn 440

5700 555036

47p138r40 17 Nov 2000

Everest 4260 8380 5840plusmn 320 31 p140r41 31 oct 2009West Nepal 3950 6870 5590plusmn 138 37 p143r39 1 Aug 2009Spiti Lahaul 3780 6360 5390plusmn 140 5103 5500 34 35 p147r37 2 Aug 2002

p147r38Hindu Kush 2717 6385 5050plusmn 160 5138 30 p150r34 2 Sep 2000

p150r35Karakoram east 3230 7770

5030plusmn 280 4845 554076

47p148r35 7 Sep 1998

Karakoram west 2910 7920 66 p149r35 29 Aug 1998Pamir 2800 7090 4580plusmn 250 NA NA 52 NA p152r33 30 Jul 2000

growth of connected lakes and for glaciers that surged andthus advanced For the few glaciers that advanced their frontposition was updated manually based on the SPOT5 ortho-image

For each study site we favored our user-defined glacierinventories instead of the global Randolph Glacier Inventory(RGI v20 Arendt et al 2012) because the latter is hetero-geneous and for some study sites did not match our accu-racy requirements for the adjustments and corrections of theDEMs

We also estimated the altitude of the equilibrium line(ELA) as the snowline on thesim 2000 Landsat data with min-imal snow cover to separate ablation and accumulation ar-eas The ELA is assumed to be identical to the altitude ofthe snowline at the end of the ablation season (eg Raba-tel et al 2005) after the end of the monsoons or before thefirst snowfalls Therefore we manually digitized snowlinesfor more than 30 glaciers for each study site and computedtheir mean elevations Our ELAs are slightly different fromthose given by Scherler et al (2011a) and Kaumlaumlb et al (2012)probably due to the sensitivity of the ELA to the choice ofthe image (Table 2) Ideally the ELA should have been es-timated from images acquired at the end of the ablation sea-son during various years covering our whole study periodThis would be a highly time-consuming task that was not ac-complished here However we stress that in our study theELA is only used to compare elevation changes on the abla-tionaccumulation areas (Sect 41) and to estimate the sen-sitivity of our mass balances to the choice of the volume-to-mass conversion factor (Sect 35) Our preferred mass bal-ance estimate uses a unique volume-to-mass conversion fac-tor for the whole glacier and is thus independent from theELA

33 Adjustments and corrections of DEM biases

Different DEM processing steps are followed to extract un-biased glacier elevation changes for each study site (eg Nuthand Kaumlaumlb 2011 Gardelle et al 2012b)

331 Planimetric adjustment of the DEMs

A horizontal shift between two DEMs results in an aspect-dependent bias of elevation differences (Nuth and Kaumlaumlb2011) Here the horizontal shift is determined by minimizingthe root mean square error of elevation differences (SPOT5-SRTM) off glaciers where the terrain is assumed to be stableover the study period (Berthier et al 2007) The SRTM DEMis then resampled (cubic convolution) according to the shift

332 Alongacross track corrections (drift of thesatellite orbit)

For DEMs derived from stereo imagery the satellite acquisi-tion geometry can induce a bias along andor across the trackdirection (Berthier et al 2007) We estimate the azimuth ofeach SPOT5 ground track and use it to rotate the coordinatesystem accordingly (Nuth and Kaumlaumlb 2011) Then we com-pute the elevation differences along and across the satellitetrack on stable areas and when necessary correct the bias us-ing a 5th order polynomial fit

333 Curvature correction

As suggested by Paul (2008) and Gardelle et al (2012b) thedifference in original spatial resolutions of the two DEMs canlead to biases related to altitude in mountainous areas Whenthe curvature the first derivative of the slope is high (sharppeaks or ridges) the altitude tends to be underestimatedby the coarse DEM unable to reproduce high-frequency

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1268 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

slope variations Thus this apparent ldquoelevation biasrdquo is cor-rected using the relation between elevation differences andthe maximum terrain curvature estimated over stable areasoff glaciers (Gardelle et al 2012b) Note that the units inFig 1b in Gardelle et al (2012b) should be 10minus2 mminus1 in-stead of mminus1 (A Gardner personal communication 2012)However this does not impact the validity of the correction

334 Radar penetration correction

The SRTM DEM acquired in C-band potentially underesti-mates glacier elevations since at this wavelength (sim 56 cm)the penetration of the radar signal into snow and ice canreach several meters (eg Rignot et al 2001) Here we esti-mate this penetration for each study site by differencing theSRTM C-band (57 GHz) and X-band (97 GHz) DEMs Thelatter was acquired simultaneously with the SRTM X-bandat higher resolution (30 m) but with a narrower swath andthus has a coverage restricted to selected swaths only Sincethe penetration of the X-band is expected to be low com-pared to the C-band (an hypothesis that still needs to be con-firmed Ulaby et al 1986) we consider the elevation differ-ence (SRTMC-bandminus SRTMX-band) to be a first approxima-tion of C-band radar penetration over glaciers (Gardelle etal 2012b) The correction is calculated as a function of alti-tude and applied to each pixel separately

For the Hindu Kush Spiti Lahaul and West Nepal sitesdue to the lack of SRTMX-band data we were unable to de-termine the value of the penetration Thus we applied thecorrection computed over the nearest study site ie Karako-ram for Spiti Lahaul and Hindu Kush and Everest for WestNepal The mean values of the SRTMC-bandpenetrations overglaciers are given for each study site in Table 3

335 Seasonality correction

SPOT5 DEMs were acquired between late October and earlyJanuary depending on the study site Since the SRTM DEMis from mid-February we need to account for possible masschanges during this 1-to-5-month period in order to esti-mate glacier mass balance over an integer number of years(Gardelle et al 2012a) For the eastern sites (from WestNepal to Hengduan Shan) where glaciers are of summer-accumulation types (Fujita 2008) the correction was set to0 This choice is confirmed by recent field observations thatshow no winter accumulation on Mera and Pokalde glaciersin Nepal (Wagnon et al 2013) For the Karakoram studysites we used the winter accumulation rate+013 m we permonth measured on Biafo Glacier (Wake 1989) For theother western study sites (the Pamir Hindu Kush and SpitiLahaul sites) the correction is derived from the mean of win-ter mass balances of 35 glaciers all in the Northern Hemi-sphere+089 m we yrminus1 over 2000ndash2005 corresponding to+015 m we per winter month (Ohmura 2011) The lattervalue is slightly lower than the average winter mass balance

Table 3 Average SRTM C-band penetration estimates (in m)in February 2000 over glaciers in this study and from Kaumlaumlb etal (2012)

Sub-region This study Kaumlaumlb et al (2012)

Hengduan Shan 17 NABhutan 24

25plusmn 05Everest 14West Nepal NA

15plusmn 04Spiti Lahaul NAHindu Kush NA 24plusmn 04Karakoram East

34 24plusmn 03Karakoram WestPamir 18 NA

(+135plusmn 031 m we yrminus1) measured on Abramov Glacier inthe Pamir during 1968ndash1998 (WGMS 2012) and in reason-ably good agreement with the modeled mean winter massbalance (+108plusmn 034 m we yrminus1) of Chhota Shigri Glacierduring 1969ndash2012 in the western Himalaya (Azam et al2013)

336 Mean elevation changes and mass balancecalculation

Before averaging elevation changes we exclude all interpo-lated pixels in the SRTM or SPOT5 DEMs as well as unex-pected elevation changes exceedingplusmn150 m for study sites(Pamir and Karakoram) that include surge-type glaciers orplusmn80 m for other sites Those thresholds were chosen aftercareful inspection of the elevation difference maps and his-tograms in order to identify the largest realistic glacier eleva-tion changes Glaciers that are truncated at the edges of theDEM are also excluded from mass balance calculations asthey may bias the final results if for example only their accu-mulation or ablation area is sampled Although they are trun-cated Siachen (Karakoram East) and Fedtchenko (Pamir)glaciers were retained because of their wide coverage andbecause their accumulation and ablation areas were correctlysampled in the DEMs

Elevation changes are then analyzed for 100 m altitudebins Within each altitude band we average the elevationchanges after excluding pixels where absolute elevation dif-ferences differ by more than three standard deviations fromthe mean (Berthier et al 2004 Gardner et al 2012) Thisis an efficient way to exclude outliers based on the assump-tion that elevation changes should be similar at a given alti-tude within each study site This assumption does not holdfor regions containing actively surging glaciers Thus forthe Pamir and Karakoram study sites surge-type glaciersare treated separately ie elevation changes are averaged by100 m altitude bands for each surge-type glacier individually

Where no elevation change is available for a pixel weassign to it the value of the mean elevation change of the

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1269

altitude band it belongs to in order to assess the mass bal-ance over the whole glacier area

The conversion of elevation changes to mass balance re-quires knowledge of the density of the material that has beenlost or gained and its evolution during the study period Giventhe lack of repeat measurement of density profiles over theentire snowfirnice column in the PKH we applied a con-stant density conversion factor of 850 kg mminus3 for all ourstudy sites (Sapiano et al 1998 Huss 2013)

The mass balance of each surge-type glacier is computedseparately and added (area-weighted) to the mass balance ofthe rest of the glaciers to estimate the mass balance of thewhole study site

For study sites with a high concentration of growingpro-glacial lakes such as Bhutan Everest and West Nepal(Gardelle et al 2011) our mass losses are slightly underes-timated because they do not take into account the glacier icethat has been replaced by water during the expansion of thelake Subaqueous ice losses are only estimated for LhotseSharImja Glacier (Everest area Fujita et al 2009) for thesake of comparison to previous studious

The region-wide mass balance of PKH glaciers as definedin Sect 2 is computed by extrapolating the mass balancefrom each study site to a wider sub-region (Fig 1) that is as-sumed to experience the same glacier mass changes becauseof its similar climate (Bolch et al 2012 Kaumlaumlb et al 2012)The total glacier area for each sub-region is computed fromthe RGI v20 (Arendt et al 2012)

The periods over which we calculate mass balance areslightly different from one site to another (Table 1) depend-ing on the year of acquisition of the SPOT5 DEM (two in2008 two in 2010 and five in 2011) In the following we willassume that all mass balances are representative of the period1999ndash2011 in order to estimate a region-wide mass budgetof PKH glaciers neglecting thus part of the inter-annual vari-ability

34 Accuracy assessment

The errorE1hi to be expected for a pixel elevation change

1hi is assumed to equal the standard deviationσ1h of themean elevation change1h of the altitude band it belongs toThe value ofσ1h can range fromplusmn4 m toplusmn 20 m dependingon the study site and elevation This is rather conservative asthe value ofσ1h contains also the intrinsic natural variabilityof elevation changes within the altitude band (ie it containsboth noise and real geophysical signal)

The errorE1h of the mean elevation change1h in eachaltitude band is then calculated according to standard princi-ples of error propagation

E1h =E1hiradicNeff

(1)

Neff represents the number of independent measurements inthe altitude bandNeff will be lower thanNtot the total num-

ber of elevation change measurements1hi in the altitudeband since the latter are correlated spatially (Bretherton etal 1999)

Neff =Ntot middot PS

2d (2)

where PS is the pixel size (here 40 m)d is the distance ofspatial autocorrelation determined using Moranrsquos I autocor-relation index on elevation differences off glaciers On theaverage over the 9 study sitesd = 492plusmn 72 m

The error on the penetration correction was assumed to besystematic due to the not fully understood physics involvedand equal toplusmn 15 m This error was obtained by compar-ing the SRTM C-band penetration estimated using two inde-pendent methods (i) SRTMC-bandndash SRTMX-band and (ii) dif-ference between ICESat and SRTM when ICESat elevationchange trends during 2003ndash2008 are extrapolated to Febru-ary 2000 (Kaumlaumlb et al 2012) The maximum difference of thepenetration inferred from the two methods is 10 m (Table 3)A 50 additional error margin was added to account for thefact that this comparison could only be made on 2 of ourstudy sites

Given the slender observational support for the seasonal-ity correction (Section 33) we assumed its uncertainty tobe plusmn100 on the western sites ieplusmn015 m we per win-ter month The same absolute uncertainty (plusmn015 m we perwinter month) was used for the eastern sites where no sea-sonality correction is applied but ablation andor accumula-tion may occur in winter (Wagnon et al 2013)

All errors are summed quadratically within each altitudeband and then summed for all altitude bands assuming con-servatively that they are 100 dependent The error on theSRTM penetration correction clearly dominates our errorbudget for volume change

During the conversion from volume to mass we assumea plusmn60 kg mminus3 error on the density conversion factor (Huss2013) Thisplusmn7 error is similar to the one estimated byBolch et al (2011) In a sensitivity test the volume to massconversion is also performed using two alternative densityscenarios (Kaumlaumlb et al 2012) (i) a density of 900 kg mminus3 isassumed everywhere and (ii) 600 kg mminus3 in the accumula-tion area and 900 kg mminus3 in the ablation area These two sce-narios take among other things into account that elevationchanges could be due to changes in glacier dynamics or dueto surface mass balance We then calculate the maximum ab-solute difference between the mass balances calculated usingour preferred scenario (850plusmn 60 kg mminus3 everywhere) and themass balances calculated using the two others scenarios Forthe eastern sites (West Nepal to Hengduan Shan) the max-imum absolute difference is 003 m we yrminus1 For the west-ern sites (Pamir to Spiti Lahaul) it reaches 006 m we yrminus1Those uncertainties due to the choice of a given density sce-nario remain negligible compared to other sources of errors

We also include an error due to extrapolation of our massbalance estimate from our study site (the extent of each

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1270 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

SPOT5 DEM) to the unmeasured glaciers within each sub-region To estimate this regional extrapolation error we com-pare the ICESat elevation change trends (computed as inKaumlaumlb et al 2012) within 3 times 3 cells centered at the studysites (Table 5) with the ICESat-derived trends for the entiresub-regions The absolute difference between both trends isless than 002ndash003 m yrminus1 for Bhutan West Nepal Spiti La-haul and Karakoram ie negligible For Everest the eleva-tion trend for the entire sub-region is 012 m yrminus1 less nega-tive than for the 3 times 3 cells around the SPOT5 DEM cen-ter for Hindu Kush the trend for the entire sub-region is017 m yrminus1 more negative than for the 3 times 3 cell due to thevariability of elevation changes within this sub-region (Kaumlaumlbet al 2012) The regional extrapolation error is obtainedas the area-weighted mean absolute difference between themass balances derived from ICESat for those 3 times 3 cellsand the whole sub-region and equalsplusmn004 m we yrminus1

Finally for sub-regions and total PKH estimates of masschange we include an error on the glacier area computedfrom the RGI v20 (Arendt et al 2012) The inventory er-ror is specific to each study site and evaluated based on thecomparison of our user-defined inventory on each study sitewith the RGI v20 (Table 1) We note the remarkable accu-racy of the RGI for all but one of our study sites The rel-ative errors are generally of a few percents and up to 12 for West Nepal The differences between our and existing in-ventories are probably due to the difficulty of delimitatingdebris-covered glacier parts (eg Frey et al 2012 Paul etal 2013) and accumulation areas The Hengduan Shan studysite is an exception There the RGI is based on the ChineseGlacier Inventory (Shi et al 2009 Arendt et al 2012) andoverestimates the ice-covered area by 88 compared to ourLandsat-based inventory

Unless stated otherwise all error bars from this study andpreviously published studies correspond to one standard er-ror

4 Results

41 Overview of the elevation changes

The maps of glacier elevation changes are given in Figs 2ndash10for our nine study sites with as inset the distribution ofthe elevation differences off glaciers The full maps of ele-vation differences off glaciers are shown in the Supplement(Figs S2ndashS10) Those off glacier maps indicate a local ran-dom noise of aboutplusmn5ndash10 m However at length scales of afew kilometers the spatial variations in elevation differencesare small typically less than 1 m

For the four eastern study sites from the HengduanShan to the West Nepal (Fig 1) the elevation changesaveraged over the accumulation areas are small (between003plusmn 014 m yrminus1 to 013plusmn 015 m yrminus1) whereas clear sur-face lowering is observed for all four study sites in their

ablation areas ranging fromminus050plusmn 014 m yrminus1 (Bhutan)to minus081plusmn 013 m yrminus1 (Hengduan Shan) Over the BhutanEverest and West Nepal study sites glaciers in contact witha proglacial lake often experienced larger thinning rates

Further west the Spiti Lahaul is the only site wheresurface lowering is measured both in the accumula-tion area (minus034plusmn 016 m yrminus1) and the ablation area(minus070plusmn 014 m yrminus1) Clear surface lowering is also ob-served in the ablation areas of the Hindu Kush glaciers(minus050plusmn 018 m yrminus1)

In the Karakoram (east and west Figs 8 and 9) andin the Pamir (Fig 10) the pattern of elevation changesis highly heterogeneous due to the occurrence of numer-ous glacier surges or similar flow instabilities The eleva-tion changes in the accumulation areas of those three studysites are slightly positive (between+022plusmn 014 m yrminus1 and+030plusmn 018 m yrminus1) In the ablation areas small surfacelowering is observed for the two Karakoram study sites (av-erage of the two sitesminus033plusmn 016 m yrminus1) whereas no el-evation change is found in Pamir (minus002plusmn 013 m yrminus1)

Note that those mean rates of elevation changes are sensi-tive to the choice of the ELA assumed to equal the snowlinealtitude in Landsat images from around year 2000

42 Influence of a debris cover

To evaluate thinning rates over debris-covered and debris-free ice we perform a histogram adjustment in order tocompare data sets with similar altitude distributions This isneeded because debris-covered parts tend to be concentratedat lower elevations The adjustment consists in randomly ex-cluding pixels over debris-free ice from each elevation bandto match a scaled version of the debris-covered pixel distribu-tion (Kaumlaumlb et al 2012) Elevation changes with altitude canthen be compared for clean and debris-covered ice (Fig 11)

The thinning is higher for debris-covered ice in the Ever-est area and on the lowermost parts of glaciers in the PamirBut on average in the Pamir western Karakoram and SpitiLahaul the lowering of debris-free and debris-covered ice issimilar The thinning is stronger over clean ice in the Heng-duan Shan Bhutan West Nepal Hindu Kush and althoughto a lesser extent in the eastern Karakoram

43 Mass balance over the nine study sites

Glacier mass changes have been heterogeneous overthe PKH since 1999 Slight mass gains or balancedmass budgets are found in the north-west in thePamir (+014plusmn 013 m we yrminus1) and for the two studysites in the central Karakoram (+009plusmn 018 m we yrminus1

and +011plusmn 014 m we yrminus1) Mass loss is moderatein the adjacent Hindu Kush with a mass balance ofminus012plusmn 016 m we yrminus1 Glaciers in all our eastern studysites (two in Nepal Bhutan and Hengduan Shan) experi-enced homogeneous mass losses with mass balances ranging

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1271

Fig 2 Glacier elevation changes over the Hengduan Shan study site Inset distribution of the elevation differences off glacier The medianand the standard deviation of the elevation difference and the number of pixels off glaciers are given (by construction the mean elevationdifference is null)

Fig 3Same as Fig 2 but for the Bhutan study site

from minus033plusmn 014 m we yrminus1 to minus022plusmn 012 m we yrminus1The most negative mass balance is measured in Spiti Lahaul(western Himalaya) atminus045plusmn 013 m we yrminus1 (Table 4)

However within a study site mass balances can be highlyvariable from one glacier to another In the Everest area twolarge glaciers (71 km2 each) experienced distinctly differentmass balance withminus016plusmn 016 m we yrminus1 for the south-flowing Ngozumpa Glacier andminus059plusmn 016 m we yrminus1

for the north-flowing Rongbuk Glacier (Fig 4) In Bhutan(Fig 3) mass loss is higher south (minus025plusmn 013 m we yrminus1)

than north (minus014plusmn 012 m we yrminus1) of the main Hi-malayan ridge though not at a statistically significant level

The region-wide mass budget of PKH glaciers isnegative minus101plusmn 55 Gt yrminus1 (minus014plusmn 008 m we yrminus1)which corresponds to a sea level rise contribution of0028plusmn 0015 mm yrminus1 over 1999ndash2011

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1272 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 4Same as Fig 2 but for the Everest study site

Fig 5Same as Fig 2 but for the West Nepal study site

44 Glacier surges

In the Pamir and the Karakoram the spatial pattern of ele-vation changes of many glaciers is typical of surge events orflow instabilities They can be identified because of their veryhigh thinning and thickening rates that can both reach up to16 m yrminus1 (Figs 8ndash10)

Most of them have already been reported to be surge-type glaciers based on their velocity (Kotlyakov et al 2008Quincey et al 2011 Copland et al 2009 Heid and Kaumlaumlb2012) morphology (Copland et al 2011 Barrand and Mur-ray 2006 Hewitt 2007) or elevation changes (Gardelle etal 2012a) Here we only identify surge-type glaciers for thesake of mass balance calculation and we do not attempt toprovide an exhaustive up-to-date surge inventory This is why

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1273

Fig 6 Same as Fig 2 but for the Spiti Lahaul study site The map of elevation changes contains many voids due to the presence ofthin clouds during the acquisition of the SPOT5 DEM The mass balance of the Chhota Shigri and Bara Shigri glaciers are respectivelyminus039plusmn 018 m we yrminus1 andminus048plusmn 018 m we yrminus1

the published surge inventories in the Pamir and the Karako-ram do not necessarily match the surge-type glaciers labeledon the elevation change maps as these refer only to our ob-servation period (1999ndash2011)

The mass balance of surge-type glaciers (respectivelynon-surge-type glaciers) is +019plusmn 022 m we yrminus1

(+012plusmn 011 m we yrminus1) in the Pamir+009plusmn 030 m we yrminus1 (+009plusmn 015 m we yrminus1) inthe western Karakoram and+010plusmn 025 m we yrminus1

(+012plusmn 012 m we yrminus1) in the eastern Karakoram Theoverall similarity between the mass budgets for surge-typeand non-surge-type glaciers shows that those flow instabili-ties seem not to affect the glacier-wide mass balance at leastover short time periods here 10 yr

5 Discussion

51 SRTM penetration

Alternative estimates of SRTMC-bandpenetration for the PKHregion were obtained by Kaumlaumlb et al (2012) by comparing el-evations of the ICESat altimeter (Zwally et al 2002) with theSRTMC-bandDEM in PKH and extrapolating the 2003ndash2008trend of elevation changes back to the acquisition date ofSRTM The difference between east and west is more dis-tinct in our case with a mean penetration in the Karakoram of34 m (24plusmn 03 m in Kaumlaumlb et al 2012) 24 m in Bhutan and14 m around the Everest area (25plusmn 05 m for a wider areaincluding east Nepal and Bhutan in Kaumlaumlb et al 2012) Thiseastwest gradient is expected given that in February 2000when the SRTM mission was flown the snowfirn thicknesswas probably larger in the western PKH where glaciers areof winter-accumulation type than in the eastern PKH where

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1274 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 7Same as Fig 2 but for the Hindu Kush study site

glaciers are of summer-accumulation type Further studieswould be necessary to investigate the possible reasons be-hind the relatively low SRTMC-bandpenetration for the Pamir(18 m) that do not follow the eastwest gradient mentionedabove

Both our and Kaumlaumlb et alrsquos (2012) estimates agree withintheir error bars and seem thus to provide robust first-orderestimates for the SRTMC-bandradar penetration in the regionHowever it remains to be verified if a penetration depth com-puted at the regional scale is appropriate for a single smallglacier whose altitude distribution is different from the oneof whole study site (eg Abramov Glacier in the far north ofour Pamir study site)

52 Thinning over debris-covered ice

PKH glaciers are heavily debris-covered in their ablation ar-eas because of steep rock walls that surround them and in-tense avalanche activity Twelve percent (12 ) of the glacierarea in our study sites are covered with debris (up to 22 inthe Everest area Table 1) which is in agreement with pre-vious estimates for the Himalaya and the Karakoram 13 in Kaumlaumlb et al (2012) and 10 in Bolch et al (2012) Thedebris layer is expected to modify the ablation of the under-

lying ice It will increase ablation if its thickness is just a fewcentimeters (dominance of the albedo decrease) or decreaseablation it if the layer is thick enough to protect the ice fromsolar radiation (Mattson et al 1993 Mihalcea et al 2006Benn et al 2012 Lejeune et al 2013)

However recent studies have reported similar overall thin-ning rates over debris-covered and debris-free ice in thePKH (Kaumlaumlb et al 2012 Gardelle et al 2012a) or evenhigher lowering rates over debris-covered parts in the Ever-est area (Nuimura et al 2012) Our findings are consistentwith Nuimura et al (2012) for the Everest study site with astronger thinning (rate of elevation change ofminus097 m yrminus1)

in debris-covered areas than over clean ice (rate of elevationchange ofminus057 m yrminus1) In the Pamir Karakoram and SpitiLahaul study sites these rates are similar while in the HinduKush West Nepal Bhutan and Hengduan Shan study sitesthinning is greater over clean ice in agreement with the com-monly assumed protective effect of debris (Fig 4) Thosedifferences between our 9 study sites suggest a complex re-lationship between debris cover and glacier wastage

As local glacier thickness changes are a result of bothmass balance processes and a dynamic component theseunexpected high thinning rates over debris-covered parts

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1275

Fig 8Same as Fig 2 but for the Karakoram east study site Surge-type glaciers in an active surge phase (resp quiescent phase) are identifiedwith a solid triangle (resp solid circle)

are potentially the result of glacier-wide processes Over aglacier tongue the ablation can be enhanced by the pres-ence of steep exposed ice cliffs or supraglacial lakes andponds (Sakai et al 2000 2002) In addition it is also pos-sible that the glacier tongues are mostly covered by thindebris layers such that the albedo effect dominates the in-sulating effect resulting in an increased tongue-wide abla-tion The prevalence of thin debris was recently suggestedfor debris-covered glaciers around the Mount Gongga in thesouth-eastern Tibetan Plateau (Zhang et al 2013b) Finallyice-flow rates could be very low at debris-covered parts asshown by Quincey et al (2007) in the Everest area Kaumlaumlb(2005) for Bhutan and Scherler et al (2011b) in the HinduKush-Karakoram-Himalaya Thus all three factors are likelyto increase the overall thinning under debris despite the un-doubted protective effect of a continuous thick debris coverat a local scale Future investigations combining surface ve-locity fields and maps of elevation changes could allow eval-uating more closely the relative role of ablation and ice fluxesin thinning rates over PKH glacier tongues (eg Berthierand Vincent 2012 Nuth et al 2012) Measurements of the

spatial distribution of debris thickness over the PKH glaciertongues are also needed

Note that in the case of debris-covered tongues the ele-vation change measurement represents the glacier thicknesschange but also the possible debris thickness evolution Thelatter remains poorly known in the PKH or in any othermountain range But based on the debris discharges givenby Bishop et al (1995) for Batura Glacier (Pakistan) 48to 90times 103 m3 yrminus1 the thickening of the debris can be es-timated between 27 and 55 mm yrminus1 which is negligiblegiven the magnitude of the glacier elevation changes

53 Comparison to previous mass balance estimates

Throughout the PKH there is only a single peer-reviewedglaciological record (on Chhota Shigri Glacier) coveringmost of our study period (Wagnon et al 2007 Vincent etal 2013) Our geodetic mass balance estimate for the SpitiLahaul study site has been recently compared to those glacio-logical measurements on Chhota Shigri Glacier (Vincent etal 2013) Mass balance records reported in Yao et al (2012)for the Tibetan Plateau and the surroundings mountain ranges

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1276 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 9Same as Fig 8 but for the Karakoram west study site

only start during glaciological year 20052006 This is whyhere we do not compare our mass balance estimates toglaciological records and limit our comparison to others re-gional estimates obtained using the geodetic or gravimetricmethods

We stress that the comparison of our mass balance mea-surements to published estimates is complicated by thefact that generally they do not cover the same time pe-riod Little is known about the inter-annual variability inthe PKH during the 21st century due to the limited num-ber of long glaciological time series The 9 yr mass balance

record on Chhota Shigri Glacier reveals a relatively highinter-annual variability of the annual mass balance (stan-dard deviationplusmn 074 m we yrminus1 during 2003ndash2011 Vin-cent et al 2013) which is similar to the one obtained be-tween 1968 and 1998 on Abramov Glacier (standard devi-ation 069 m we yrminus1 WGMS 2012) Thus inter-annualvariability can explain part of the difference between two cu-mulative mass balance estimates over 5ndash10 yr even if they areoffset by only one or two years

Based on ICESat repeat track measurements and theSRTM DEM Kaumlaumlb et al (2012) measured a region-wide

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1277

Fig 10Same as Fig 8 but for the Pamir study site

mass balance ofminus021plusmn 005 m we yrminus1 between 2003and 2008 for Hindu Kush-Karakoram-Himalaya glaciersFor that same area (ie excluding the Pamir and theHengduan Shan study sites) we found a mass balance ofminus015plusmn 007 m we yrminus1 Thus our result agrees with Kaumlaumlbet al (2012) within the error bars although our study periodis longer (1999ndash2011) and our measurement principle andextrapolation approach are different In addition we havecalculated updated mass balances from Kaumlaumlb et al (2012)ie averages over 3times 3 degree cells centered over our studysites and find that they are consistent within error barswith our mass balance estimates (Table 5) Similar resultsare found when our mass balances are compared to a spa-tially extended ICESat analysis for 2003ndash2009 (Gardner etal 2013) Mass losses measured with ICESat (Kaumlaumlb et al2012 Gardner et al 2013) tend to be slightly larger thanours This could be due to the difference in time periodsAlso our smaller mass losses could be partly explained if wesystematically underestimated the poorly constrained pene-

tration of the C-band andor X-band radar signal into ice andsnow

At a more local scale Bolch et al (2011) reported a massbalance ofminus079plusmn 052 m yrminus1 between 2002 and 2007 byDEM differencing over a glacier area of 62 km2 includingten glaciers south and west of Mt Everest For these sameglaciers Nuimura et al (2012) found a mean mass balanceof minus045plusmn 060 m yrminus1 during 2000ndash2008 while they com-puted a mass balance ofminus040plusmn 025 m yrminus1 between 1992and 2008 over a glacier area of 183 km2 around Mt EverestIn the present study we found an average mass balance ofminus041plusmn 021 m yrminus1 over the ten glaciers mentioned above(Table 5 Fig S1) and a value ofminus026plusmn 013 m yrminus1 overthe whole Everest study site between 2000 and 2011 Ourvalues are comparable to those of Nuimura et al (2012)Our mass balance is less negative than the 2002ndash2007 massbalance of Bolch et al (2011) but not statistically differentwhen error bars are considered (Fig 12) Indeed Bolch etal (2011) acknowledged some high uncertainties for their

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1278 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Table 4Glacier mass budget of the nine sub-regions to which the results of the study sites (mass balances) have been extrapolated For eachof them the total glacierized area as well as the percentage of the glacier area over which the mass balance was computed (ie study sites)are given The total glacier area is derived from the RGI v20 (Arendt et al 2012) The error for the mass balance in each sub-region is theroot of sum of squares (RSS) of the mass balance error of each study site and the regional extrapolation error The error for the mass budgetin each sub-region includes additionally the error from the total glacier area in the RGI v20

Sub-region Total glacier Measured glacier Mass balance Mass budgetarea (km2) area () (m we yrminus1) (Gt yrminus1)

Hengduan Shan 11584 12 minus033plusmn 014 minus38plusmn 38Bhutan 4021 34 minus022plusmn 013 minus09plusmn 05Everest 6226 23 minus026plusmn 014 minus16plusmn 08West Nepal 6849 13 minus032plusmn 014 minus22plusmn 10Spiti Lahaul 9043 23 minus045plusmn 014 minus41plusmn 13Hindu Kush 6135 13 minus012plusmn 016 minus07plusmn 10Karakoram East

1902428 +011 plusmn 014 +19 plusmn 31

Karakoram West 30 +009 plusmn 018Pamir 9369 34 +014 plusmn 014 +13 plusmn 13

TotalArea- 72251 30 minus014plusmn 008 minus101plusmn 55weighted average

Table 5Comparison of geodetic mass balances between this study and previous published results with overlapping study periods and similargeographic locations Updated figures from Kaumlaumlb et al (2012) are averaged over 3 times 3 cells centered over the study sites of the presentstudy

GlacierSite name This study Bolch et al (2011) Nuimura et al (2012) Kaumlaumlb et al (2012 updated)2002ndash2007 2000ndash2008 2003ndash2008

Study Mass balance Area (km2)a Mass balance Area Mass balance Mass balanceperiod (m we yrminus1) (m we yrminus1) (km2) (m we yrminus1) (m we yrminus1)

Hengduan San 1999ndash2010 minus022 plusmn 014 1303 (55 )Bhutan 1999ndash2010 minus022 plusmn 012 1384 (64 ) minus052 plusmn 016b

Everest

1999ndash2011

minus026 plusmn 013 1461 (58 ) minus039 plusmn 011AX010 minus090plusmn 034 04Changri SharNup minus042plusmn 017 161 (79 ) minus029plusmn 052 130 minus055plusmn 038Khumbu minus051plusmn 019 203 (47 ) minus045plusmn 052 170 minus076plusmn 052Nuptse minus037plusmn 020 50 (74 ) minus040plusmn 053 40 minus034plusmn 027Lhotse Nup minus021plusmn 027 24 (70 ) minus103plusmn 051 19 minus022plusmn 047Lhotse minus043plusmn 018 85 (83 ) minus110plusmn 052 65 minus067plusmn 051Lhotse SharImja minus070plusmn 052 98 (44 ) minus145plusmn 052 107 minus093plusmn 060Amphu Laptse minus046plusmn 034 25 (38 ) minus077plusmn 052 15 minus018plusmn 094Chukhung +044 plusmn 024 42 (47 ) +004 plusmn 054 38 +043 plusmn 081Ama Dablam minus049plusmn 017 36 (54 ) minus056plusmn 052 22 minus056plusmn 073Duwo minus016plusmn 026 19 (18 ) minus196plusmn 053 10 minus068plusmn 074Total Khumbu minus041plusmn 021 744 minus079plusmn 052 617 minus045plusmn 060(10 Glaciers above)

West Nepal 1999ndash2011 minus032 plusmn 013 908 (40 ) minus032 plusmn 012Spiti Lahaul 1999ndash2011 minus045 plusmn 013 2110 (46 ) minus038 plusmn 006Karakoram East 1999ndash2010 +011 plusmn 014 5328 (42 ) minus004 plusmn 004Karakoram West 1999ndash2008 +009 plusmn 018 5434 (45 )Hindu Kush 1999ndash2008 minus012 plusmn 016 793 (80 ) minus020 plusmn 006Pamir 1999ndash2011 +014 plusmn 013 3178 (50 )

a The total glacier area is given in km2 and in parenthesis the of the glacier area actually covered with measurementsb For this cell the ICESat coverage is insufficient and does not sample all glacier elevations

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1279

Fig 11 Elevation changes (SPOT5-SRTM) in the ablation area for clean ice (blue circles) and debris-covered ice (black circles) for eachstudy site For the Karakoram (east and west) and the Pamir study sites only non-surge-type glaciers are considered Standard error of theelevation differences are typically less thanplusmn2 m with each 100 m elevation band (not shown for the sake of clarity) Note elevation changesover separate sections of a glacier cannot be treated as mass changes due to the disregard of glacier dynamics

estimate over this short time period The differences in sur-vey periods and in the glacier outlines may also explainpart of these discrepancies In particular the delimitation ofthe accumulation area is notoriously difficult and Bolch etal (2011) did not include some of the steep high altitudeslopes which we included due to their potential avalanchecontribution For the 10 glaciers mentioned above our massbalances would become 012 m we yrminus1 more negative ifwe used the exact same outlines as Bolch et al (2011)Our positive mass balance for the small Chukhung Glacier(+044plusmn 024 m we yrminus1) is in agreement with Nuimura etal (2012) but enigmatic in a context of regional glacier lossand deserves further investigation If the 2002ndash2007 massbalance estimate by Bolch et al (2011) is excluded all otherrecent mass balance estimates suggest no acceleration of themass loss in the Everest area when compared to the long-term(1970ndash2007) mass balance measured by Bolch et al (2011)in this regionminus032plusmn 008 m weyrminus1

In Bhutan our results indicate a stronger mass loss forsouthern glaciers than for northern ones Interestingly in thisarea Kaumlaumlb (2005) measured high speeds (up to 200 m yrminus1)

for glaciers north of the Himalayan main ridge and ratherlow velocities over southern glacier tongues by using repeat

ASTER data This is consistent with the more negative massbalance that we report for the southern glaciers in Bhutanie for the slow flowing presumably downwasting glaciers

Berthier et al (2007) reported a mass balance betweenminus069 andminus085 m we yrminus1 for an ice-covered area of915 km2 around Chhota Shigri Glacier between 1999 (SRTMDEM) and 2004 (SPOT5-HRG DEM) For the same areawe found a mass balance ofminus045plusmn 013 m we yrminus1 over1999ndash2011 The difference in the study period may explainpart of the differences Also the methodology by Berthieret al (2007) did not explicitly take into account the SRTMradar penetration over glaciers and also corrected an appar-ent elevation-dependent bias according to glacier elevationswhich is now known as a resolution issue and is best cor-rected according to the terrain maximum curvature (Gardelleet al 2012b) More details and discussions about these dif-ferences can be found in Vincent et al (2013)

Importantly the results of the eastern Karakoram studysite confirm the balanced or slightly positive mass budgetreported by Gardelle et al (2012a) in the western Karako-ram over the past decade The two Karakoram study sitesare adjacent with a small overlapping area (sim 1100 km2 ofglaciers) where elevation differences agree within 1 m Both

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1280 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 12 Comparison of geodetic mass balances for 10 glaciers inthe Mount Everest area 1999ndash2010 mass balances measured inthe present study are compared with published estimates during2002ndash2007 (Bolch et al 2011) and during 2000ndash2008 (Nuimuraet al 2012) The 1-to-1 line is drawn The mean difference (plusmn1standard deviation) between (i) our values and Bolch et al (2011)are (047plusmn 058 m we yrminus1) and (ii) our values and Nuimura etal (2012) are (012plusmn 021 m we yrminus1) Part of the differences be-tween estimates may be due to the different periods surveyed andthe different glacier outlines used

sites show similar mass balances over slightly different pe-riods +011plusmn 014 m yrminus1 over 1999ndash2010 in the east and+009plusmn 018 m yrminus1 over 1999ndash2008 in the west This con-firms a ldquoKarakoram anomalyrdquo that was first suggested basedon field observations (Hewitt 2005) We report a mass bal-ance of+010plusmn 013 m yrminus1 for Baltoro Glacier (see loca-tion in Fig 8 size= 304 km2) between 1999 and 2010which is consistent with its gradual speed-up reported since2003 (Quincey et al 2009) However given the completelydifferent methods used in our and the latter study we can-not conclude that this demonstrates a real shift from nega-tive to positive mass balance The mass budget of SiachenGlacier (1145 km2 sometimes referred to as the highestbattlefield in the world) is also balanced or slightly pos-itive (+014plusmn 014 m we yrminus1) Thus the alarming state-ment by Rasul (2012) that this glacier retreated by 59 kmand lost 17 of its mass during 1989ndash2009 is very likelyerroneous Again within error bars our regional mass bal-ance in the Karakoram agrees with the mass balance ofminus003plusmn 004 m yrminus1 found by Kaumlaumlb et al (2012) over 2003ndash2008 and with the mass balance ofminus010plusmn 018 m we yrminus1

(2-sigma error level) obtained by Gardner et al (2013) for aregion including both the Karakoram and the Hindu Kush

In the northernmost part of the Pamir in the Alay moun-tain range the mass balance of Abramov Glacier has beenmeasured in the field for 30 yr between 1968 and 1998(WGMS 2012) with a mean value ofminus046 m yrminus1 Herewe report a mass balance ofminus003plusmn 014 m yrminus1 for this

glacier over 1999ndash2011 Our near zero mass budget estimatefor this glacier disagrees with negative mass balances recon-structed for the same period with a model run using biascorrected NCEPNCAR re-analysis data and calibrated withfield data (Barundun et al 2013) An underestimate of ourSRTM C-Band penetration in the Pamir as a whole and inparticular for the small Abramov Glacier may contribute tothese differences

For the whole Pamir study site our mass budget is bal-anced or slightly positive (+014plusmn 013 m we yrminus1) In theeastern Pamir the mass balance of Muztag Ata Glacier was+025 m we yrminus1 between 2005 and 2010 (Yao et al 2012)Although Muztag Ata Glacier is located outside of our Pamirstudy site and his glaciological records only cover the sec-ond half of our study period this measurement is consistentwith our remotely sensed mass balance Thus the ldquoKarako-ram anomalyrdquo should probably be renamed the ldquoPamir-Karakoram anomalyrdquo

This slightly positive or balanced mass budget measuredin the Pamir seems to contradict previous findings by Heidand Kaumlaumlb (2012) They reported rapidly decreasing glacierspeeds in this region between two snapshots of annual ve-locities in 20002001 and 20092010 an observation thatwould fit with negative mass balances (Span and Kuhn 2003Azam et al 2012) We note though that the relationship be-tween glacier mass balance and ice fluxes is not straightfor-ward and might in particular involve a time delay betweena change in mass balance and the related dynamic response(Heid and Kaumlaumlb 2012) Thus the decrease in speed couldwell be due to negative mass balances before or partiallybefore 1999ndash2011 We acknowledge that this discrepancyneeds to be investigated further in particular by examiningcontinuous changes in annual glacier speed through the firstdecade of the 21st Century Indeed Heid and Kaumlaumlb (2012)measured differences between two annual periods which maynot represent mean decadal velocity changes

Jacob et al (2012) using satellite gravimetry (GRACE)measured a mass budget ofminus5plusmn 6 Gt yrminus1 over 2003ndash2010for the ldquoKarakoram-Himalayardquo their region 8c which cor-responds to our study area excluding the Pamir Karakoramand Hindu Kush sub-regions but including some glaciersnorth of the Himalayan ridge already on the Tibetan PlateauFor this subset of our PKH study area we measured amass budget ofminus126plusmn 42 Gt yrminus1 The two estimates over-lap within their error bars especially if our error is dou-bled to match the two standard error level used by Jacob etal (2012) The estimate of Jacob et al (2012) over our com-plete study region (PKH) ieminus6plusmn 8 Gt yrminus1 (sum of theirregions 8b and 8c) also includes mass changes for the KunlunShan sub-region for which we did not measure glacier massbalances Therefore the comparison with our PKH value(minus101plusmn 55 Gt yrminus1) is not straightforward but suggests areasonable agreement especially if we consider the slightmass gain (15plusmn 17 Gt yrminus1) measured with ICESat in theWest Kunlun (Gardner et al 2013)

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1281

Table 6 Annual average of the glacier seasonal and decadal mass loss contributions to Indus Ganges and Brahmaputra discharges Basinarea and seasonal contributions are given by Kaser et al (2010) glacier area in each basin is derived from the RGI v20 (Arendt et al 2012)Numbers in parentheses indicate the percentage of glacierized area within the river basin Seasonal contributions were modeled by Kaser etal (2010) based on the CRU 20 dataset over 1961ndash1990 assuming a balanced glacier mass budget Glacier decadal mass loss contributionsare given as annual average over 1999ndash2011

River Basin area Glacier area Glacier contribution Mean discharge(km2) (km2) (m3 sminus1) (m3 sminus1)

Seasonally delayed Decadal mass lossKaser et al (2010) (this study)

Indus 1 139 814 25 598 (2 ) 105 103plusmn 72 2146 (Chevallier personalcommunication 2012)

Ganges 1 023 609 11 168 (1 ) 47 103plusmn 49 11739 Papa et al (2012Brahmaputra 527 666 15 296 (3 ) 33 147plusmn 64 19710 updated)

54 Reasons for the heterogeneous pattern of massbalance

The PKH glacier mass balance is slightly negative(minus014plusmn 008 m we yrminus1) but changes are not homoge-neous from one sub-region to another and seem to coincidewith the climatic settings of the mountain range For the east-ern sites (from the Hengduan Shan to West Nepal) which areunder the influence of the Indian and south-east Asian sum-mer monsoons the mass balances are moderately negative(sim minus026 m we yrminus1) In the northwest of the PKH at thePamir and Karakoram sites where glacier climate is char-acterized by the westerlies in winter mass balances are bal-anced or slightly positive (sim +010 m we yrminus1) In betweenthese two influences the glaciers of the Spiti Lahaul sitein a monsoon-arid transition zone experienced the strongestmass loss (minus045plusmn 013 m we yrminus1)

This heterogeneous pattern of mass balances seems to berelated to trends in precipitation throughout the PKH Archerand Fowler (2004) reported an increase in winter precipi-tation over the Upper Indus Basin since 1961 a trend con-firmed by Yao et al (2012) over the Pamir and the Karako-ram based on the analysis of the Global Precipitation Cli-matology Project data set (GPCP Adler et al 2003) Thisprecipitation increase is a source of greater accumulation forglaciers and thus a possible explanation for their slightly pos-itive mass budgets In addition in the Upper Indus Basin themean summer temperature has been decreasing since 1961(Fowler and Archer 2006) which is consistent with the find-ings of Shekhar et al (2010) in the Karakoram who reporteda decrease in both maximum and minimum temperature since1988 These increases in winter precipitation and summertemperature likely translate in greater accumulation and re-duced ablation changes which are consistent with the bal-anced or slightly positive glacier mass budget

On the other hand in the east of the PKH the Indiansummer monsoon has been weakening since the 1950s (Bol-lasina et al 2011) which could have reduced the amount

of snowfall over the eastern study sites and thus resultedin negative mass balances In addition maximum and min-imum temperatures increased since 1988 in the western Hi-malaya (Shekhar et al 2010) and maximum temperaturesincreased in the central Himalaya (Nepal) since the mid-1970s (Shrestha et al 1999) Thus both precipitation andtemperature trends in the Himalaya likely contributed to thenegative mass balances measured over the correspondingstudy sites (from Spiti Lahaul to Hengduan Shan Fig 1)

However gridded data sets such as GPCP are not neces-sarily representative of the climate at high altitude IndeedHewitt (2005) reported a 5-to-10-fold increase in precipi-tation between the glacier front and the accumulation areain the Karakoram that is not captured by reanalysis dataas recently confirmed by Immerzeel et al (2012) Thus di-rect measurements of accumulation on High Mountain Asiaglaciers (eg Azam et al 2013) and high-altitude weatherstations are eagerly needed to validate the large scale grid-ded data set (eg reanalysis) test their ability to describe thespecific climate near to PKH glaciers and assess the relativerole played by temperature and precipitation changes

55 Contribution to water resources

Our glacier mass balance can be averaged over large riverbasins to estimate the mean contribution to river runoffdue to a decade of glacier mass loss We assume therebyonly direct river runoff without loss terms For the threerivers for which we cover the entire glacierized area of theircatchments within our sub-regions (Fig 1) these contribu-tions to the river discharge are equal to 103 m3 sminus1 (Indus)103 m3 sminus1 (Ganges) and 147 m3 sminus1 (Brahmaputra) for theperiod 1999ndash2011 (Table 6)

Rivers of PKH being key water resources measures oftheir discharges are highly sensitive and difficult to access forpolitical reasons Papa et al (2012) built recent time series ofdischarges for Ganges and Brahmaputra by using altimetrymeasurements The locations of the corresponding dischargeestimates in Bangladesh are given in Fig 1 We can therefore

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1282 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

evaluate the relative contribution of glacier decadal mass loss(Table 6) to annual river run-off at 09 for the Gangesat Hardinge (basin area ofsim 1 020 000 km2) and 07 forBrahmaputra at Bahadurabad (basin area ofsim 530 000 km2)between 1999 and 2011 Given the discharge measured at theGuddu station by the Surface Water Hydrology Project of theWater and Power Development Authority (Pakistan see lo-cation on Fig 1) between 1999 and 2003 we can estimatethe contribution of glacier mass loss at 48 for the IndusRiver at this station

The seasonal contributions of glaciers to river run-off (iethe part of the annual precipitation that experiences season-ally delayed release from the glaciers) directly results fromseasonal variations of meteorological quantities in contrastto the glacier mass loss contribution which results from thelong-term climatic change Both runoff components directlyaffect the amount of water available in a river at a given lo-cation and at a given time so that their comparison could beof interest Even if the seasonal timing of the glacier massloss contribution is unclear it could in general peak at asimilar time of the year than the seasonal glacier contribu-tion so that both components rather enlarge each other thancancel Seasonal glacier contributions have been modeled byKaser et al (2010) The latter study assumed glaciers to bein equilibrium with the present climate and thus prescribedglacier mass balances equal to 0 Thus their seasonally de-layed glacier contributions do not include the part due todecadal glacier mass loss and is complementary to our es-timates (Table 6) For the eastern basins under the influenceof the Indian summer monsoon (Ganges and Brahmaputra)the contribution of glacier mass loss is higher than the sea-sonal contribution In other words for the first decade of the21st century the mass loss of glaciers is on average moreimportant for water availability in those two rivers than theirseasonal water storage effect The situation is different for theIndus Basin where the glacier melt season is also the dry sea-son which results in a relatively higher seasonally delayedglacier contribution to the river discharge There the season-ally delayed water release from the glaciers and the decadalglacier mass loss contribute on average equally to the riverdischarge Both contributions are strongly (and equally) de-pendent on the location of the discharge measurement andwill be significantly larger (in relative term) in more heav-ily glacierized and smaller mountain catchments located inthe upper part of these major river basins (Bookhagen andBurbank 2010)

6 Conclusion

In this study we assessed the spatial pattern of glaciermass balance over the PKH by comparing the SRTM andSPOT5 DEMs over 9 well-distributed study sites between1999 and 2011 We found slightly positive or balanced massbudgets in the western part (Pamir Karakoram) moder-ate mass loss in the east (Nepal Bhutan Hengduan Shan)

and the most negative mass balance in the monsoon-aridtransition zone (Spiti Lahaul in the western Himalaya)By extrapolating these values to climatically homogeneoussub-regions we estimate the PKH overall mass balanceto have beenminus014plusmn 008 m we yrminus1 between 1999 and2011 which corresponds to a contribution to sea level riseof 0028plusmn 0015 mm yrminus1 This is about 4 of the globalglacier mass loss estimated by Gardner et al (2013) thoughover a slightly shorter time period (2003ndash2009) The largestsource of uncertainties in those geodetic mass balance esti-mates is the poorly constrained penetration of the C-BandSRTM radar signal into snow and ice

Our technique of DEM differencing allows mapping in de-tail the spatial pattern of glacier elevation changes Thosemaps reveal many surge-type behaviors both in the Pamirand the Karakoram It also appears that the glacier-wide massbalance does not seem to be considerably affected by themass transfer from surges over thesim 10 yr time period stud-ied Similar lowering rates over debris-covered and clean iceare observed for four study sites despite the commonly as-sumed insulating effect of debris covers This suggests forthe scale of entire glacier tongues a spatial mixture of re-duced ablation under intact thick debris covers and enhancedablation by supraglacial lakes or ice cliffs or due to thin de-bris layer and very low glacier velocities in ablation areasthat reduce the ice advection downstream

The regionally heterogeneous pattern of ice wastage is inbroad agreement with contrasted trends in large-scale pre-cipitation and temperature However glaciological (eg ac-cumulation) and meteorological records are needed at highaltitude in the vicinity of glaciers to evaluate more closelythe local influence of atmospheric variables on glacier massbalance and fully understand the reasons behind those con-trasted mass budgets in the PKH

Supplementary material related to this article isavailable online at httpwwwthe-cryospherenet712632013tc-7-1263-2013-supplementpdf

Acknowledgements We thank G Cogley A Gardner T NuimuraT Bolch P Wagnon M Barandun M Hoelzle M Huss and ananonymous referee for their constructive comments that led to aconsiderably improved manuscript We thank the Water and PowerDevelopment Authority (WAPDA) for their hydrological data(processed by P Chevallier) and F Papa for sharing his updatedrun-off data ahead of publication We thank T Bolch for sharinghis glacier outlines from the Everest area We thank the USGSfor allowing free access to their Landsat archive CGIAR-SCI forSRTM C-band data DLR for SRTM X-band data and GLIMS forglacier inventories J Gardelle acknowledges a PhD fellowshipfrom the French Space Agency (CNES) and the French NationalResearch Center (CNRS) E Berthier acknowledges support fromCNES through the TOSCA and ISIS proposal no 397 and fromthe Programme National de Teacuteleacutedeacutetection Spatiale E Berthieracknowledges M Masiokas and R Villalba for welcoming him

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1283

at the IANIGLA (Mendoza Argentina) Y Arnaud acknowledgessupport from French National Research Agency through ANR-09-CEP-005-01PAPRIKA A Kaumlaumlb acknowledges support fromESA through Glaciers_cci (400010177810I-AM) and from theEuropean Research Council under the European Unionrsquos SeventhFramework Programme (FP2007-2013)ERC Grant Agreementn 320816

Edited by I M Howat

The publication of this articleis financed by CNRS-INSU

References

Adler R Huffman G Chang A Ferraro R Xie P JanowiakJ Rudolf B Schneider U Curtis S Bolvin D Gruber ASusskind J Arkin P and Nelkin E The Version-2 GlobalPrecipitation Climatology Project (GPCP) Monthly Precipita-tion Analysis (1979ndashPresent) J Hydrometeorol 4 1147ndash11672003

Archer D R and Fowler H J Spatial and temporal variationsin precipitation in the Upper Indus Basin global teleconnectionsand hydrological implications Hydrol Earth Syst Sci 8 47ndash61doi105194hess-8-47-2004 2004

Arendt A Bolch T Cogley J G Gardner A Hagen J-OHock R Kaser G Pfeffer W T Moholdt G Paul F RadicV Andreassen L Bajracharya S Beedle M Berthier EBhambri R Bliss A Brown I Burgess E Burgess DCawkwell F Chinn T Copland L Davies B de AngelisH Dolgova E Filbert K Forester R Fountain A Frey HGiffen B Glasser N Gurney S Hagg W Hall D Hari-tashya U K Hartmann G Helm C Herreid S Howat IKapustin G Khromova T Kienholz C Koenig M KohlerJ Kriegel D Kutuzov S Lavrentiev I LeBris R Lund JManley W Mayer C Miles E Li X Menounos B Mer-cer A Moelg N Mool P Nosenko G Negrete A NuthC Pettersson R Racoviteanu A Ranzi R Rastner P RauF Rich J Rott H Schneider C Seliverstov Y Sharp MSigurethsson O Stokes C Wheate R Winsvold S WolkenG Wyatt F and Zheltyhina N Randolph Glacier Inventory[v20] A Dataset of Global Glacier Outlines Global Land IceMeasurements from Space Boulder Colorado USA Digital Me-dia httpwwwglimsorgRGIrandolphhtml 2012

Azam M Wagnon P Ramanathan A Vincent C Sharma PArnaud Y Linda A Pottakkal J Chevallier P Singh V andBerthier E From balance to imbalance a shift in the dynamicbehaviour of Chhota Shigri glacier western Himalaya India JGlaciol 58 315ndash324 doi1031892012JoG11J123 2012

Azam M F Wagnon P Vincent C Ramanathan A Linda Aand Singh V B Reconstruction of the annual mass balanceof Chhota Shigri Glacier (Western Himalaya India) since 1969Ann Glaciol in revision 2013

Barrand N and Murray T Multivariate Controls on the In-cidence of Glacier Surging in the Karakoram Himalaya

Arct Antarct Alp Res 38 489ndash498 doi1016571523-0430(2006)38[489MCOTIO]20CO2 2006

Barundun M Huss M Azisov E Gafurov A HoelzleM Merkushkin A Salzmann N and Usubaliev R Re-establishing seasonal mass balance observation at AbramovGlacier Kyrgyzstan from 1968ndash2012 Abstract EGU2013-5668European Geophysical Union Vienna 2013

Benn D Bolch T Hands K Gulley J Luckman ANicholson L Quincey D Thompson S Toumi R andWiseman S Response of debris-covered glaciers in theMount Everest region to recent warming and implicationsfor outburst flood hazards Earth-Sci Rev 114 156ndash174doi101016jearscirev201203008 2012

Berthier E and Vincent C Relative contribution of surface mass-balance and ice-flux changes to the accelerated thinning of Merde Glace French Alps over 1979ndash2008 J Glaciol 58 501ndash512doi1031892012JoG11J083 2012

Berthier E Arnaud Y Baratoux D Vincent C and Reacutemy FRecent rapid thinning of the ldquo Mer de Glace rdquo glacier derivedfrom satellite optical images Geophys Res Lett 31 L17401doi1010292004GL020706 2004

Berthier E Arnaud Y Kumar R Ahmad S Wagnon P andChevallier P Remote sensing estimates of glacier mass balancesin the Himachal Pradesh (Western Himalaya India) RemoteSens Environ 108 327ndash338 doi101016jrse2006110172007

Bhambri R Bolch T Kawishwar P Dobhal D P SrivastavaD and Pratap B Heterogeneity in Glacier response from 1973to 2011 in the Shyok valley Karakoram India The CryosphereDiscuss 6 3049ndash3078 doi105194tcd-6-3049-2012 2012

Bhutiyani M Mass-balance studies on Siachen Glacier in theNubra valley Karakoram Himalaya India J Glaciol 45 112ndash118 1999

Bishop M P Schroder J F and Ward J L SPOT multispec-tral analysis for producing supraglacial debris-load estimates forBatura glacier Pakistan Geocarto Int 10 81ndash90 1995

Bolch T Pieczonka T and Benn D I Multi-decadal mass lossof glaciers in the Everest area (Nepal Himalaya) derived fromstereo imagery The Cryosphere 5 349ndash358 doi105194tc-5-349-2011 2011

Bolch T Kulkarni A Kaumlaumlb A Huggel C Paul F Cogley JFrey H Kargel J Fujita K Scheel M Bajracharya S andStoffel M The state and fate of Himalayan Glaciers Science336 310ndash314 doi101126science1215828 2012

Bollasina M Ming Y and Ramaswamy V Anthropogenicaerosols and the weakening of the South Asian summer mon-soon Science 334 502ndash505 doi101126science12049942011

Bookhagen B and Burbank D Toward a complete Himalayan hy-drological budget spatiotemporal distribution of snowmelt andrainfall and their impact on river discharge J Geophys Res115 F03019 doi1010292009JF001426 2010

Bretherton C Widmann M Dymnikov V Wallace J and BladeacuteI The effective number of spatial degrees of freedom of a time-varying field J Climate 12 1990ndash2009 1999

Copland L Pope S Bishop M Shroder J Clendon PBush A Kamp U Seong Y and Owen L Glacier veloc-ities across the central Karakoram Ann Glaciol 50 41ndash49doi103189172756409789624229 2009

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1265

Fig 1 Overview of the extent and location of the nine study sites (black polygons) and corresponding sub-regions (red polygons) along thePKH range Brownish-filled background polygons represent the major river basins of the area Triangles indicate the location of dischargemeasurements discussed in Sect 55

PKH range could not be achieved because of funding restric-tions (SPOT5 is a commercial satellite) the busy acquisitionschedule of this satellite and cloud coverage A further con-strain is that the optical stereo images used to compute theDEMs should be acquired as close as possible to the end ofthe glacier ablation season In practice a time window of2ndash3 months is allowed for the acquisitions which is shortgiven the 26-day orbital cycle of SPOT5 Thus a maximumof 3 to 4 acquisition attempts are possible each year and onlyif the satellite is not programmed for higher priority targets(eg crisis situations) Despite these difficulties our SPOT5DEMs sampled in total 22 200 km2 of glaciers about 30 of the total glacierized area in the PKH (Arendt et al 2012)

The eastern-most sites (Hengduan Shan Bhutan Everestand West Nepal) are strongly influenced by the Indian andsouth-east Asian summer monsoons Ablation and accumula-tion of monsoon-type glaciers occur during the summer sea-son (eg Fujita 2008) Towards the other end of the PKHin the north-west (Pamir Hindu Kush and Karakoram) theclimate is dominated by the westerlies and glaciers are ofwinter-accumulation type The Spiti Lahaul site lies in a tran-sition zone influenced both by the monsoon and the wester-lies The topography also plays a strong role in the moisturetransfer as it dries out the southerly air flow and can preventthe air mass from travelling further north which results ina northward decrease of precipitation (Bookhagen and Bur-bank 2010)

Monsoon-type glaciers are expected to be more sensitivethan winter-accumulation type glaciers to a change in therainsnow limit driven by an increase in temperature (Fujita2008) Indeed a summer warming would increase the alti-

tude of the rainsnow limit and reduce snow accumulation onglaciers as well as decrease their surface albedo (less freshsnow with a high albedo) and thus enhance ablation

The Karakoram and the Pamir are known to host numer-ous surge-type glaciers (Barrand and Murray 2006 Hewitt2007 Kotlyakov et al 2008 Gardelle et al 2012a) char-acterized by the alternation between short active phases in-volving rapid mass transfer from high to low elevationsand longer quiescent phases of low mass fluxes Coplandet al (2011) reported a doubling in the number of glaciersurges after 1990 in the Karakoram with a total of 90 surge-type glaciers observed in the region since the late 1960sIn the Pamir an inventory from 1991 revealed 215 glacierswith signs of repeated surging (looped moraines fast ad-vance of the front) and 51 additional actively surging glaciers(Kotlyakov et al 2008)

Many PKH glaciers are heavily debris-covered in their ab-lation areas because of steep rock walls that surround themand intense avalanche activity (eg Bolch et al 2012) Thesize of these debris zones is highly variable and debris thick-ness can range from a few centimeters (dust or sand) to sev-eral meters (Mihalcea et al 2008 Zhang et al 2013b) Themean debris cover is about 10 of the total glacier area inthe Karakoram and Himalaya (Bolch et al 2012) and culmi-nate at 36 in the Central Himalaya (Scherler et al 2011a)

In the eastern PKH glacier termini are often connected topro-glacial lakes storing meltwater behind frontal morainesor dead-ice dams whereas in the western PKH pro-glaciallakes are less numerous and glacier ablation areas are oftenscattered with supra-glacial lakes (Gardelle et al 2011) re-sulting from successive coalescence of small melting ponds

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1266 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Table 1 Characteristics of the nine study sites Numbers between parentheses indicate the percentage difference between our user-definedglacier inventory and the RGI v20 (Arendt et al 2012)

Site name Glacier area Debris ( of Glacier area Arendt Date Landsat imagesthis study (km2) glacier area) et al (2012) (km2) (SPOT5 DEM)

PathRow Date

Hengduan Shan 1303 11 2541 (+88 ) 24 Nov 2011 p135r39 23 Sep 1999Bhutan 1384 11 1429 (+3 ) 20 Dec 2010 p138r40 19 Dec 2000

p138r41Everest 1461 22 1400 (minus4 ) 4 Jan 2011 p140r41 30 Oct 2000West Nepal 908 17 802 (minus12 ) 3 Jan 2011 p143r39 1 Aug 2009

p143r40Spiti Lahaul 2110 13 2257 (+7 ) 20 Oct 2011 p147r37 15 Oct 2000

p147r38Hindu Kush 793 7 724 (minus11 ) 17ndash21 Oct 2008 p150r34 15 Aug amp

p150r35 16 Sep 1999Karakoram East 5328 10 5384 (+1 ) 31 Oct 2010 p148r35 7 Sep 1998Karakoram West 5434 12 5687 (+1 ) 3 Dec 2008 p149r35 29 Aug 1998Pamir 3178 11 3210 (minus1 ) 29 Nov 2011 p152r33 16 Sep 2000

3 Data and methods

31 Digital elevation models

For each study site we used the Shuttle Radar TopographicMission (SRTM) version 4 DEM as the reference topogra-phy (Rabus et al 2003) This data set acquired in Febru-ary 2000 comes along with a mask to identify the pixelswhich have been interpolated (both were downloaded fromhttpsrtmcsicgiarorg) On each site the SRTM DEM issubtracted from a more recent DEM built from a stereo pairacquired by the HRS sensor onboard the SPOT5 satellite(Korona et al 2009) except for the Hindu Kush study sitewhere the recent DEM has been created from a stereo pairacquired by the HRG sensor also onboard SPOT5 (Berthieret al 2007) The date of each SPOT5 DEM acquisition dif-fers depending on the study site (Table 1) but 7 out of 9 wereacquired between October 2010 and November 2011 EachSPOT5-HRS DEM is also provided with a correlation maskand an ortho-image generated from the rear HRS image

SPOT5-HRS DEMs were produced by the French Map-ping Agency (IGN) using two sets of correlation parametersdefined during the SPIRIT (SPOT 5 stereoscopic survey ofPolar Ice Reference Images and Topographies) internationalpolar year project (Korona et al 2009) Thus two versionsof the DEM are delivered with their respective mask one op-timized (v1) for a gentle topography and the second one (v2)for a more rugged relief Earlier work has shown that v2 hasa slightly larger percentage of interpolated pixels but for thenon-interpolated pixels smaller errors (Korona et al 2009)This version (v2) is preferred here Given that both DEMsare calculated from the same image pair little difference isexpected between v1 and v2 However locally we identifiedsome large bull eye elevation differences between the two

versions of the DEM both off and on glaciers Comparisonwith the SRTM DEM shows that both versions of the DEMare erroneous for those areas and thus we preferred to ex-clude from further analysis all pixels for which the absoluteelevation difference between the two versions of the DEMis larger than 5 m About 20 of the valid DEM pixels areexcluded through that procedure

The SRTM DEM and mask originally at a 3 arcsecondresolution (sim 90 m) are resampled bilinearly to 40 m (UTMprojection WGS84 ellipsoid) to match the posting and pro-jection of the SPOT5 DEM Altitudes are defined above theEGM96 geoid for both DEMs

32 Glacier mask

Where available glacier outlines were downloaded fromthe GLIMS database (Raup et al 2007 Supplement) andmanually corrected if needed Otherwise the glacier maskwas derived from Landsat-TM and Landsat-ETM+ imagesPathrow and dates of Landsat acquisition are given in Ta-ble 1 for each study site A threshold varying between 05and 07 depending on the study site was applied to the Nor-malized Difference Snow Index (TM2-TM5)(TM2+TM5)to detect clean ice and snow automatically whereas debris-covered parts were digitized manually by visual interpreta-tion (eg Racoviteanu et al 2009) The total glacier areaand debris-covered part can be found for each study site inTable 1

Most inventories are derived from Landsat images withminimal snow cover from around the year 2000 prior to the2003 failure of the Scan Line Corrector of the ETM+ sensoronboard Landsat 7 that resulted in image striping Glacierareas are expected to have experienced minor changes sincethen except for glacier fronts that retreated due to the rapid

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1267

Table 2 Elevation ranges equilibrium line altitudes (ELA) and accumulation area ratios (AAR) for the nine study sites The dates ofacquisition of Landsat images used for ELA determination are also given

Sub-region Glaciers elevation ELA (m) AAR Landsat imagesrange (m)

Min Max This study Scherler et Kaumlaumlb et This study Kaumlaumlb et PathRow Dateal (2011a) al (2012) al (2012)

Hengduan Shan 2630 6860 4970plusmn 320 NA NA 55 NA p135r39 23 Sep 1999Bhutan 4390 7480 5690plusmn 440

5700 555036

47p138r40 17 Nov 2000

Everest 4260 8380 5840plusmn 320 31 p140r41 31 oct 2009West Nepal 3950 6870 5590plusmn 138 37 p143r39 1 Aug 2009Spiti Lahaul 3780 6360 5390plusmn 140 5103 5500 34 35 p147r37 2 Aug 2002

p147r38Hindu Kush 2717 6385 5050plusmn 160 5138 30 p150r34 2 Sep 2000

p150r35Karakoram east 3230 7770

5030plusmn 280 4845 554076

47p148r35 7 Sep 1998

Karakoram west 2910 7920 66 p149r35 29 Aug 1998Pamir 2800 7090 4580plusmn 250 NA NA 52 NA p152r33 30 Jul 2000

growth of connected lakes and for glaciers that surged andthus advanced For the few glaciers that advanced their frontposition was updated manually based on the SPOT5 ortho-image

For each study site we favored our user-defined glacierinventories instead of the global Randolph Glacier Inventory(RGI v20 Arendt et al 2012) because the latter is hetero-geneous and for some study sites did not match our accu-racy requirements for the adjustments and corrections of theDEMs

We also estimated the altitude of the equilibrium line(ELA) as the snowline on thesim 2000 Landsat data with min-imal snow cover to separate ablation and accumulation ar-eas The ELA is assumed to be identical to the altitude ofthe snowline at the end of the ablation season (eg Raba-tel et al 2005) after the end of the monsoons or before thefirst snowfalls Therefore we manually digitized snowlinesfor more than 30 glaciers for each study site and computedtheir mean elevations Our ELAs are slightly different fromthose given by Scherler et al (2011a) and Kaumlaumlb et al (2012)probably due to the sensitivity of the ELA to the choice ofthe image (Table 2) Ideally the ELA should have been es-timated from images acquired at the end of the ablation sea-son during various years covering our whole study periodThis would be a highly time-consuming task that was not ac-complished here However we stress that in our study theELA is only used to compare elevation changes on the abla-tionaccumulation areas (Sect 41) and to estimate the sen-sitivity of our mass balances to the choice of the volume-to-mass conversion factor (Sect 35) Our preferred mass bal-ance estimate uses a unique volume-to-mass conversion fac-tor for the whole glacier and is thus independent from theELA

33 Adjustments and corrections of DEM biases

Different DEM processing steps are followed to extract un-biased glacier elevation changes for each study site (eg Nuthand Kaumlaumlb 2011 Gardelle et al 2012b)

331 Planimetric adjustment of the DEMs

A horizontal shift between two DEMs results in an aspect-dependent bias of elevation differences (Nuth and Kaumlaumlb2011) Here the horizontal shift is determined by minimizingthe root mean square error of elevation differences (SPOT5-SRTM) off glaciers where the terrain is assumed to be stableover the study period (Berthier et al 2007) The SRTM DEMis then resampled (cubic convolution) according to the shift

332 Alongacross track corrections (drift of thesatellite orbit)

For DEMs derived from stereo imagery the satellite acquisi-tion geometry can induce a bias along andor across the trackdirection (Berthier et al 2007) We estimate the azimuth ofeach SPOT5 ground track and use it to rotate the coordinatesystem accordingly (Nuth and Kaumlaumlb 2011) Then we com-pute the elevation differences along and across the satellitetrack on stable areas and when necessary correct the bias us-ing a 5th order polynomial fit

333 Curvature correction

As suggested by Paul (2008) and Gardelle et al (2012b) thedifference in original spatial resolutions of the two DEMs canlead to biases related to altitude in mountainous areas Whenthe curvature the first derivative of the slope is high (sharppeaks or ridges) the altitude tends to be underestimatedby the coarse DEM unable to reproduce high-frequency

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1268 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

slope variations Thus this apparent ldquoelevation biasrdquo is cor-rected using the relation between elevation differences andthe maximum terrain curvature estimated over stable areasoff glaciers (Gardelle et al 2012b) Note that the units inFig 1b in Gardelle et al (2012b) should be 10minus2 mminus1 in-stead of mminus1 (A Gardner personal communication 2012)However this does not impact the validity of the correction

334 Radar penetration correction

The SRTM DEM acquired in C-band potentially underesti-mates glacier elevations since at this wavelength (sim 56 cm)the penetration of the radar signal into snow and ice canreach several meters (eg Rignot et al 2001) Here we esti-mate this penetration for each study site by differencing theSRTM C-band (57 GHz) and X-band (97 GHz) DEMs Thelatter was acquired simultaneously with the SRTM X-bandat higher resolution (30 m) but with a narrower swath andthus has a coverage restricted to selected swaths only Sincethe penetration of the X-band is expected to be low com-pared to the C-band (an hypothesis that still needs to be con-firmed Ulaby et al 1986) we consider the elevation differ-ence (SRTMC-bandminus SRTMX-band) to be a first approxima-tion of C-band radar penetration over glaciers (Gardelle etal 2012b) The correction is calculated as a function of alti-tude and applied to each pixel separately

For the Hindu Kush Spiti Lahaul and West Nepal sitesdue to the lack of SRTMX-band data we were unable to de-termine the value of the penetration Thus we applied thecorrection computed over the nearest study site ie Karako-ram for Spiti Lahaul and Hindu Kush and Everest for WestNepal The mean values of the SRTMC-bandpenetrations overglaciers are given for each study site in Table 3

335 Seasonality correction

SPOT5 DEMs were acquired between late October and earlyJanuary depending on the study site Since the SRTM DEMis from mid-February we need to account for possible masschanges during this 1-to-5-month period in order to esti-mate glacier mass balance over an integer number of years(Gardelle et al 2012a) For the eastern sites (from WestNepal to Hengduan Shan) where glaciers are of summer-accumulation types (Fujita 2008) the correction was set to0 This choice is confirmed by recent field observations thatshow no winter accumulation on Mera and Pokalde glaciersin Nepal (Wagnon et al 2013) For the Karakoram studysites we used the winter accumulation rate+013 m we permonth measured on Biafo Glacier (Wake 1989) For theother western study sites (the Pamir Hindu Kush and SpitiLahaul sites) the correction is derived from the mean of win-ter mass balances of 35 glaciers all in the Northern Hemi-sphere+089 m we yrminus1 over 2000ndash2005 corresponding to+015 m we per winter month (Ohmura 2011) The lattervalue is slightly lower than the average winter mass balance

Table 3 Average SRTM C-band penetration estimates (in m)in February 2000 over glaciers in this study and from Kaumlaumlb etal (2012)

Sub-region This study Kaumlaumlb et al (2012)

Hengduan Shan 17 NABhutan 24

25plusmn 05Everest 14West Nepal NA

15plusmn 04Spiti Lahaul NAHindu Kush NA 24plusmn 04Karakoram East

34 24plusmn 03Karakoram WestPamir 18 NA

(+135plusmn 031 m we yrminus1) measured on Abramov Glacier inthe Pamir during 1968ndash1998 (WGMS 2012) and in reason-ably good agreement with the modeled mean winter massbalance (+108plusmn 034 m we yrminus1) of Chhota Shigri Glacierduring 1969ndash2012 in the western Himalaya (Azam et al2013)

336 Mean elevation changes and mass balancecalculation

Before averaging elevation changes we exclude all interpo-lated pixels in the SRTM or SPOT5 DEMs as well as unex-pected elevation changes exceedingplusmn150 m for study sites(Pamir and Karakoram) that include surge-type glaciers orplusmn80 m for other sites Those thresholds were chosen aftercareful inspection of the elevation difference maps and his-tograms in order to identify the largest realistic glacier eleva-tion changes Glaciers that are truncated at the edges of theDEM are also excluded from mass balance calculations asthey may bias the final results if for example only their accu-mulation or ablation area is sampled Although they are trun-cated Siachen (Karakoram East) and Fedtchenko (Pamir)glaciers were retained because of their wide coverage andbecause their accumulation and ablation areas were correctlysampled in the DEMs

Elevation changes are then analyzed for 100 m altitudebins Within each altitude band we average the elevationchanges after excluding pixels where absolute elevation dif-ferences differ by more than three standard deviations fromthe mean (Berthier et al 2004 Gardner et al 2012) Thisis an efficient way to exclude outliers based on the assump-tion that elevation changes should be similar at a given alti-tude within each study site This assumption does not holdfor regions containing actively surging glaciers Thus forthe Pamir and Karakoram study sites surge-type glaciersare treated separately ie elevation changes are averaged by100 m altitude bands for each surge-type glacier individually

Where no elevation change is available for a pixel weassign to it the value of the mean elevation change of the

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1269

altitude band it belongs to in order to assess the mass bal-ance over the whole glacier area

The conversion of elevation changes to mass balance re-quires knowledge of the density of the material that has beenlost or gained and its evolution during the study period Giventhe lack of repeat measurement of density profiles over theentire snowfirnice column in the PKH we applied a con-stant density conversion factor of 850 kg mminus3 for all ourstudy sites (Sapiano et al 1998 Huss 2013)

The mass balance of each surge-type glacier is computedseparately and added (area-weighted) to the mass balance ofthe rest of the glaciers to estimate the mass balance of thewhole study site

For study sites with a high concentration of growingpro-glacial lakes such as Bhutan Everest and West Nepal(Gardelle et al 2011) our mass losses are slightly underes-timated because they do not take into account the glacier icethat has been replaced by water during the expansion of thelake Subaqueous ice losses are only estimated for LhotseSharImja Glacier (Everest area Fujita et al 2009) for thesake of comparison to previous studious

The region-wide mass balance of PKH glaciers as definedin Sect 2 is computed by extrapolating the mass balancefrom each study site to a wider sub-region (Fig 1) that is as-sumed to experience the same glacier mass changes becauseof its similar climate (Bolch et al 2012 Kaumlaumlb et al 2012)The total glacier area for each sub-region is computed fromthe RGI v20 (Arendt et al 2012)

The periods over which we calculate mass balance areslightly different from one site to another (Table 1) depend-ing on the year of acquisition of the SPOT5 DEM (two in2008 two in 2010 and five in 2011) In the following we willassume that all mass balances are representative of the period1999ndash2011 in order to estimate a region-wide mass budgetof PKH glaciers neglecting thus part of the inter-annual vari-ability

34 Accuracy assessment

The errorE1hi to be expected for a pixel elevation change

1hi is assumed to equal the standard deviationσ1h of themean elevation change1h of the altitude band it belongs toThe value ofσ1h can range fromplusmn4 m toplusmn 20 m dependingon the study site and elevation This is rather conservative asthe value ofσ1h contains also the intrinsic natural variabilityof elevation changes within the altitude band (ie it containsboth noise and real geophysical signal)

The errorE1h of the mean elevation change1h in eachaltitude band is then calculated according to standard princi-ples of error propagation

E1h =E1hiradicNeff

(1)

Neff represents the number of independent measurements inthe altitude bandNeff will be lower thanNtot the total num-

ber of elevation change measurements1hi in the altitudeband since the latter are correlated spatially (Bretherton etal 1999)

Neff =Ntot middot PS

2d (2)

where PS is the pixel size (here 40 m)d is the distance ofspatial autocorrelation determined using Moranrsquos I autocor-relation index on elevation differences off glaciers On theaverage over the 9 study sitesd = 492plusmn 72 m

The error on the penetration correction was assumed to besystematic due to the not fully understood physics involvedand equal toplusmn 15 m This error was obtained by compar-ing the SRTM C-band penetration estimated using two inde-pendent methods (i) SRTMC-bandndash SRTMX-band and (ii) dif-ference between ICESat and SRTM when ICESat elevationchange trends during 2003ndash2008 are extrapolated to Febru-ary 2000 (Kaumlaumlb et al 2012) The maximum difference of thepenetration inferred from the two methods is 10 m (Table 3)A 50 additional error margin was added to account for thefact that this comparison could only be made on 2 of ourstudy sites

Given the slender observational support for the seasonal-ity correction (Section 33) we assumed its uncertainty tobe plusmn100 on the western sites ieplusmn015 m we per win-ter month The same absolute uncertainty (plusmn015 m we perwinter month) was used for the eastern sites where no sea-sonality correction is applied but ablation andor accumula-tion may occur in winter (Wagnon et al 2013)

All errors are summed quadratically within each altitudeband and then summed for all altitude bands assuming con-servatively that they are 100 dependent The error on theSRTM penetration correction clearly dominates our errorbudget for volume change

During the conversion from volume to mass we assumea plusmn60 kg mminus3 error on the density conversion factor (Huss2013) Thisplusmn7 error is similar to the one estimated byBolch et al (2011) In a sensitivity test the volume to massconversion is also performed using two alternative densityscenarios (Kaumlaumlb et al 2012) (i) a density of 900 kg mminus3 isassumed everywhere and (ii) 600 kg mminus3 in the accumula-tion area and 900 kg mminus3 in the ablation area These two sce-narios take among other things into account that elevationchanges could be due to changes in glacier dynamics or dueto surface mass balance We then calculate the maximum ab-solute difference between the mass balances calculated usingour preferred scenario (850plusmn 60 kg mminus3 everywhere) and themass balances calculated using the two others scenarios Forthe eastern sites (West Nepal to Hengduan Shan) the max-imum absolute difference is 003 m we yrminus1 For the west-ern sites (Pamir to Spiti Lahaul) it reaches 006 m we yrminus1Those uncertainties due to the choice of a given density sce-nario remain negligible compared to other sources of errors

We also include an error due to extrapolation of our massbalance estimate from our study site (the extent of each

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1270 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

SPOT5 DEM) to the unmeasured glaciers within each sub-region To estimate this regional extrapolation error we com-pare the ICESat elevation change trends (computed as inKaumlaumlb et al 2012) within 3 times 3 cells centered at the studysites (Table 5) with the ICESat-derived trends for the entiresub-regions The absolute difference between both trends isless than 002ndash003 m yrminus1 for Bhutan West Nepal Spiti La-haul and Karakoram ie negligible For Everest the eleva-tion trend for the entire sub-region is 012 m yrminus1 less nega-tive than for the 3 times 3 cells around the SPOT5 DEM cen-ter for Hindu Kush the trend for the entire sub-region is017 m yrminus1 more negative than for the 3 times 3 cell due to thevariability of elevation changes within this sub-region (Kaumlaumlbet al 2012) The regional extrapolation error is obtainedas the area-weighted mean absolute difference between themass balances derived from ICESat for those 3 times 3 cellsand the whole sub-region and equalsplusmn004 m we yrminus1

Finally for sub-regions and total PKH estimates of masschange we include an error on the glacier area computedfrom the RGI v20 (Arendt et al 2012) The inventory er-ror is specific to each study site and evaluated based on thecomparison of our user-defined inventory on each study sitewith the RGI v20 (Table 1) We note the remarkable accu-racy of the RGI for all but one of our study sites The rel-ative errors are generally of a few percents and up to 12 for West Nepal The differences between our and existing in-ventories are probably due to the difficulty of delimitatingdebris-covered glacier parts (eg Frey et al 2012 Paul etal 2013) and accumulation areas The Hengduan Shan studysite is an exception There the RGI is based on the ChineseGlacier Inventory (Shi et al 2009 Arendt et al 2012) andoverestimates the ice-covered area by 88 compared to ourLandsat-based inventory

Unless stated otherwise all error bars from this study andpreviously published studies correspond to one standard er-ror

4 Results

41 Overview of the elevation changes

The maps of glacier elevation changes are given in Figs 2ndash10for our nine study sites with as inset the distribution ofthe elevation differences off glaciers The full maps of ele-vation differences off glaciers are shown in the Supplement(Figs S2ndashS10) Those off glacier maps indicate a local ran-dom noise of aboutplusmn5ndash10 m However at length scales of afew kilometers the spatial variations in elevation differencesare small typically less than 1 m

For the four eastern study sites from the HengduanShan to the West Nepal (Fig 1) the elevation changesaveraged over the accumulation areas are small (between003plusmn 014 m yrminus1 to 013plusmn 015 m yrminus1) whereas clear sur-face lowering is observed for all four study sites in their

ablation areas ranging fromminus050plusmn 014 m yrminus1 (Bhutan)to minus081plusmn 013 m yrminus1 (Hengduan Shan) Over the BhutanEverest and West Nepal study sites glaciers in contact witha proglacial lake often experienced larger thinning rates

Further west the Spiti Lahaul is the only site wheresurface lowering is measured both in the accumula-tion area (minus034plusmn 016 m yrminus1) and the ablation area(minus070plusmn 014 m yrminus1) Clear surface lowering is also ob-served in the ablation areas of the Hindu Kush glaciers(minus050plusmn 018 m yrminus1)

In the Karakoram (east and west Figs 8 and 9) andin the Pamir (Fig 10) the pattern of elevation changesis highly heterogeneous due to the occurrence of numer-ous glacier surges or similar flow instabilities The eleva-tion changes in the accumulation areas of those three studysites are slightly positive (between+022plusmn 014 m yrminus1 and+030plusmn 018 m yrminus1) In the ablation areas small surfacelowering is observed for the two Karakoram study sites (av-erage of the two sitesminus033plusmn 016 m yrminus1) whereas no el-evation change is found in Pamir (minus002plusmn 013 m yrminus1)

Note that those mean rates of elevation changes are sensi-tive to the choice of the ELA assumed to equal the snowlinealtitude in Landsat images from around year 2000

42 Influence of a debris cover

To evaluate thinning rates over debris-covered and debris-free ice we perform a histogram adjustment in order tocompare data sets with similar altitude distributions This isneeded because debris-covered parts tend to be concentratedat lower elevations The adjustment consists in randomly ex-cluding pixels over debris-free ice from each elevation bandto match a scaled version of the debris-covered pixel distribu-tion (Kaumlaumlb et al 2012) Elevation changes with altitude canthen be compared for clean and debris-covered ice (Fig 11)

The thinning is higher for debris-covered ice in the Ever-est area and on the lowermost parts of glaciers in the PamirBut on average in the Pamir western Karakoram and SpitiLahaul the lowering of debris-free and debris-covered ice issimilar The thinning is stronger over clean ice in the Heng-duan Shan Bhutan West Nepal Hindu Kush and althoughto a lesser extent in the eastern Karakoram

43 Mass balance over the nine study sites

Glacier mass changes have been heterogeneous overthe PKH since 1999 Slight mass gains or balancedmass budgets are found in the north-west in thePamir (+014plusmn 013 m we yrminus1) and for the two studysites in the central Karakoram (+009plusmn 018 m we yrminus1

and +011plusmn 014 m we yrminus1) Mass loss is moderatein the adjacent Hindu Kush with a mass balance ofminus012plusmn 016 m we yrminus1 Glaciers in all our eastern studysites (two in Nepal Bhutan and Hengduan Shan) experi-enced homogeneous mass losses with mass balances ranging

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1271

Fig 2 Glacier elevation changes over the Hengduan Shan study site Inset distribution of the elevation differences off glacier The medianand the standard deviation of the elevation difference and the number of pixels off glaciers are given (by construction the mean elevationdifference is null)

Fig 3Same as Fig 2 but for the Bhutan study site

from minus033plusmn 014 m we yrminus1 to minus022plusmn 012 m we yrminus1The most negative mass balance is measured in Spiti Lahaul(western Himalaya) atminus045plusmn 013 m we yrminus1 (Table 4)

However within a study site mass balances can be highlyvariable from one glacier to another In the Everest area twolarge glaciers (71 km2 each) experienced distinctly differentmass balance withminus016plusmn 016 m we yrminus1 for the south-flowing Ngozumpa Glacier andminus059plusmn 016 m we yrminus1

for the north-flowing Rongbuk Glacier (Fig 4) In Bhutan(Fig 3) mass loss is higher south (minus025plusmn 013 m we yrminus1)

than north (minus014plusmn 012 m we yrminus1) of the main Hi-malayan ridge though not at a statistically significant level

The region-wide mass budget of PKH glaciers isnegative minus101plusmn 55 Gt yrminus1 (minus014plusmn 008 m we yrminus1)which corresponds to a sea level rise contribution of0028plusmn 0015 mm yrminus1 over 1999ndash2011

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1272 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 4Same as Fig 2 but for the Everest study site

Fig 5Same as Fig 2 but for the West Nepal study site

44 Glacier surges

In the Pamir and the Karakoram the spatial pattern of ele-vation changes of many glaciers is typical of surge events orflow instabilities They can be identified because of their veryhigh thinning and thickening rates that can both reach up to16 m yrminus1 (Figs 8ndash10)

Most of them have already been reported to be surge-type glaciers based on their velocity (Kotlyakov et al 2008Quincey et al 2011 Copland et al 2009 Heid and Kaumlaumlb2012) morphology (Copland et al 2011 Barrand and Mur-ray 2006 Hewitt 2007) or elevation changes (Gardelle etal 2012a) Here we only identify surge-type glaciers for thesake of mass balance calculation and we do not attempt toprovide an exhaustive up-to-date surge inventory This is why

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1273

Fig 6 Same as Fig 2 but for the Spiti Lahaul study site The map of elevation changes contains many voids due to the presence ofthin clouds during the acquisition of the SPOT5 DEM The mass balance of the Chhota Shigri and Bara Shigri glaciers are respectivelyminus039plusmn 018 m we yrminus1 andminus048plusmn 018 m we yrminus1

the published surge inventories in the Pamir and the Karako-ram do not necessarily match the surge-type glaciers labeledon the elevation change maps as these refer only to our ob-servation period (1999ndash2011)

The mass balance of surge-type glaciers (respectivelynon-surge-type glaciers) is +019plusmn 022 m we yrminus1

(+012plusmn 011 m we yrminus1) in the Pamir+009plusmn 030 m we yrminus1 (+009plusmn 015 m we yrminus1) inthe western Karakoram and+010plusmn 025 m we yrminus1

(+012plusmn 012 m we yrminus1) in the eastern Karakoram Theoverall similarity between the mass budgets for surge-typeand non-surge-type glaciers shows that those flow instabili-ties seem not to affect the glacier-wide mass balance at leastover short time periods here 10 yr

5 Discussion

51 SRTM penetration

Alternative estimates of SRTMC-bandpenetration for the PKHregion were obtained by Kaumlaumlb et al (2012) by comparing el-evations of the ICESat altimeter (Zwally et al 2002) with theSRTMC-bandDEM in PKH and extrapolating the 2003ndash2008trend of elevation changes back to the acquisition date ofSRTM The difference between east and west is more dis-tinct in our case with a mean penetration in the Karakoram of34 m (24plusmn 03 m in Kaumlaumlb et al 2012) 24 m in Bhutan and14 m around the Everest area (25plusmn 05 m for a wider areaincluding east Nepal and Bhutan in Kaumlaumlb et al 2012) Thiseastwest gradient is expected given that in February 2000when the SRTM mission was flown the snowfirn thicknesswas probably larger in the western PKH where glaciers areof winter-accumulation type than in the eastern PKH where

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1274 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 7Same as Fig 2 but for the Hindu Kush study site

glaciers are of summer-accumulation type Further studieswould be necessary to investigate the possible reasons be-hind the relatively low SRTMC-bandpenetration for the Pamir(18 m) that do not follow the eastwest gradient mentionedabove

Both our and Kaumlaumlb et alrsquos (2012) estimates agree withintheir error bars and seem thus to provide robust first-orderestimates for the SRTMC-bandradar penetration in the regionHowever it remains to be verified if a penetration depth com-puted at the regional scale is appropriate for a single smallglacier whose altitude distribution is different from the oneof whole study site (eg Abramov Glacier in the far north ofour Pamir study site)

52 Thinning over debris-covered ice

PKH glaciers are heavily debris-covered in their ablation ar-eas because of steep rock walls that surround them and in-tense avalanche activity Twelve percent (12 ) of the glacierarea in our study sites are covered with debris (up to 22 inthe Everest area Table 1) which is in agreement with pre-vious estimates for the Himalaya and the Karakoram 13 in Kaumlaumlb et al (2012) and 10 in Bolch et al (2012) Thedebris layer is expected to modify the ablation of the under-

lying ice It will increase ablation if its thickness is just a fewcentimeters (dominance of the albedo decrease) or decreaseablation it if the layer is thick enough to protect the ice fromsolar radiation (Mattson et al 1993 Mihalcea et al 2006Benn et al 2012 Lejeune et al 2013)

However recent studies have reported similar overall thin-ning rates over debris-covered and debris-free ice in thePKH (Kaumlaumlb et al 2012 Gardelle et al 2012a) or evenhigher lowering rates over debris-covered parts in the Ever-est area (Nuimura et al 2012) Our findings are consistentwith Nuimura et al (2012) for the Everest study site with astronger thinning (rate of elevation change ofminus097 m yrminus1)

in debris-covered areas than over clean ice (rate of elevationchange ofminus057 m yrminus1) In the Pamir Karakoram and SpitiLahaul study sites these rates are similar while in the HinduKush West Nepal Bhutan and Hengduan Shan study sitesthinning is greater over clean ice in agreement with the com-monly assumed protective effect of debris (Fig 4) Thosedifferences between our 9 study sites suggest a complex re-lationship between debris cover and glacier wastage

As local glacier thickness changes are a result of bothmass balance processes and a dynamic component theseunexpected high thinning rates over debris-covered parts

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1275

Fig 8Same as Fig 2 but for the Karakoram east study site Surge-type glaciers in an active surge phase (resp quiescent phase) are identifiedwith a solid triangle (resp solid circle)

are potentially the result of glacier-wide processes Over aglacier tongue the ablation can be enhanced by the pres-ence of steep exposed ice cliffs or supraglacial lakes andponds (Sakai et al 2000 2002) In addition it is also pos-sible that the glacier tongues are mostly covered by thindebris layers such that the albedo effect dominates the in-sulating effect resulting in an increased tongue-wide abla-tion The prevalence of thin debris was recently suggestedfor debris-covered glaciers around the Mount Gongga in thesouth-eastern Tibetan Plateau (Zhang et al 2013b) Finallyice-flow rates could be very low at debris-covered parts asshown by Quincey et al (2007) in the Everest area Kaumlaumlb(2005) for Bhutan and Scherler et al (2011b) in the HinduKush-Karakoram-Himalaya Thus all three factors are likelyto increase the overall thinning under debris despite the un-doubted protective effect of a continuous thick debris coverat a local scale Future investigations combining surface ve-locity fields and maps of elevation changes could allow eval-uating more closely the relative role of ablation and ice fluxesin thinning rates over PKH glacier tongues (eg Berthierand Vincent 2012 Nuth et al 2012) Measurements of the

spatial distribution of debris thickness over the PKH glaciertongues are also needed

Note that in the case of debris-covered tongues the ele-vation change measurement represents the glacier thicknesschange but also the possible debris thickness evolution Thelatter remains poorly known in the PKH or in any othermountain range But based on the debris discharges givenby Bishop et al (1995) for Batura Glacier (Pakistan) 48to 90times 103 m3 yrminus1 the thickening of the debris can be es-timated between 27 and 55 mm yrminus1 which is negligiblegiven the magnitude of the glacier elevation changes

53 Comparison to previous mass balance estimates

Throughout the PKH there is only a single peer-reviewedglaciological record (on Chhota Shigri Glacier) coveringmost of our study period (Wagnon et al 2007 Vincent etal 2013) Our geodetic mass balance estimate for the SpitiLahaul study site has been recently compared to those glacio-logical measurements on Chhota Shigri Glacier (Vincent etal 2013) Mass balance records reported in Yao et al (2012)for the Tibetan Plateau and the surroundings mountain ranges

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1276 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 9Same as Fig 8 but for the Karakoram west study site

only start during glaciological year 20052006 This is whyhere we do not compare our mass balance estimates toglaciological records and limit our comparison to others re-gional estimates obtained using the geodetic or gravimetricmethods

We stress that the comparison of our mass balance mea-surements to published estimates is complicated by thefact that generally they do not cover the same time pe-riod Little is known about the inter-annual variability inthe PKH during the 21st century due to the limited num-ber of long glaciological time series The 9 yr mass balance

record on Chhota Shigri Glacier reveals a relatively highinter-annual variability of the annual mass balance (stan-dard deviationplusmn 074 m we yrminus1 during 2003ndash2011 Vin-cent et al 2013) which is similar to the one obtained be-tween 1968 and 1998 on Abramov Glacier (standard devi-ation 069 m we yrminus1 WGMS 2012) Thus inter-annualvariability can explain part of the difference between two cu-mulative mass balance estimates over 5ndash10 yr even if they areoffset by only one or two years

Based on ICESat repeat track measurements and theSRTM DEM Kaumlaumlb et al (2012) measured a region-wide

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1277

Fig 10Same as Fig 8 but for the Pamir study site

mass balance ofminus021plusmn 005 m we yrminus1 between 2003and 2008 for Hindu Kush-Karakoram-Himalaya glaciersFor that same area (ie excluding the Pamir and theHengduan Shan study sites) we found a mass balance ofminus015plusmn 007 m we yrminus1 Thus our result agrees with Kaumlaumlbet al (2012) within the error bars although our study periodis longer (1999ndash2011) and our measurement principle andextrapolation approach are different In addition we havecalculated updated mass balances from Kaumlaumlb et al (2012)ie averages over 3times 3 degree cells centered over our studysites and find that they are consistent within error barswith our mass balance estimates (Table 5) Similar resultsare found when our mass balances are compared to a spa-tially extended ICESat analysis for 2003ndash2009 (Gardner etal 2013) Mass losses measured with ICESat (Kaumlaumlb et al2012 Gardner et al 2013) tend to be slightly larger thanours This could be due to the difference in time periodsAlso our smaller mass losses could be partly explained if wesystematically underestimated the poorly constrained pene-

tration of the C-band andor X-band radar signal into ice andsnow

At a more local scale Bolch et al (2011) reported a massbalance ofminus079plusmn 052 m yrminus1 between 2002 and 2007 byDEM differencing over a glacier area of 62 km2 includingten glaciers south and west of Mt Everest For these sameglaciers Nuimura et al (2012) found a mean mass balanceof minus045plusmn 060 m yrminus1 during 2000ndash2008 while they com-puted a mass balance ofminus040plusmn 025 m yrminus1 between 1992and 2008 over a glacier area of 183 km2 around Mt EverestIn the present study we found an average mass balance ofminus041plusmn 021 m yrminus1 over the ten glaciers mentioned above(Table 5 Fig S1) and a value ofminus026plusmn 013 m yrminus1 overthe whole Everest study site between 2000 and 2011 Ourvalues are comparable to those of Nuimura et al (2012)Our mass balance is less negative than the 2002ndash2007 massbalance of Bolch et al (2011) but not statistically differentwhen error bars are considered (Fig 12) Indeed Bolch etal (2011) acknowledged some high uncertainties for their

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1278 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Table 4Glacier mass budget of the nine sub-regions to which the results of the study sites (mass balances) have been extrapolated For eachof them the total glacierized area as well as the percentage of the glacier area over which the mass balance was computed (ie study sites)are given The total glacier area is derived from the RGI v20 (Arendt et al 2012) The error for the mass balance in each sub-region is theroot of sum of squares (RSS) of the mass balance error of each study site and the regional extrapolation error The error for the mass budgetin each sub-region includes additionally the error from the total glacier area in the RGI v20

Sub-region Total glacier Measured glacier Mass balance Mass budgetarea (km2) area () (m we yrminus1) (Gt yrminus1)

Hengduan Shan 11584 12 minus033plusmn 014 minus38plusmn 38Bhutan 4021 34 minus022plusmn 013 minus09plusmn 05Everest 6226 23 minus026plusmn 014 minus16plusmn 08West Nepal 6849 13 minus032plusmn 014 minus22plusmn 10Spiti Lahaul 9043 23 minus045plusmn 014 minus41plusmn 13Hindu Kush 6135 13 minus012plusmn 016 minus07plusmn 10Karakoram East

1902428 +011 plusmn 014 +19 plusmn 31

Karakoram West 30 +009 plusmn 018Pamir 9369 34 +014 plusmn 014 +13 plusmn 13

TotalArea- 72251 30 minus014plusmn 008 minus101plusmn 55weighted average

Table 5Comparison of geodetic mass balances between this study and previous published results with overlapping study periods and similargeographic locations Updated figures from Kaumlaumlb et al (2012) are averaged over 3 times 3 cells centered over the study sites of the presentstudy

GlacierSite name This study Bolch et al (2011) Nuimura et al (2012) Kaumlaumlb et al (2012 updated)2002ndash2007 2000ndash2008 2003ndash2008

Study Mass balance Area (km2)a Mass balance Area Mass balance Mass balanceperiod (m we yrminus1) (m we yrminus1) (km2) (m we yrminus1) (m we yrminus1)

Hengduan San 1999ndash2010 minus022 plusmn 014 1303 (55 )Bhutan 1999ndash2010 minus022 plusmn 012 1384 (64 ) minus052 plusmn 016b

Everest

1999ndash2011

minus026 plusmn 013 1461 (58 ) minus039 plusmn 011AX010 minus090plusmn 034 04Changri SharNup minus042plusmn 017 161 (79 ) minus029plusmn 052 130 minus055plusmn 038Khumbu minus051plusmn 019 203 (47 ) minus045plusmn 052 170 minus076plusmn 052Nuptse minus037plusmn 020 50 (74 ) minus040plusmn 053 40 minus034plusmn 027Lhotse Nup minus021plusmn 027 24 (70 ) minus103plusmn 051 19 minus022plusmn 047Lhotse minus043plusmn 018 85 (83 ) minus110plusmn 052 65 minus067plusmn 051Lhotse SharImja minus070plusmn 052 98 (44 ) minus145plusmn 052 107 minus093plusmn 060Amphu Laptse minus046plusmn 034 25 (38 ) minus077plusmn 052 15 minus018plusmn 094Chukhung +044 plusmn 024 42 (47 ) +004 plusmn 054 38 +043 plusmn 081Ama Dablam minus049plusmn 017 36 (54 ) minus056plusmn 052 22 minus056plusmn 073Duwo minus016plusmn 026 19 (18 ) minus196plusmn 053 10 minus068plusmn 074Total Khumbu minus041plusmn 021 744 minus079plusmn 052 617 minus045plusmn 060(10 Glaciers above)

West Nepal 1999ndash2011 minus032 plusmn 013 908 (40 ) minus032 plusmn 012Spiti Lahaul 1999ndash2011 minus045 plusmn 013 2110 (46 ) minus038 plusmn 006Karakoram East 1999ndash2010 +011 plusmn 014 5328 (42 ) minus004 plusmn 004Karakoram West 1999ndash2008 +009 plusmn 018 5434 (45 )Hindu Kush 1999ndash2008 minus012 plusmn 016 793 (80 ) minus020 plusmn 006Pamir 1999ndash2011 +014 plusmn 013 3178 (50 )

a The total glacier area is given in km2 and in parenthesis the of the glacier area actually covered with measurementsb For this cell the ICESat coverage is insufficient and does not sample all glacier elevations

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1279

Fig 11 Elevation changes (SPOT5-SRTM) in the ablation area for clean ice (blue circles) and debris-covered ice (black circles) for eachstudy site For the Karakoram (east and west) and the Pamir study sites only non-surge-type glaciers are considered Standard error of theelevation differences are typically less thanplusmn2 m with each 100 m elevation band (not shown for the sake of clarity) Note elevation changesover separate sections of a glacier cannot be treated as mass changes due to the disregard of glacier dynamics

estimate over this short time period The differences in sur-vey periods and in the glacier outlines may also explainpart of these discrepancies In particular the delimitation ofthe accumulation area is notoriously difficult and Bolch etal (2011) did not include some of the steep high altitudeslopes which we included due to their potential avalanchecontribution For the 10 glaciers mentioned above our massbalances would become 012 m we yrminus1 more negative ifwe used the exact same outlines as Bolch et al (2011)Our positive mass balance for the small Chukhung Glacier(+044plusmn 024 m we yrminus1) is in agreement with Nuimura etal (2012) but enigmatic in a context of regional glacier lossand deserves further investigation If the 2002ndash2007 massbalance estimate by Bolch et al (2011) is excluded all otherrecent mass balance estimates suggest no acceleration of themass loss in the Everest area when compared to the long-term(1970ndash2007) mass balance measured by Bolch et al (2011)in this regionminus032plusmn 008 m weyrminus1

In Bhutan our results indicate a stronger mass loss forsouthern glaciers than for northern ones Interestingly in thisarea Kaumlaumlb (2005) measured high speeds (up to 200 m yrminus1)

for glaciers north of the Himalayan main ridge and ratherlow velocities over southern glacier tongues by using repeat

ASTER data This is consistent with the more negative massbalance that we report for the southern glaciers in Bhutanie for the slow flowing presumably downwasting glaciers

Berthier et al (2007) reported a mass balance betweenminus069 andminus085 m we yrminus1 for an ice-covered area of915 km2 around Chhota Shigri Glacier between 1999 (SRTMDEM) and 2004 (SPOT5-HRG DEM) For the same areawe found a mass balance ofminus045plusmn 013 m we yrminus1 over1999ndash2011 The difference in the study period may explainpart of the differences Also the methodology by Berthieret al (2007) did not explicitly take into account the SRTMradar penetration over glaciers and also corrected an appar-ent elevation-dependent bias according to glacier elevationswhich is now known as a resolution issue and is best cor-rected according to the terrain maximum curvature (Gardelleet al 2012b) More details and discussions about these dif-ferences can be found in Vincent et al (2013)

Importantly the results of the eastern Karakoram studysite confirm the balanced or slightly positive mass budgetreported by Gardelle et al (2012a) in the western Karako-ram over the past decade The two Karakoram study sitesare adjacent with a small overlapping area (sim 1100 km2 ofglaciers) where elevation differences agree within 1 m Both

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1280 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 12 Comparison of geodetic mass balances for 10 glaciers inthe Mount Everest area 1999ndash2010 mass balances measured inthe present study are compared with published estimates during2002ndash2007 (Bolch et al 2011) and during 2000ndash2008 (Nuimuraet al 2012) The 1-to-1 line is drawn The mean difference (plusmn1standard deviation) between (i) our values and Bolch et al (2011)are (047plusmn 058 m we yrminus1) and (ii) our values and Nuimura etal (2012) are (012plusmn 021 m we yrminus1) Part of the differences be-tween estimates may be due to the different periods surveyed andthe different glacier outlines used

sites show similar mass balances over slightly different pe-riods +011plusmn 014 m yrminus1 over 1999ndash2010 in the east and+009plusmn 018 m yrminus1 over 1999ndash2008 in the west This con-firms a ldquoKarakoram anomalyrdquo that was first suggested basedon field observations (Hewitt 2005) We report a mass bal-ance of+010plusmn 013 m yrminus1 for Baltoro Glacier (see loca-tion in Fig 8 size= 304 km2) between 1999 and 2010which is consistent with its gradual speed-up reported since2003 (Quincey et al 2009) However given the completelydifferent methods used in our and the latter study we can-not conclude that this demonstrates a real shift from nega-tive to positive mass balance The mass budget of SiachenGlacier (1145 km2 sometimes referred to as the highestbattlefield in the world) is also balanced or slightly pos-itive (+014plusmn 014 m we yrminus1) Thus the alarming state-ment by Rasul (2012) that this glacier retreated by 59 kmand lost 17 of its mass during 1989ndash2009 is very likelyerroneous Again within error bars our regional mass bal-ance in the Karakoram agrees with the mass balance ofminus003plusmn 004 m yrminus1 found by Kaumlaumlb et al (2012) over 2003ndash2008 and with the mass balance ofminus010plusmn 018 m we yrminus1

(2-sigma error level) obtained by Gardner et al (2013) for aregion including both the Karakoram and the Hindu Kush

In the northernmost part of the Pamir in the Alay moun-tain range the mass balance of Abramov Glacier has beenmeasured in the field for 30 yr between 1968 and 1998(WGMS 2012) with a mean value ofminus046 m yrminus1 Herewe report a mass balance ofminus003plusmn 014 m yrminus1 for this

glacier over 1999ndash2011 Our near zero mass budget estimatefor this glacier disagrees with negative mass balances recon-structed for the same period with a model run using biascorrected NCEPNCAR re-analysis data and calibrated withfield data (Barundun et al 2013) An underestimate of ourSRTM C-Band penetration in the Pamir as a whole and inparticular for the small Abramov Glacier may contribute tothese differences

For the whole Pamir study site our mass budget is bal-anced or slightly positive (+014plusmn 013 m we yrminus1) In theeastern Pamir the mass balance of Muztag Ata Glacier was+025 m we yrminus1 between 2005 and 2010 (Yao et al 2012)Although Muztag Ata Glacier is located outside of our Pamirstudy site and his glaciological records only cover the sec-ond half of our study period this measurement is consistentwith our remotely sensed mass balance Thus the ldquoKarako-ram anomalyrdquo should probably be renamed the ldquoPamir-Karakoram anomalyrdquo

This slightly positive or balanced mass budget measuredin the Pamir seems to contradict previous findings by Heidand Kaumlaumlb (2012) They reported rapidly decreasing glacierspeeds in this region between two snapshots of annual ve-locities in 20002001 and 20092010 an observation thatwould fit with negative mass balances (Span and Kuhn 2003Azam et al 2012) We note though that the relationship be-tween glacier mass balance and ice fluxes is not straightfor-ward and might in particular involve a time delay betweena change in mass balance and the related dynamic response(Heid and Kaumlaumlb 2012) Thus the decrease in speed couldwell be due to negative mass balances before or partiallybefore 1999ndash2011 We acknowledge that this discrepancyneeds to be investigated further in particular by examiningcontinuous changes in annual glacier speed through the firstdecade of the 21st Century Indeed Heid and Kaumlaumlb (2012)measured differences between two annual periods which maynot represent mean decadal velocity changes

Jacob et al (2012) using satellite gravimetry (GRACE)measured a mass budget ofminus5plusmn 6 Gt yrminus1 over 2003ndash2010for the ldquoKarakoram-Himalayardquo their region 8c which cor-responds to our study area excluding the Pamir Karakoramand Hindu Kush sub-regions but including some glaciersnorth of the Himalayan ridge already on the Tibetan PlateauFor this subset of our PKH study area we measured amass budget ofminus126plusmn 42 Gt yrminus1 The two estimates over-lap within their error bars especially if our error is dou-bled to match the two standard error level used by Jacob etal (2012) The estimate of Jacob et al (2012) over our com-plete study region (PKH) ieminus6plusmn 8 Gt yrminus1 (sum of theirregions 8b and 8c) also includes mass changes for the KunlunShan sub-region for which we did not measure glacier massbalances Therefore the comparison with our PKH value(minus101plusmn 55 Gt yrminus1) is not straightforward but suggests areasonable agreement especially if we consider the slightmass gain (15plusmn 17 Gt yrminus1) measured with ICESat in theWest Kunlun (Gardner et al 2013)

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1281

Table 6 Annual average of the glacier seasonal and decadal mass loss contributions to Indus Ganges and Brahmaputra discharges Basinarea and seasonal contributions are given by Kaser et al (2010) glacier area in each basin is derived from the RGI v20 (Arendt et al 2012)Numbers in parentheses indicate the percentage of glacierized area within the river basin Seasonal contributions were modeled by Kaser etal (2010) based on the CRU 20 dataset over 1961ndash1990 assuming a balanced glacier mass budget Glacier decadal mass loss contributionsare given as annual average over 1999ndash2011

River Basin area Glacier area Glacier contribution Mean discharge(km2) (km2) (m3 sminus1) (m3 sminus1)

Seasonally delayed Decadal mass lossKaser et al (2010) (this study)

Indus 1 139 814 25 598 (2 ) 105 103plusmn 72 2146 (Chevallier personalcommunication 2012)

Ganges 1 023 609 11 168 (1 ) 47 103plusmn 49 11739 Papa et al (2012Brahmaputra 527 666 15 296 (3 ) 33 147plusmn 64 19710 updated)

54 Reasons for the heterogeneous pattern of massbalance

The PKH glacier mass balance is slightly negative(minus014plusmn 008 m we yrminus1) but changes are not homoge-neous from one sub-region to another and seem to coincidewith the climatic settings of the mountain range For the east-ern sites (from the Hengduan Shan to West Nepal) which areunder the influence of the Indian and south-east Asian sum-mer monsoons the mass balances are moderately negative(sim minus026 m we yrminus1) In the northwest of the PKH at thePamir and Karakoram sites where glacier climate is char-acterized by the westerlies in winter mass balances are bal-anced or slightly positive (sim +010 m we yrminus1) In betweenthese two influences the glaciers of the Spiti Lahaul sitein a monsoon-arid transition zone experienced the strongestmass loss (minus045plusmn 013 m we yrminus1)

This heterogeneous pattern of mass balances seems to berelated to trends in precipitation throughout the PKH Archerand Fowler (2004) reported an increase in winter precipi-tation over the Upper Indus Basin since 1961 a trend con-firmed by Yao et al (2012) over the Pamir and the Karako-ram based on the analysis of the Global Precipitation Cli-matology Project data set (GPCP Adler et al 2003) Thisprecipitation increase is a source of greater accumulation forglaciers and thus a possible explanation for their slightly pos-itive mass budgets In addition in the Upper Indus Basin themean summer temperature has been decreasing since 1961(Fowler and Archer 2006) which is consistent with the find-ings of Shekhar et al (2010) in the Karakoram who reporteda decrease in both maximum and minimum temperature since1988 These increases in winter precipitation and summertemperature likely translate in greater accumulation and re-duced ablation changes which are consistent with the bal-anced or slightly positive glacier mass budget

On the other hand in the east of the PKH the Indiansummer monsoon has been weakening since the 1950s (Bol-lasina et al 2011) which could have reduced the amount

of snowfall over the eastern study sites and thus resultedin negative mass balances In addition maximum and min-imum temperatures increased since 1988 in the western Hi-malaya (Shekhar et al 2010) and maximum temperaturesincreased in the central Himalaya (Nepal) since the mid-1970s (Shrestha et al 1999) Thus both precipitation andtemperature trends in the Himalaya likely contributed to thenegative mass balances measured over the correspondingstudy sites (from Spiti Lahaul to Hengduan Shan Fig 1)

However gridded data sets such as GPCP are not neces-sarily representative of the climate at high altitude IndeedHewitt (2005) reported a 5-to-10-fold increase in precipi-tation between the glacier front and the accumulation areain the Karakoram that is not captured by reanalysis dataas recently confirmed by Immerzeel et al (2012) Thus di-rect measurements of accumulation on High Mountain Asiaglaciers (eg Azam et al 2013) and high-altitude weatherstations are eagerly needed to validate the large scale grid-ded data set (eg reanalysis) test their ability to describe thespecific climate near to PKH glaciers and assess the relativerole played by temperature and precipitation changes

55 Contribution to water resources

Our glacier mass balance can be averaged over large riverbasins to estimate the mean contribution to river runoffdue to a decade of glacier mass loss We assume therebyonly direct river runoff without loss terms For the threerivers for which we cover the entire glacierized area of theircatchments within our sub-regions (Fig 1) these contribu-tions to the river discharge are equal to 103 m3 sminus1 (Indus)103 m3 sminus1 (Ganges) and 147 m3 sminus1 (Brahmaputra) for theperiod 1999ndash2011 (Table 6)

Rivers of PKH being key water resources measures oftheir discharges are highly sensitive and difficult to access forpolitical reasons Papa et al (2012) built recent time series ofdischarges for Ganges and Brahmaputra by using altimetrymeasurements The locations of the corresponding dischargeestimates in Bangladesh are given in Fig 1 We can therefore

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1282 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

evaluate the relative contribution of glacier decadal mass loss(Table 6) to annual river run-off at 09 for the Gangesat Hardinge (basin area ofsim 1 020 000 km2) and 07 forBrahmaputra at Bahadurabad (basin area ofsim 530 000 km2)between 1999 and 2011 Given the discharge measured at theGuddu station by the Surface Water Hydrology Project of theWater and Power Development Authority (Pakistan see lo-cation on Fig 1) between 1999 and 2003 we can estimatethe contribution of glacier mass loss at 48 for the IndusRiver at this station

The seasonal contributions of glaciers to river run-off (iethe part of the annual precipitation that experiences season-ally delayed release from the glaciers) directly results fromseasonal variations of meteorological quantities in contrastto the glacier mass loss contribution which results from thelong-term climatic change Both runoff components directlyaffect the amount of water available in a river at a given lo-cation and at a given time so that their comparison could beof interest Even if the seasonal timing of the glacier massloss contribution is unclear it could in general peak at asimilar time of the year than the seasonal glacier contribu-tion so that both components rather enlarge each other thancancel Seasonal glacier contributions have been modeled byKaser et al (2010) The latter study assumed glaciers to bein equilibrium with the present climate and thus prescribedglacier mass balances equal to 0 Thus their seasonally de-layed glacier contributions do not include the part due todecadal glacier mass loss and is complementary to our es-timates (Table 6) For the eastern basins under the influenceof the Indian summer monsoon (Ganges and Brahmaputra)the contribution of glacier mass loss is higher than the sea-sonal contribution In other words for the first decade of the21st century the mass loss of glaciers is on average moreimportant for water availability in those two rivers than theirseasonal water storage effect The situation is different for theIndus Basin where the glacier melt season is also the dry sea-son which results in a relatively higher seasonally delayedglacier contribution to the river discharge There the season-ally delayed water release from the glaciers and the decadalglacier mass loss contribute on average equally to the riverdischarge Both contributions are strongly (and equally) de-pendent on the location of the discharge measurement andwill be significantly larger (in relative term) in more heav-ily glacierized and smaller mountain catchments located inthe upper part of these major river basins (Bookhagen andBurbank 2010)

6 Conclusion

In this study we assessed the spatial pattern of glaciermass balance over the PKH by comparing the SRTM andSPOT5 DEMs over 9 well-distributed study sites between1999 and 2011 We found slightly positive or balanced massbudgets in the western part (Pamir Karakoram) moder-ate mass loss in the east (Nepal Bhutan Hengduan Shan)

and the most negative mass balance in the monsoon-aridtransition zone (Spiti Lahaul in the western Himalaya)By extrapolating these values to climatically homogeneoussub-regions we estimate the PKH overall mass balanceto have beenminus014plusmn 008 m we yrminus1 between 1999 and2011 which corresponds to a contribution to sea level riseof 0028plusmn 0015 mm yrminus1 This is about 4 of the globalglacier mass loss estimated by Gardner et al (2013) thoughover a slightly shorter time period (2003ndash2009) The largestsource of uncertainties in those geodetic mass balance esti-mates is the poorly constrained penetration of the C-BandSRTM radar signal into snow and ice

Our technique of DEM differencing allows mapping in de-tail the spatial pattern of glacier elevation changes Thosemaps reveal many surge-type behaviors both in the Pamirand the Karakoram It also appears that the glacier-wide massbalance does not seem to be considerably affected by themass transfer from surges over thesim 10 yr time period stud-ied Similar lowering rates over debris-covered and clean iceare observed for four study sites despite the commonly as-sumed insulating effect of debris covers This suggests forthe scale of entire glacier tongues a spatial mixture of re-duced ablation under intact thick debris covers and enhancedablation by supraglacial lakes or ice cliffs or due to thin de-bris layer and very low glacier velocities in ablation areasthat reduce the ice advection downstream

The regionally heterogeneous pattern of ice wastage is inbroad agreement with contrasted trends in large-scale pre-cipitation and temperature However glaciological (eg ac-cumulation) and meteorological records are needed at highaltitude in the vicinity of glaciers to evaluate more closelythe local influence of atmospheric variables on glacier massbalance and fully understand the reasons behind those con-trasted mass budgets in the PKH

Supplementary material related to this article isavailable online at httpwwwthe-cryospherenet712632013tc-7-1263-2013-supplementpdf

Acknowledgements We thank G Cogley A Gardner T NuimuraT Bolch P Wagnon M Barandun M Hoelzle M Huss and ananonymous referee for their constructive comments that led to aconsiderably improved manuscript We thank the Water and PowerDevelopment Authority (WAPDA) for their hydrological data(processed by P Chevallier) and F Papa for sharing his updatedrun-off data ahead of publication We thank T Bolch for sharinghis glacier outlines from the Everest area We thank the USGSfor allowing free access to their Landsat archive CGIAR-SCI forSRTM C-band data DLR for SRTM X-band data and GLIMS forglacier inventories J Gardelle acknowledges a PhD fellowshipfrom the French Space Agency (CNES) and the French NationalResearch Center (CNRS) E Berthier acknowledges support fromCNES through the TOSCA and ISIS proposal no 397 and fromthe Programme National de Teacuteleacutedeacutetection Spatiale E Berthieracknowledges M Masiokas and R Villalba for welcoming him

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1283

at the IANIGLA (Mendoza Argentina) Y Arnaud acknowledgessupport from French National Research Agency through ANR-09-CEP-005-01PAPRIKA A Kaumlaumlb acknowledges support fromESA through Glaciers_cci (400010177810I-AM) and from theEuropean Research Council under the European Unionrsquos SeventhFramework Programme (FP2007-2013)ERC Grant Agreementn 320816

Edited by I M Howat

The publication of this articleis financed by CNRS-INSU

References

Adler R Huffman G Chang A Ferraro R Xie P JanowiakJ Rudolf B Schneider U Curtis S Bolvin D Gruber ASusskind J Arkin P and Nelkin E The Version-2 GlobalPrecipitation Climatology Project (GPCP) Monthly Precipita-tion Analysis (1979ndashPresent) J Hydrometeorol 4 1147ndash11672003

Archer D R and Fowler H J Spatial and temporal variationsin precipitation in the Upper Indus Basin global teleconnectionsand hydrological implications Hydrol Earth Syst Sci 8 47ndash61doi105194hess-8-47-2004 2004

Arendt A Bolch T Cogley J G Gardner A Hagen J-OHock R Kaser G Pfeffer W T Moholdt G Paul F RadicV Andreassen L Bajracharya S Beedle M Berthier EBhambri R Bliss A Brown I Burgess E Burgess DCawkwell F Chinn T Copland L Davies B de AngelisH Dolgova E Filbert K Forester R Fountain A Frey HGiffen B Glasser N Gurney S Hagg W Hall D Hari-tashya U K Hartmann G Helm C Herreid S Howat IKapustin G Khromova T Kienholz C Koenig M KohlerJ Kriegel D Kutuzov S Lavrentiev I LeBris R Lund JManley W Mayer C Miles E Li X Menounos B Mer-cer A Moelg N Mool P Nosenko G Negrete A NuthC Pettersson R Racoviteanu A Ranzi R Rastner P RauF Rich J Rott H Schneider C Seliverstov Y Sharp MSigurethsson O Stokes C Wheate R Winsvold S WolkenG Wyatt F and Zheltyhina N Randolph Glacier Inventory[v20] A Dataset of Global Glacier Outlines Global Land IceMeasurements from Space Boulder Colorado USA Digital Me-dia httpwwwglimsorgRGIrandolphhtml 2012

Azam M Wagnon P Ramanathan A Vincent C Sharma PArnaud Y Linda A Pottakkal J Chevallier P Singh V andBerthier E From balance to imbalance a shift in the dynamicbehaviour of Chhota Shigri glacier western Himalaya India JGlaciol 58 315ndash324 doi1031892012JoG11J123 2012

Azam M F Wagnon P Vincent C Ramanathan A Linda Aand Singh V B Reconstruction of the annual mass balanceof Chhota Shigri Glacier (Western Himalaya India) since 1969Ann Glaciol in revision 2013

Barrand N and Murray T Multivariate Controls on the In-cidence of Glacier Surging in the Karakoram Himalaya

Arct Antarct Alp Res 38 489ndash498 doi1016571523-0430(2006)38[489MCOTIO]20CO2 2006

Barundun M Huss M Azisov E Gafurov A HoelzleM Merkushkin A Salzmann N and Usubaliev R Re-establishing seasonal mass balance observation at AbramovGlacier Kyrgyzstan from 1968ndash2012 Abstract EGU2013-5668European Geophysical Union Vienna 2013

Benn D Bolch T Hands K Gulley J Luckman ANicholson L Quincey D Thompson S Toumi R andWiseman S Response of debris-covered glaciers in theMount Everest region to recent warming and implicationsfor outburst flood hazards Earth-Sci Rev 114 156ndash174doi101016jearscirev201203008 2012

Berthier E and Vincent C Relative contribution of surface mass-balance and ice-flux changes to the accelerated thinning of Merde Glace French Alps over 1979ndash2008 J Glaciol 58 501ndash512doi1031892012JoG11J083 2012

Berthier E Arnaud Y Baratoux D Vincent C and Reacutemy FRecent rapid thinning of the ldquo Mer de Glace rdquo glacier derivedfrom satellite optical images Geophys Res Lett 31 L17401doi1010292004GL020706 2004

Berthier E Arnaud Y Kumar R Ahmad S Wagnon P andChevallier P Remote sensing estimates of glacier mass balancesin the Himachal Pradesh (Western Himalaya India) RemoteSens Environ 108 327ndash338 doi101016jrse2006110172007

Bhambri R Bolch T Kawishwar P Dobhal D P SrivastavaD and Pratap B Heterogeneity in Glacier response from 1973to 2011 in the Shyok valley Karakoram India The CryosphereDiscuss 6 3049ndash3078 doi105194tcd-6-3049-2012 2012

Bhutiyani M Mass-balance studies on Siachen Glacier in theNubra valley Karakoram Himalaya India J Glaciol 45 112ndash118 1999

Bishop M P Schroder J F and Ward J L SPOT multispec-tral analysis for producing supraglacial debris-load estimates forBatura glacier Pakistan Geocarto Int 10 81ndash90 1995

Bolch T Pieczonka T and Benn D I Multi-decadal mass lossof glaciers in the Everest area (Nepal Himalaya) derived fromstereo imagery The Cryosphere 5 349ndash358 doi105194tc-5-349-2011 2011

Bolch T Kulkarni A Kaumlaumlb A Huggel C Paul F Cogley JFrey H Kargel J Fujita K Scheel M Bajracharya S andStoffel M The state and fate of Himalayan Glaciers Science336 310ndash314 doi101126science1215828 2012

Bollasina M Ming Y and Ramaswamy V Anthropogenicaerosols and the weakening of the South Asian summer mon-soon Science 334 502ndash505 doi101126science12049942011

Bookhagen B and Burbank D Toward a complete Himalayan hy-drological budget spatiotemporal distribution of snowmelt andrainfall and their impact on river discharge J Geophys Res115 F03019 doi1010292009JF001426 2010

Bretherton C Widmann M Dymnikov V Wallace J and BladeacuteI The effective number of spatial degrees of freedom of a time-varying field J Climate 12 1990ndash2009 1999

Copland L Pope S Bishop M Shroder J Clendon PBush A Kamp U Seong Y and Owen L Glacier veloc-ities across the central Karakoram Ann Glaciol 50 41ndash49doi103189172756409789624229 2009

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

1266 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Table 1 Characteristics of the nine study sites Numbers between parentheses indicate the percentage difference between our user-definedglacier inventory and the RGI v20 (Arendt et al 2012)

Site name Glacier area Debris ( of Glacier area Arendt Date Landsat imagesthis study (km2) glacier area) et al (2012) (km2) (SPOT5 DEM)

PathRow Date

Hengduan Shan 1303 11 2541 (+88 ) 24 Nov 2011 p135r39 23 Sep 1999Bhutan 1384 11 1429 (+3 ) 20 Dec 2010 p138r40 19 Dec 2000

p138r41Everest 1461 22 1400 (minus4 ) 4 Jan 2011 p140r41 30 Oct 2000West Nepal 908 17 802 (minus12 ) 3 Jan 2011 p143r39 1 Aug 2009

p143r40Spiti Lahaul 2110 13 2257 (+7 ) 20 Oct 2011 p147r37 15 Oct 2000

p147r38Hindu Kush 793 7 724 (minus11 ) 17ndash21 Oct 2008 p150r34 15 Aug amp

p150r35 16 Sep 1999Karakoram East 5328 10 5384 (+1 ) 31 Oct 2010 p148r35 7 Sep 1998Karakoram West 5434 12 5687 (+1 ) 3 Dec 2008 p149r35 29 Aug 1998Pamir 3178 11 3210 (minus1 ) 29 Nov 2011 p152r33 16 Sep 2000

3 Data and methods

31 Digital elevation models

For each study site we used the Shuttle Radar TopographicMission (SRTM) version 4 DEM as the reference topogra-phy (Rabus et al 2003) This data set acquired in Febru-ary 2000 comes along with a mask to identify the pixelswhich have been interpolated (both were downloaded fromhttpsrtmcsicgiarorg) On each site the SRTM DEM issubtracted from a more recent DEM built from a stereo pairacquired by the HRS sensor onboard the SPOT5 satellite(Korona et al 2009) except for the Hindu Kush study sitewhere the recent DEM has been created from a stereo pairacquired by the HRG sensor also onboard SPOT5 (Berthieret al 2007) The date of each SPOT5 DEM acquisition dif-fers depending on the study site (Table 1) but 7 out of 9 wereacquired between October 2010 and November 2011 EachSPOT5-HRS DEM is also provided with a correlation maskand an ortho-image generated from the rear HRS image

SPOT5-HRS DEMs were produced by the French Map-ping Agency (IGN) using two sets of correlation parametersdefined during the SPIRIT (SPOT 5 stereoscopic survey ofPolar Ice Reference Images and Topographies) internationalpolar year project (Korona et al 2009) Thus two versionsof the DEM are delivered with their respective mask one op-timized (v1) for a gentle topography and the second one (v2)for a more rugged relief Earlier work has shown that v2 hasa slightly larger percentage of interpolated pixels but for thenon-interpolated pixels smaller errors (Korona et al 2009)This version (v2) is preferred here Given that both DEMsare calculated from the same image pair little difference isexpected between v1 and v2 However locally we identifiedsome large bull eye elevation differences between the two

versions of the DEM both off and on glaciers Comparisonwith the SRTM DEM shows that both versions of the DEMare erroneous for those areas and thus we preferred to ex-clude from further analysis all pixels for which the absoluteelevation difference between the two versions of the DEMis larger than 5 m About 20 of the valid DEM pixels areexcluded through that procedure

The SRTM DEM and mask originally at a 3 arcsecondresolution (sim 90 m) are resampled bilinearly to 40 m (UTMprojection WGS84 ellipsoid) to match the posting and pro-jection of the SPOT5 DEM Altitudes are defined above theEGM96 geoid for both DEMs

32 Glacier mask

Where available glacier outlines were downloaded fromthe GLIMS database (Raup et al 2007 Supplement) andmanually corrected if needed Otherwise the glacier maskwas derived from Landsat-TM and Landsat-ETM+ imagesPathrow and dates of Landsat acquisition are given in Ta-ble 1 for each study site A threshold varying between 05and 07 depending on the study site was applied to the Nor-malized Difference Snow Index (TM2-TM5)(TM2+TM5)to detect clean ice and snow automatically whereas debris-covered parts were digitized manually by visual interpreta-tion (eg Racoviteanu et al 2009) The total glacier areaand debris-covered part can be found for each study site inTable 1

Most inventories are derived from Landsat images withminimal snow cover from around the year 2000 prior to the2003 failure of the Scan Line Corrector of the ETM+ sensoronboard Landsat 7 that resulted in image striping Glacierareas are expected to have experienced minor changes sincethen except for glacier fronts that retreated due to the rapid

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1267

Table 2 Elevation ranges equilibrium line altitudes (ELA) and accumulation area ratios (AAR) for the nine study sites The dates ofacquisition of Landsat images used for ELA determination are also given

Sub-region Glaciers elevation ELA (m) AAR Landsat imagesrange (m)

Min Max This study Scherler et Kaumlaumlb et This study Kaumlaumlb et PathRow Dateal (2011a) al (2012) al (2012)

Hengduan Shan 2630 6860 4970plusmn 320 NA NA 55 NA p135r39 23 Sep 1999Bhutan 4390 7480 5690plusmn 440

5700 555036

47p138r40 17 Nov 2000

Everest 4260 8380 5840plusmn 320 31 p140r41 31 oct 2009West Nepal 3950 6870 5590plusmn 138 37 p143r39 1 Aug 2009Spiti Lahaul 3780 6360 5390plusmn 140 5103 5500 34 35 p147r37 2 Aug 2002

p147r38Hindu Kush 2717 6385 5050plusmn 160 5138 30 p150r34 2 Sep 2000

p150r35Karakoram east 3230 7770

5030plusmn 280 4845 554076

47p148r35 7 Sep 1998

Karakoram west 2910 7920 66 p149r35 29 Aug 1998Pamir 2800 7090 4580plusmn 250 NA NA 52 NA p152r33 30 Jul 2000

growth of connected lakes and for glaciers that surged andthus advanced For the few glaciers that advanced their frontposition was updated manually based on the SPOT5 ortho-image

For each study site we favored our user-defined glacierinventories instead of the global Randolph Glacier Inventory(RGI v20 Arendt et al 2012) because the latter is hetero-geneous and for some study sites did not match our accu-racy requirements for the adjustments and corrections of theDEMs

We also estimated the altitude of the equilibrium line(ELA) as the snowline on thesim 2000 Landsat data with min-imal snow cover to separate ablation and accumulation ar-eas The ELA is assumed to be identical to the altitude ofthe snowline at the end of the ablation season (eg Raba-tel et al 2005) after the end of the monsoons or before thefirst snowfalls Therefore we manually digitized snowlinesfor more than 30 glaciers for each study site and computedtheir mean elevations Our ELAs are slightly different fromthose given by Scherler et al (2011a) and Kaumlaumlb et al (2012)probably due to the sensitivity of the ELA to the choice ofthe image (Table 2) Ideally the ELA should have been es-timated from images acquired at the end of the ablation sea-son during various years covering our whole study periodThis would be a highly time-consuming task that was not ac-complished here However we stress that in our study theELA is only used to compare elevation changes on the abla-tionaccumulation areas (Sect 41) and to estimate the sen-sitivity of our mass balances to the choice of the volume-to-mass conversion factor (Sect 35) Our preferred mass bal-ance estimate uses a unique volume-to-mass conversion fac-tor for the whole glacier and is thus independent from theELA

33 Adjustments and corrections of DEM biases

Different DEM processing steps are followed to extract un-biased glacier elevation changes for each study site (eg Nuthand Kaumlaumlb 2011 Gardelle et al 2012b)

331 Planimetric adjustment of the DEMs

A horizontal shift between two DEMs results in an aspect-dependent bias of elevation differences (Nuth and Kaumlaumlb2011) Here the horizontal shift is determined by minimizingthe root mean square error of elevation differences (SPOT5-SRTM) off glaciers where the terrain is assumed to be stableover the study period (Berthier et al 2007) The SRTM DEMis then resampled (cubic convolution) according to the shift

332 Alongacross track corrections (drift of thesatellite orbit)

For DEMs derived from stereo imagery the satellite acquisi-tion geometry can induce a bias along andor across the trackdirection (Berthier et al 2007) We estimate the azimuth ofeach SPOT5 ground track and use it to rotate the coordinatesystem accordingly (Nuth and Kaumlaumlb 2011) Then we com-pute the elevation differences along and across the satellitetrack on stable areas and when necessary correct the bias us-ing a 5th order polynomial fit

333 Curvature correction

As suggested by Paul (2008) and Gardelle et al (2012b) thedifference in original spatial resolutions of the two DEMs canlead to biases related to altitude in mountainous areas Whenthe curvature the first derivative of the slope is high (sharppeaks or ridges) the altitude tends to be underestimatedby the coarse DEM unable to reproduce high-frequency

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1268 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

slope variations Thus this apparent ldquoelevation biasrdquo is cor-rected using the relation between elevation differences andthe maximum terrain curvature estimated over stable areasoff glaciers (Gardelle et al 2012b) Note that the units inFig 1b in Gardelle et al (2012b) should be 10minus2 mminus1 in-stead of mminus1 (A Gardner personal communication 2012)However this does not impact the validity of the correction

334 Radar penetration correction

The SRTM DEM acquired in C-band potentially underesti-mates glacier elevations since at this wavelength (sim 56 cm)the penetration of the radar signal into snow and ice canreach several meters (eg Rignot et al 2001) Here we esti-mate this penetration for each study site by differencing theSRTM C-band (57 GHz) and X-band (97 GHz) DEMs Thelatter was acquired simultaneously with the SRTM X-bandat higher resolution (30 m) but with a narrower swath andthus has a coverage restricted to selected swaths only Sincethe penetration of the X-band is expected to be low com-pared to the C-band (an hypothesis that still needs to be con-firmed Ulaby et al 1986) we consider the elevation differ-ence (SRTMC-bandminus SRTMX-band) to be a first approxima-tion of C-band radar penetration over glaciers (Gardelle etal 2012b) The correction is calculated as a function of alti-tude and applied to each pixel separately

For the Hindu Kush Spiti Lahaul and West Nepal sitesdue to the lack of SRTMX-band data we were unable to de-termine the value of the penetration Thus we applied thecorrection computed over the nearest study site ie Karako-ram for Spiti Lahaul and Hindu Kush and Everest for WestNepal The mean values of the SRTMC-bandpenetrations overglaciers are given for each study site in Table 3

335 Seasonality correction

SPOT5 DEMs were acquired between late October and earlyJanuary depending on the study site Since the SRTM DEMis from mid-February we need to account for possible masschanges during this 1-to-5-month period in order to esti-mate glacier mass balance over an integer number of years(Gardelle et al 2012a) For the eastern sites (from WestNepal to Hengduan Shan) where glaciers are of summer-accumulation types (Fujita 2008) the correction was set to0 This choice is confirmed by recent field observations thatshow no winter accumulation on Mera and Pokalde glaciersin Nepal (Wagnon et al 2013) For the Karakoram studysites we used the winter accumulation rate+013 m we permonth measured on Biafo Glacier (Wake 1989) For theother western study sites (the Pamir Hindu Kush and SpitiLahaul sites) the correction is derived from the mean of win-ter mass balances of 35 glaciers all in the Northern Hemi-sphere+089 m we yrminus1 over 2000ndash2005 corresponding to+015 m we per winter month (Ohmura 2011) The lattervalue is slightly lower than the average winter mass balance

Table 3 Average SRTM C-band penetration estimates (in m)in February 2000 over glaciers in this study and from Kaumlaumlb etal (2012)

Sub-region This study Kaumlaumlb et al (2012)

Hengduan Shan 17 NABhutan 24

25plusmn 05Everest 14West Nepal NA

15plusmn 04Spiti Lahaul NAHindu Kush NA 24plusmn 04Karakoram East

34 24plusmn 03Karakoram WestPamir 18 NA

(+135plusmn 031 m we yrminus1) measured on Abramov Glacier inthe Pamir during 1968ndash1998 (WGMS 2012) and in reason-ably good agreement with the modeled mean winter massbalance (+108plusmn 034 m we yrminus1) of Chhota Shigri Glacierduring 1969ndash2012 in the western Himalaya (Azam et al2013)

336 Mean elevation changes and mass balancecalculation

Before averaging elevation changes we exclude all interpo-lated pixels in the SRTM or SPOT5 DEMs as well as unex-pected elevation changes exceedingplusmn150 m for study sites(Pamir and Karakoram) that include surge-type glaciers orplusmn80 m for other sites Those thresholds were chosen aftercareful inspection of the elevation difference maps and his-tograms in order to identify the largest realistic glacier eleva-tion changes Glaciers that are truncated at the edges of theDEM are also excluded from mass balance calculations asthey may bias the final results if for example only their accu-mulation or ablation area is sampled Although they are trun-cated Siachen (Karakoram East) and Fedtchenko (Pamir)glaciers were retained because of their wide coverage andbecause their accumulation and ablation areas were correctlysampled in the DEMs

Elevation changes are then analyzed for 100 m altitudebins Within each altitude band we average the elevationchanges after excluding pixels where absolute elevation dif-ferences differ by more than three standard deviations fromthe mean (Berthier et al 2004 Gardner et al 2012) Thisis an efficient way to exclude outliers based on the assump-tion that elevation changes should be similar at a given alti-tude within each study site This assumption does not holdfor regions containing actively surging glaciers Thus forthe Pamir and Karakoram study sites surge-type glaciersare treated separately ie elevation changes are averaged by100 m altitude bands for each surge-type glacier individually

Where no elevation change is available for a pixel weassign to it the value of the mean elevation change of the

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1269

altitude band it belongs to in order to assess the mass bal-ance over the whole glacier area

The conversion of elevation changes to mass balance re-quires knowledge of the density of the material that has beenlost or gained and its evolution during the study period Giventhe lack of repeat measurement of density profiles over theentire snowfirnice column in the PKH we applied a con-stant density conversion factor of 850 kg mminus3 for all ourstudy sites (Sapiano et al 1998 Huss 2013)

The mass balance of each surge-type glacier is computedseparately and added (area-weighted) to the mass balance ofthe rest of the glaciers to estimate the mass balance of thewhole study site

For study sites with a high concentration of growingpro-glacial lakes such as Bhutan Everest and West Nepal(Gardelle et al 2011) our mass losses are slightly underes-timated because they do not take into account the glacier icethat has been replaced by water during the expansion of thelake Subaqueous ice losses are only estimated for LhotseSharImja Glacier (Everest area Fujita et al 2009) for thesake of comparison to previous studious

The region-wide mass balance of PKH glaciers as definedin Sect 2 is computed by extrapolating the mass balancefrom each study site to a wider sub-region (Fig 1) that is as-sumed to experience the same glacier mass changes becauseof its similar climate (Bolch et al 2012 Kaumlaumlb et al 2012)The total glacier area for each sub-region is computed fromthe RGI v20 (Arendt et al 2012)

The periods over which we calculate mass balance areslightly different from one site to another (Table 1) depend-ing on the year of acquisition of the SPOT5 DEM (two in2008 two in 2010 and five in 2011) In the following we willassume that all mass balances are representative of the period1999ndash2011 in order to estimate a region-wide mass budgetof PKH glaciers neglecting thus part of the inter-annual vari-ability

34 Accuracy assessment

The errorE1hi to be expected for a pixel elevation change

1hi is assumed to equal the standard deviationσ1h of themean elevation change1h of the altitude band it belongs toThe value ofσ1h can range fromplusmn4 m toplusmn 20 m dependingon the study site and elevation This is rather conservative asthe value ofσ1h contains also the intrinsic natural variabilityof elevation changes within the altitude band (ie it containsboth noise and real geophysical signal)

The errorE1h of the mean elevation change1h in eachaltitude band is then calculated according to standard princi-ples of error propagation

E1h =E1hiradicNeff

(1)

Neff represents the number of independent measurements inthe altitude bandNeff will be lower thanNtot the total num-

ber of elevation change measurements1hi in the altitudeband since the latter are correlated spatially (Bretherton etal 1999)

Neff =Ntot middot PS

2d (2)

where PS is the pixel size (here 40 m)d is the distance ofspatial autocorrelation determined using Moranrsquos I autocor-relation index on elevation differences off glaciers On theaverage over the 9 study sitesd = 492plusmn 72 m

The error on the penetration correction was assumed to besystematic due to the not fully understood physics involvedand equal toplusmn 15 m This error was obtained by compar-ing the SRTM C-band penetration estimated using two inde-pendent methods (i) SRTMC-bandndash SRTMX-band and (ii) dif-ference between ICESat and SRTM when ICESat elevationchange trends during 2003ndash2008 are extrapolated to Febru-ary 2000 (Kaumlaumlb et al 2012) The maximum difference of thepenetration inferred from the two methods is 10 m (Table 3)A 50 additional error margin was added to account for thefact that this comparison could only be made on 2 of ourstudy sites

Given the slender observational support for the seasonal-ity correction (Section 33) we assumed its uncertainty tobe plusmn100 on the western sites ieplusmn015 m we per win-ter month The same absolute uncertainty (plusmn015 m we perwinter month) was used for the eastern sites where no sea-sonality correction is applied but ablation andor accumula-tion may occur in winter (Wagnon et al 2013)

All errors are summed quadratically within each altitudeband and then summed for all altitude bands assuming con-servatively that they are 100 dependent The error on theSRTM penetration correction clearly dominates our errorbudget for volume change

During the conversion from volume to mass we assumea plusmn60 kg mminus3 error on the density conversion factor (Huss2013) Thisplusmn7 error is similar to the one estimated byBolch et al (2011) In a sensitivity test the volume to massconversion is also performed using two alternative densityscenarios (Kaumlaumlb et al 2012) (i) a density of 900 kg mminus3 isassumed everywhere and (ii) 600 kg mminus3 in the accumula-tion area and 900 kg mminus3 in the ablation area These two sce-narios take among other things into account that elevationchanges could be due to changes in glacier dynamics or dueto surface mass balance We then calculate the maximum ab-solute difference between the mass balances calculated usingour preferred scenario (850plusmn 60 kg mminus3 everywhere) and themass balances calculated using the two others scenarios Forthe eastern sites (West Nepal to Hengduan Shan) the max-imum absolute difference is 003 m we yrminus1 For the west-ern sites (Pamir to Spiti Lahaul) it reaches 006 m we yrminus1Those uncertainties due to the choice of a given density sce-nario remain negligible compared to other sources of errors

We also include an error due to extrapolation of our massbalance estimate from our study site (the extent of each

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1270 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

SPOT5 DEM) to the unmeasured glaciers within each sub-region To estimate this regional extrapolation error we com-pare the ICESat elevation change trends (computed as inKaumlaumlb et al 2012) within 3 times 3 cells centered at the studysites (Table 5) with the ICESat-derived trends for the entiresub-regions The absolute difference between both trends isless than 002ndash003 m yrminus1 for Bhutan West Nepal Spiti La-haul and Karakoram ie negligible For Everest the eleva-tion trend for the entire sub-region is 012 m yrminus1 less nega-tive than for the 3 times 3 cells around the SPOT5 DEM cen-ter for Hindu Kush the trend for the entire sub-region is017 m yrminus1 more negative than for the 3 times 3 cell due to thevariability of elevation changes within this sub-region (Kaumlaumlbet al 2012) The regional extrapolation error is obtainedas the area-weighted mean absolute difference between themass balances derived from ICESat for those 3 times 3 cellsand the whole sub-region and equalsplusmn004 m we yrminus1

Finally for sub-regions and total PKH estimates of masschange we include an error on the glacier area computedfrom the RGI v20 (Arendt et al 2012) The inventory er-ror is specific to each study site and evaluated based on thecomparison of our user-defined inventory on each study sitewith the RGI v20 (Table 1) We note the remarkable accu-racy of the RGI for all but one of our study sites The rel-ative errors are generally of a few percents and up to 12 for West Nepal The differences between our and existing in-ventories are probably due to the difficulty of delimitatingdebris-covered glacier parts (eg Frey et al 2012 Paul etal 2013) and accumulation areas The Hengduan Shan studysite is an exception There the RGI is based on the ChineseGlacier Inventory (Shi et al 2009 Arendt et al 2012) andoverestimates the ice-covered area by 88 compared to ourLandsat-based inventory

Unless stated otherwise all error bars from this study andpreviously published studies correspond to one standard er-ror

4 Results

41 Overview of the elevation changes

The maps of glacier elevation changes are given in Figs 2ndash10for our nine study sites with as inset the distribution ofthe elevation differences off glaciers The full maps of ele-vation differences off glaciers are shown in the Supplement(Figs S2ndashS10) Those off glacier maps indicate a local ran-dom noise of aboutplusmn5ndash10 m However at length scales of afew kilometers the spatial variations in elevation differencesare small typically less than 1 m

For the four eastern study sites from the HengduanShan to the West Nepal (Fig 1) the elevation changesaveraged over the accumulation areas are small (between003plusmn 014 m yrminus1 to 013plusmn 015 m yrminus1) whereas clear sur-face lowering is observed for all four study sites in their

ablation areas ranging fromminus050plusmn 014 m yrminus1 (Bhutan)to minus081plusmn 013 m yrminus1 (Hengduan Shan) Over the BhutanEverest and West Nepal study sites glaciers in contact witha proglacial lake often experienced larger thinning rates

Further west the Spiti Lahaul is the only site wheresurface lowering is measured both in the accumula-tion area (minus034plusmn 016 m yrminus1) and the ablation area(minus070plusmn 014 m yrminus1) Clear surface lowering is also ob-served in the ablation areas of the Hindu Kush glaciers(minus050plusmn 018 m yrminus1)

In the Karakoram (east and west Figs 8 and 9) andin the Pamir (Fig 10) the pattern of elevation changesis highly heterogeneous due to the occurrence of numer-ous glacier surges or similar flow instabilities The eleva-tion changes in the accumulation areas of those three studysites are slightly positive (between+022plusmn 014 m yrminus1 and+030plusmn 018 m yrminus1) In the ablation areas small surfacelowering is observed for the two Karakoram study sites (av-erage of the two sitesminus033plusmn 016 m yrminus1) whereas no el-evation change is found in Pamir (minus002plusmn 013 m yrminus1)

Note that those mean rates of elevation changes are sensi-tive to the choice of the ELA assumed to equal the snowlinealtitude in Landsat images from around year 2000

42 Influence of a debris cover

To evaluate thinning rates over debris-covered and debris-free ice we perform a histogram adjustment in order tocompare data sets with similar altitude distributions This isneeded because debris-covered parts tend to be concentratedat lower elevations The adjustment consists in randomly ex-cluding pixels over debris-free ice from each elevation bandto match a scaled version of the debris-covered pixel distribu-tion (Kaumlaumlb et al 2012) Elevation changes with altitude canthen be compared for clean and debris-covered ice (Fig 11)

The thinning is higher for debris-covered ice in the Ever-est area and on the lowermost parts of glaciers in the PamirBut on average in the Pamir western Karakoram and SpitiLahaul the lowering of debris-free and debris-covered ice issimilar The thinning is stronger over clean ice in the Heng-duan Shan Bhutan West Nepal Hindu Kush and althoughto a lesser extent in the eastern Karakoram

43 Mass balance over the nine study sites

Glacier mass changes have been heterogeneous overthe PKH since 1999 Slight mass gains or balancedmass budgets are found in the north-west in thePamir (+014plusmn 013 m we yrminus1) and for the two studysites in the central Karakoram (+009plusmn 018 m we yrminus1

and +011plusmn 014 m we yrminus1) Mass loss is moderatein the adjacent Hindu Kush with a mass balance ofminus012plusmn 016 m we yrminus1 Glaciers in all our eastern studysites (two in Nepal Bhutan and Hengduan Shan) experi-enced homogeneous mass losses with mass balances ranging

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1271

Fig 2 Glacier elevation changes over the Hengduan Shan study site Inset distribution of the elevation differences off glacier The medianand the standard deviation of the elevation difference and the number of pixels off glaciers are given (by construction the mean elevationdifference is null)

Fig 3Same as Fig 2 but for the Bhutan study site

from minus033plusmn 014 m we yrminus1 to minus022plusmn 012 m we yrminus1The most negative mass balance is measured in Spiti Lahaul(western Himalaya) atminus045plusmn 013 m we yrminus1 (Table 4)

However within a study site mass balances can be highlyvariable from one glacier to another In the Everest area twolarge glaciers (71 km2 each) experienced distinctly differentmass balance withminus016plusmn 016 m we yrminus1 for the south-flowing Ngozumpa Glacier andminus059plusmn 016 m we yrminus1

for the north-flowing Rongbuk Glacier (Fig 4) In Bhutan(Fig 3) mass loss is higher south (minus025plusmn 013 m we yrminus1)

than north (minus014plusmn 012 m we yrminus1) of the main Hi-malayan ridge though not at a statistically significant level

The region-wide mass budget of PKH glaciers isnegative minus101plusmn 55 Gt yrminus1 (minus014plusmn 008 m we yrminus1)which corresponds to a sea level rise contribution of0028plusmn 0015 mm yrminus1 over 1999ndash2011

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1272 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 4Same as Fig 2 but for the Everest study site

Fig 5Same as Fig 2 but for the West Nepal study site

44 Glacier surges

In the Pamir and the Karakoram the spatial pattern of ele-vation changes of many glaciers is typical of surge events orflow instabilities They can be identified because of their veryhigh thinning and thickening rates that can both reach up to16 m yrminus1 (Figs 8ndash10)

Most of them have already been reported to be surge-type glaciers based on their velocity (Kotlyakov et al 2008Quincey et al 2011 Copland et al 2009 Heid and Kaumlaumlb2012) morphology (Copland et al 2011 Barrand and Mur-ray 2006 Hewitt 2007) or elevation changes (Gardelle etal 2012a) Here we only identify surge-type glaciers for thesake of mass balance calculation and we do not attempt toprovide an exhaustive up-to-date surge inventory This is why

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1273

Fig 6 Same as Fig 2 but for the Spiti Lahaul study site The map of elevation changes contains many voids due to the presence ofthin clouds during the acquisition of the SPOT5 DEM The mass balance of the Chhota Shigri and Bara Shigri glaciers are respectivelyminus039plusmn 018 m we yrminus1 andminus048plusmn 018 m we yrminus1

the published surge inventories in the Pamir and the Karako-ram do not necessarily match the surge-type glaciers labeledon the elevation change maps as these refer only to our ob-servation period (1999ndash2011)

The mass balance of surge-type glaciers (respectivelynon-surge-type glaciers) is +019plusmn 022 m we yrminus1

(+012plusmn 011 m we yrminus1) in the Pamir+009plusmn 030 m we yrminus1 (+009plusmn 015 m we yrminus1) inthe western Karakoram and+010plusmn 025 m we yrminus1

(+012plusmn 012 m we yrminus1) in the eastern Karakoram Theoverall similarity between the mass budgets for surge-typeand non-surge-type glaciers shows that those flow instabili-ties seem not to affect the glacier-wide mass balance at leastover short time periods here 10 yr

5 Discussion

51 SRTM penetration

Alternative estimates of SRTMC-bandpenetration for the PKHregion were obtained by Kaumlaumlb et al (2012) by comparing el-evations of the ICESat altimeter (Zwally et al 2002) with theSRTMC-bandDEM in PKH and extrapolating the 2003ndash2008trend of elevation changes back to the acquisition date ofSRTM The difference between east and west is more dis-tinct in our case with a mean penetration in the Karakoram of34 m (24plusmn 03 m in Kaumlaumlb et al 2012) 24 m in Bhutan and14 m around the Everest area (25plusmn 05 m for a wider areaincluding east Nepal and Bhutan in Kaumlaumlb et al 2012) Thiseastwest gradient is expected given that in February 2000when the SRTM mission was flown the snowfirn thicknesswas probably larger in the western PKH where glaciers areof winter-accumulation type than in the eastern PKH where

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1274 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 7Same as Fig 2 but for the Hindu Kush study site

glaciers are of summer-accumulation type Further studieswould be necessary to investigate the possible reasons be-hind the relatively low SRTMC-bandpenetration for the Pamir(18 m) that do not follow the eastwest gradient mentionedabove

Both our and Kaumlaumlb et alrsquos (2012) estimates agree withintheir error bars and seem thus to provide robust first-orderestimates for the SRTMC-bandradar penetration in the regionHowever it remains to be verified if a penetration depth com-puted at the regional scale is appropriate for a single smallglacier whose altitude distribution is different from the oneof whole study site (eg Abramov Glacier in the far north ofour Pamir study site)

52 Thinning over debris-covered ice

PKH glaciers are heavily debris-covered in their ablation ar-eas because of steep rock walls that surround them and in-tense avalanche activity Twelve percent (12 ) of the glacierarea in our study sites are covered with debris (up to 22 inthe Everest area Table 1) which is in agreement with pre-vious estimates for the Himalaya and the Karakoram 13 in Kaumlaumlb et al (2012) and 10 in Bolch et al (2012) Thedebris layer is expected to modify the ablation of the under-

lying ice It will increase ablation if its thickness is just a fewcentimeters (dominance of the albedo decrease) or decreaseablation it if the layer is thick enough to protect the ice fromsolar radiation (Mattson et al 1993 Mihalcea et al 2006Benn et al 2012 Lejeune et al 2013)

However recent studies have reported similar overall thin-ning rates over debris-covered and debris-free ice in thePKH (Kaumlaumlb et al 2012 Gardelle et al 2012a) or evenhigher lowering rates over debris-covered parts in the Ever-est area (Nuimura et al 2012) Our findings are consistentwith Nuimura et al (2012) for the Everest study site with astronger thinning (rate of elevation change ofminus097 m yrminus1)

in debris-covered areas than over clean ice (rate of elevationchange ofminus057 m yrminus1) In the Pamir Karakoram and SpitiLahaul study sites these rates are similar while in the HinduKush West Nepal Bhutan and Hengduan Shan study sitesthinning is greater over clean ice in agreement with the com-monly assumed protective effect of debris (Fig 4) Thosedifferences between our 9 study sites suggest a complex re-lationship between debris cover and glacier wastage

As local glacier thickness changes are a result of bothmass balance processes and a dynamic component theseunexpected high thinning rates over debris-covered parts

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1275

Fig 8Same as Fig 2 but for the Karakoram east study site Surge-type glaciers in an active surge phase (resp quiescent phase) are identifiedwith a solid triangle (resp solid circle)

are potentially the result of glacier-wide processes Over aglacier tongue the ablation can be enhanced by the pres-ence of steep exposed ice cliffs or supraglacial lakes andponds (Sakai et al 2000 2002) In addition it is also pos-sible that the glacier tongues are mostly covered by thindebris layers such that the albedo effect dominates the in-sulating effect resulting in an increased tongue-wide abla-tion The prevalence of thin debris was recently suggestedfor debris-covered glaciers around the Mount Gongga in thesouth-eastern Tibetan Plateau (Zhang et al 2013b) Finallyice-flow rates could be very low at debris-covered parts asshown by Quincey et al (2007) in the Everest area Kaumlaumlb(2005) for Bhutan and Scherler et al (2011b) in the HinduKush-Karakoram-Himalaya Thus all three factors are likelyto increase the overall thinning under debris despite the un-doubted protective effect of a continuous thick debris coverat a local scale Future investigations combining surface ve-locity fields and maps of elevation changes could allow eval-uating more closely the relative role of ablation and ice fluxesin thinning rates over PKH glacier tongues (eg Berthierand Vincent 2012 Nuth et al 2012) Measurements of the

spatial distribution of debris thickness over the PKH glaciertongues are also needed

Note that in the case of debris-covered tongues the ele-vation change measurement represents the glacier thicknesschange but also the possible debris thickness evolution Thelatter remains poorly known in the PKH or in any othermountain range But based on the debris discharges givenby Bishop et al (1995) for Batura Glacier (Pakistan) 48to 90times 103 m3 yrminus1 the thickening of the debris can be es-timated between 27 and 55 mm yrminus1 which is negligiblegiven the magnitude of the glacier elevation changes

53 Comparison to previous mass balance estimates

Throughout the PKH there is only a single peer-reviewedglaciological record (on Chhota Shigri Glacier) coveringmost of our study period (Wagnon et al 2007 Vincent etal 2013) Our geodetic mass balance estimate for the SpitiLahaul study site has been recently compared to those glacio-logical measurements on Chhota Shigri Glacier (Vincent etal 2013) Mass balance records reported in Yao et al (2012)for the Tibetan Plateau and the surroundings mountain ranges

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1276 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 9Same as Fig 8 but for the Karakoram west study site

only start during glaciological year 20052006 This is whyhere we do not compare our mass balance estimates toglaciological records and limit our comparison to others re-gional estimates obtained using the geodetic or gravimetricmethods

We stress that the comparison of our mass balance mea-surements to published estimates is complicated by thefact that generally they do not cover the same time pe-riod Little is known about the inter-annual variability inthe PKH during the 21st century due to the limited num-ber of long glaciological time series The 9 yr mass balance

record on Chhota Shigri Glacier reveals a relatively highinter-annual variability of the annual mass balance (stan-dard deviationplusmn 074 m we yrminus1 during 2003ndash2011 Vin-cent et al 2013) which is similar to the one obtained be-tween 1968 and 1998 on Abramov Glacier (standard devi-ation 069 m we yrminus1 WGMS 2012) Thus inter-annualvariability can explain part of the difference between two cu-mulative mass balance estimates over 5ndash10 yr even if they areoffset by only one or two years

Based on ICESat repeat track measurements and theSRTM DEM Kaumlaumlb et al (2012) measured a region-wide

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1277

Fig 10Same as Fig 8 but for the Pamir study site

mass balance ofminus021plusmn 005 m we yrminus1 between 2003and 2008 for Hindu Kush-Karakoram-Himalaya glaciersFor that same area (ie excluding the Pamir and theHengduan Shan study sites) we found a mass balance ofminus015plusmn 007 m we yrminus1 Thus our result agrees with Kaumlaumlbet al (2012) within the error bars although our study periodis longer (1999ndash2011) and our measurement principle andextrapolation approach are different In addition we havecalculated updated mass balances from Kaumlaumlb et al (2012)ie averages over 3times 3 degree cells centered over our studysites and find that they are consistent within error barswith our mass balance estimates (Table 5) Similar resultsare found when our mass balances are compared to a spa-tially extended ICESat analysis for 2003ndash2009 (Gardner etal 2013) Mass losses measured with ICESat (Kaumlaumlb et al2012 Gardner et al 2013) tend to be slightly larger thanours This could be due to the difference in time periodsAlso our smaller mass losses could be partly explained if wesystematically underestimated the poorly constrained pene-

tration of the C-band andor X-band radar signal into ice andsnow

At a more local scale Bolch et al (2011) reported a massbalance ofminus079plusmn 052 m yrminus1 between 2002 and 2007 byDEM differencing over a glacier area of 62 km2 includingten glaciers south and west of Mt Everest For these sameglaciers Nuimura et al (2012) found a mean mass balanceof minus045plusmn 060 m yrminus1 during 2000ndash2008 while they com-puted a mass balance ofminus040plusmn 025 m yrminus1 between 1992and 2008 over a glacier area of 183 km2 around Mt EverestIn the present study we found an average mass balance ofminus041plusmn 021 m yrminus1 over the ten glaciers mentioned above(Table 5 Fig S1) and a value ofminus026plusmn 013 m yrminus1 overthe whole Everest study site between 2000 and 2011 Ourvalues are comparable to those of Nuimura et al (2012)Our mass balance is less negative than the 2002ndash2007 massbalance of Bolch et al (2011) but not statistically differentwhen error bars are considered (Fig 12) Indeed Bolch etal (2011) acknowledged some high uncertainties for their

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1278 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Table 4Glacier mass budget of the nine sub-regions to which the results of the study sites (mass balances) have been extrapolated For eachof them the total glacierized area as well as the percentage of the glacier area over which the mass balance was computed (ie study sites)are given The total glacier area is derived from the RGI v20 (Arendt et al 2012) The error for the mass balance in each sub-region is theroot of sum of squares (RSS) of the mass balance error of each study site and the regional extrapolation error The error for the mass budgetin each sub-region includes additionally the error from the total glacier area in the RGI v20

Sub-region Total glacier Measured glacier Mass balance Mass budgetarea (km2) area () (m we yrminus1) (Gt yrminus1)

Hengduan Shan 11584 12 minus033plusmn 014 minus38plusmn 38Bhutan 4021 34 minus022plusmn 013 minus09plusmn 05Everest 6226 23 minus026plusmn 014 minus16plusmn 08West Nepal 6849 13 minus032plusmn 014 minus22plusmn 10Spiti Lahaul 9043 23 minus045plusmn 014 minus41plusmn 13Hindu Kush 6135 13 minus012plusmn 016 minus07plusmn 10Karakoram East

1902428 +011 plusmn 014 +19 plusmn 31

Karakoram West 30 +009 plusmn 018Pamir 9369 34 +014 plusmn 014 +13 plusmn 13

TotalArea- 72251 30 minus014plusmn 008 minus101plusmn 55weighted average

Table 5Comparison of geodetic mass balances between this study and previous published results with overlapping study periods and similargeographic locations Updated figures from Kaumlaumlb et al (2012) are averaged over 3 times 3 cells centered over the study sites of the presentstudy

GlacierSite name This study Bolch et al (2011) Nuimura et al (2012) Kaumlaumlb et al (2012 updated)2002ndash2007 2000ndash2008 2003ndash2008

Study Mass balance Area (km2)a Mass balance Area Mass balance Mass balanceperiod (m we yrminus1) (m we yrminus1) (km2) (m we yrminus1) (m we yrminus1)

Hengduan San 1999ndash2010 minus022 plusmn 014 1303 (55 )Bhutan 1999ndash2010 minus022 plusmn 012 1384 (64 ) minus052 plusmn 016b

Everest

1999ndash2011

minus026 plusmn 013 1461 (58 ) minus039 plusmn 011AX010 minus090plusmn 034 04Changri SharNup minus042plusmn 017 161 (79 ) minus029plusmn 052 130 minus055plusmn 038Khumbu minus051plusmn 019 203 (47 ) minus045plusmn 052 170 minus076plusmn 052Nuptse minus037plusmn 020 50 (74 ) minus040plusmn 053 40 minus034plusmn 027Lhotse Nup minus021plusmn 027 24 (70 ) minus103plusmn 051 19 minus022plusmn 047Lhotse minus043plusmn 018 85 (83 ) minus110plusmn 052 65 minus067plusmn 051Lhotse SharImja minus070plusmn 052 98 (44 ) minus145plusmn 052 107 minus093plusmn 060Amphu Laptse minus046plusmn 034 25 (38 ) minus077plusmn 052 15 minus018plusmn 094Chukhung +044 plusmn 024 42 (47 ) +004 plusmn 054 38 +043 plusmn 081Ama Dablam minus049plusmn 017 36 (54 ) minus056plusmn 052 22 minus056plusmn 073Duwo minus016plusmn 026 19 (18 ) minus196plusmn 053 10 minus068plusmn 074Total Khumbu minus041plusmn 021 744 minus079plusmn 052 617 minus045plusmn 060(10 Glaciers above)

West Nepal 1999ndash2011 minus032 plusmn 013 908 (40 ) minus032 plusmn 012Spiti Lahaul 1999ndash2011 minus045 plusmn 013 2110 (46 ) minus038 plusmn 006Karakoram East 1999ndash2010 +011 plusmn 014 5328 (42 ) minus004 plusmn 004Karakoram West 1999ndash2008 +009 plusmn 018 5434 (45 )Hindu Kush 1999ndash2008 minus012 plusmn 016 793 (80 ) minus020 plusmn 006Pamir 1999ndash2011 +014 plusmn 013 3178 (50 )

a The total glacier area is given in km2 and in parenthesis the of the glacier area actually covered with measurementsb For this cell the ICESat coverage is insufficient and does not sample all glacier elevations

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1279

Fig 11 Elevation changes (SPOT5-SRTM) in the ablation area for clean ice (blue circles) and debris-covered ice (black circles) for eachstudy site For the Karakoram (east and west) and the Pamir study sites only non-surge-type glaciers are considered Standard error of theelevation differences are typically less thanplusmn2 m with each 100 m elevation band (not shown for the sake of clarity) Note elevation changesover separate sections of a glacier cannot be treated as mass changes due to the disregard of glacier dynamics

estimate over this short time period The differences in sur-vey periods and in the glacier outlines may also explainpart of these discrepancies In particular the delimitation ofthe accumulation area is notoriously difficult and Bolch etal (2011) did not include some of the steep high altitudeslopes which we included due to their potential avalanchecontribution For the 10 glaciers mentioned above our massbalances would become 012 m we yrminus1 more negative ifwe used the exact same outlines as Bolch et al (2011)Our positive mass balance for the small Chukhung Glacier(+044plusmn 024 m we yrminus1) is in agreement with Nuimura etal (2012) but enigmatic in a context of regional glacier lossand deserves further investigation If the 2002ndash2007 massbalance estimate by Bolch et al (2011) is excluded all otherrecent mass balance estimates suggest no acceleration of themass loss in the Everest area when compared to the long-term(1970ndash2007) mass balance measured by Bolch et al (2011)in this regionminus032plusmn 008 m weyrminus1

In Bhutan our results indicate a stronger mass loss forsouthern glaciers than for northern ones Interestingly in thisarea Kaumlaumlb (2005) measured high speeds (up to 200 m yrminus1)

for glaciers north of the Himalayan main ridge and ratherlow velocities over southern glacier tongues by using repeat

ASTER data This is consistent with the more negative massbalance that we report for the southern glaciers in Bhutanie for the slow flowing presumably downwasting glaciers

Berthier et al (2007) reported a mass balance betweenminus069 andminus085 m we yrminus1 for an ice-covered area of915 km2 around Chhota Shigri Glacier between 1999 (SRTMDEM) and 2004 (SPOT5-HRG DEM) For the same areawe found a mass balance ofminus045plusmn 013 m we yrminus1 over1999ndash2011 The difference in the study period may explainpart of the differences Also the methodology by Berthieret al (2007) did not explicitly take into account the SRTMradar penetration over glaciers and also corrected an appar-ent elevation-dependent bias according to glacier elevationswhich is now known as a resolution issue and is best cor-rected according to the terrain maximum curvature (Gardelleet al 2012b) More details and discussions about these dif-ferences can be found in Vincent et al (2013)

Importantly the results of the eastern Karakoram studysite confirm the balanced or slightly positive mass budgetreported by Gardelle et al (2012a) in the western Karako-ram over the past decade The two Karakoram study sitesare adjacent with a small overlapping area (sim 1100 km2 ofglaciers) where elevation differences agree within 1 m Both

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1280 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 12 Comparison of geodetic mass balances for 10 glaciers inthe Mount Everest area 1999ndash2010 mass balances measured inthe present study are compared with published estimates during2002ndash2007 (Bolch et al 2011) and during 2000ndash2008 (Nuimuraet al 2012) The 1-to-1 line is drawn The mean difference (plusmn1standard deviation) between (i) our values and Bolch et al (2011)are (047plusmn 058 m we yrminus1) and (ii) our values and Nuimura etal (2012) are (012plusmn 021 m we yrminus1) Part of the differences be-tween estimates may be due to the different periods surveyed andthe different glacier outlines used

sites show similar mass balances over slightly different pe-riods +011plusmn 014 m yrminus1 over 1999ndash2010 in the east and+009plusmn 018 m yrminus1 over 1999ndash2008 in the west This con-firms a ldquoKarakoram anomalyrdquo that was first suggested basedon field observations (Hewitt 2005) We report a mass bal-ance of+010plusmn 013 m yrminus1 for Baltoro Glacier (see loca-tion in Fig 8 size= 304 km2) between 1999 and 2010which is consistent with its gradual speed-up reported since2003 (Quincey et al 2009) However given the completelydifferent methods used in our and the latter study we can-not conclude that this demonstrates a real shift from nega-tive to positive mass balance The mass budget of SiachenGlacier (1145 km2 sometimes referred to as the highestbattlefield in the world) is also balanced or slightly pos-itive (+014plusmn 014 m we yrminus1) Thus the alarming state-ment by Rasul (2012) that this glacier retreated by 59 kmand lost 17 of its mass during 1989ndash2009 is very likelyerroneous Again within error bars our regional mass bal-ance in the Karakoram agrees with the mass balance ofminus003plusmn 004 m yrminus1 found by Kaumlaumlb et al (2012) over 2003ndash2008 and with the mass balance ofminus010plusmn 018 m we yrminus1

(2-sigma error level) obtained by Gardner et al (2013) for aregion including both the Karakoram and the Hindu Kush

In the northernmost part of the Pamir in the Alay moun-tain range the mass balance of Abramov Glacier has beenmeasured in the field for 30 yr between 1968 and 1998(WGMS 2012) with a mean value ofminus046 m yrminus1 Herewe report a mass balance ofminus003plusmn 014 m yrminus1 for this

glacier over 1999ndash2011 Our near zero mass budget estimatefor this glacier disagrees with negative mass balances recon-structed for the same period with a model run using biascorrected NCEPNCAR re-analysis data and calibrated withfield data (Barundun et al 2013) An underestimate of ourSRTM C-Band penetration in the Pamir as a whole and inparticular for the small Abramov Glacier may contribute tothese differences

For the whole Pamir study site our mass budget is bal-anced or slightly positive (+014plusmn 013 m we yrminus1) In theeastern Pamir the mass balance of Muztag Ata Glacier was+025 m we yrminus1 between 2005 and 2010 (Yao et al 2012)Although Muztag Ata Glacier is located outside of our Pamirstudy site and his glaciological records only cover the sec-ond half of our study period this measurement is consistentwith our remotely sensed mass balance Thus the ldquoKarako-ram anomalyrdquo should probably be renamed the ldquoPamir-Karakoram anomalyrdquo

This slightly positive or balanced mass budget measuredin the Pamir seems to contradict previous findings by Heidand Kaumlaumlb (2012) They reported rapidly decreasing glacierspeeds in this region between two snapshots of annual ve-locities in 20002001 and 20092010 an observation thatwould fit with negative mass balances (Span and Kuhn 2003Azam et al 2012) We note though that the relationship be-tween glacier mass balance and ice fluxes is not straightfor-ward and might in particular involve a time delay betweena change in mass balance and the related dynamic response(Heid and Kaumlaumlb 2012) Thus the decrease in speed couldwell be due to negative mass balances before or partiallybefore 1999ndash2011 We acknowledge that this discrepancyneeds to be investigated further in particular by examiningcontinuous changes in annual glacier speed through the firstdecade of the 21st Century Indeed Heid and Kaumlaumlb (2012)measured differences between two annual periods which maynot represent mean decadal velocity changes

Jacob et al (2012) using satellite gravimetry (GRACE)measured a mass budget ofminus5plusmn 6 Gt yrminus1 over 2003ndash2010for the ldquoKarakoram-Himalayardquo their region 8c which cor-responds to our study area excluding the Pamir Karakoramand Hindu Kush sub-regions but including some glaciersnorth of the Himalayan ridge already on the Tibetan PlateauFor this subset of our PKH study area we measured amass budget ofminus126plusmn 42 Gt yrminus1 The two estimates over-lap within their error bars especially if our error is dou-bled to match the two standard error level used by Jacob etal (2012) The estimate of Jacob et al (2012) over our com-plete study region (PKH) ieminus6plusmn 8 Gt yrminus1 (sum of theirregions 8b and 8c) also includes mass changes for the KunlunShan sub-region for which we did not measure glacier massbalances Therefore the comparison with our PKH value(minus101plusmn 55 Gt yrminus1) is not straightforward but suggests areasonable agreement especially if we consider the slightmass gain (15plusmn 17 Gt yrminus1) measured with ICESat in theWest Kunlun (Gardner et al 2013)

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1281

Table 6 Annual average of the glacier seasonal and decadal mass loss contributions to Indus Ganges and Brahmaputra discharges Basinarea and seasonal contributions are given by Kaser et al (2010) glacier area in each basin is derived from the RGI v20 (Arendt et al 2012)Numbers in parentheses indicate the percentage of glacierized area within the river basin Seasonal contributions were modeled by Kaser etal (2010) based on the CRU 20 dataset over 1961ndash1990 assuming a balanced glacier mass budget Glacier decadal mass loss contributionsare given as annual average over 1999ndash2011

River Basin area Glacier area Glacier contribution Mean discharge(km2) (km2) (m3 sminus1) (m3 sminus1)

Seasonally delayed Decadal mass lossKaser et al (2010) (this study)

Indus 1 139 814 25 598 (2 ) 105 103plusmn 72 2146 (Chevallier personalcommunication 2012)

Ganges 1 023 609 11 168 (1 ) 47 103plusmn 49 11739 Papa et al (2012Brahmaputra 527 666 15 296 (3 ) 33 147plusmn 64 19710 updated)

54 Reasons for the heterogeneous pattern of massbalance

The PKH glacier mass balance is slightly negative(minus014plusmn 008 m we yrminus1) but changes are not homoge-neous from one sub-region to another and seem to coincidewith the climatic settings of the mountain range For the east-ern sites (from the Hengduan Shan to West Nepal) which areunder the influence of the Indian and south-east Asian sum-mer monsoons the mass balances are moderately negative(sim minus026 m we yrminus1) In the northwest of the PKH at thePamir and Karakoram sites where glacier climate is char-acterized by the westerlies in winter mass balances are bal-anced or slightly positive (sim +010 m we yrminus1) In betweenthese two influences the glaciers of the Spiti Lahaul sitein a monsoon-arid transition zone experienced the strongestmass loss (minus045plusmn 013 m we yrminus1)

This heterogeneous pattern of mass balances seems to berelated to trends in precipitation throughout the PKH Archerand Fowler (2004) reported an increase in winter precipi-tation over the Upper Indus Basin since 1961 a trend con-firmed by Yao et al (2012) over the Pamir and the Karako-ram based on the analysis of the Global Precipitation Cli-matology Project data set (GPCP Adler et al 2003) Thisprecipitation increase is a source of greater accumulation forglaciers and thus a possible explanation for their slightly pos-itive mass budgets In addition in the Upper Indus Basin themean summer temperature has been decreasing since 1961(Fowler and Archer 2006) which is consistent with the find-ings of Shekhar et al (2010) in the Karakoram who reporteda decrease in both maximum and minimum temperature since1988 These increases in winter precipitation and summertemperature likely translate in greater accumulation and re-duced ablation changes which are consistent with the bal-anced or slightly positive glacier mass budget

On the other hand in the east of the PKH the Indiansummer monsoon has been weakening since the 1950s (Bol-lasina et al 2011) which could have reduced the amount

of snowfall over the eastern study sites and thus resultedin negative mass balances In addition maximum and min-imum temperatures increased since 1988 in the western Hi-malaya (Shekhar et al 2010) and maximum temperaturesincreased in the central Himalaya (Nepal) since the mid-1970s (Shrestha et al 1999) Thus both precipitation andtemperature trends in the Himalaya likely contributed to thenegative mass balances measured over the correspondingstudy sites (from Spiti Lahaul to Hengduan Shan Fig 1)

However gridded data sets such as GPCP are not neces-sarily representative of the climate at high altitude IndeedHewitt (2005) reported a 5-to-10-fold increase in precipi-tation between the glacier front and the accumulation areain the Karakoram that is not captured by reanalysis dataas recently confirmed by Immerzeel et al (2012) Thus di-rect measurements of accumulation on High Mountain Asiaglaciers (eg Azam et al 2013) and high-altitude weatherstations are eagerly needed to validate the large scale grid-ded data set (eg reanalysis) test their ability to describe thespecific climate near to PKH glaciers and assess the relativerole played by temperature and precipitation changes

55 Contribution to water resources

Our glacier mass balance can be averaged over large riverbasins to estimate the mean contribution to river runoffdue to a decade of glacier mass loss We assume therebyonly direct river runoff without loss terms For the threerivers for which we cover the entire glacierized area of theircatchments within our sub-regions (Fig 1) these contribu-tions to the river discharge are equal to 103 m3 sminus1 (Indus)103 m3 sminus1 (Ganges) and 147 m3 sminus1 (Brahmaputra) for theperiod 1999ndash2011 (Table 6)

Rivers of PKH being key water resources measures oftheir discharges are highly sensitive and difficult to access forpolitical reasons Papa et al (2012) built recent time series ofdischarges for Ganges and Brahmaputra by using altimetrymeasurements The locations of the corresponding dischargeestimates in Bangladesh are given in Fig 1 We can therefore

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1282 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

evaluate the relative contribution of glacier decadal mass loss(Table 6) to annual river run-off at 09 for the Gangesat Hardinge (basin area ofsim 1 020 000 km2) and 07 forBrahmaputra at Bahadurabad (basin area ofsim 530 000 km2)between 1999 and 2011 Given the discharge measured at theGuddu station by the Surface Water Hydrology Project of theWater and Power Development Authority (Pakistan see lo-cation on Fig 1) between 1999 and 2003 we can estimatethe contribution of glacier mass loss at 48 for the IndusRiver at this station

The seasonal contributions of glaciers to river run-off (iethe part of the annual precipitation that experiences season-ally delayed release from the glaciers) directly results fromseasonal variations of meteorological quantities in contrastto the glacier mass loss contribution which results from thelong-term climatic change Both runoff components directlyaffect the amount of water available in a river at a given lo-cation and at a given time so that their comparison could beof interest Even if the seasonal timing of the glacier massloss contribution is unclear it could in general peak at asimilar time of the year than the seasonal glacier contribu-tion so that both components rather enlarge each other thancancel Seasonal glacier contributions have been modeled byKaser et al (2010) The latter study assumed glaciers to bein equilibrium with the present climate and thus prescribedglacier mass balances equal to 0 Thus their seasonally de-layed glacier contributions do not include the part due todecadal glacier mass loss and is complementary to our es-timates (Table 6) For the eastern basins under the influenceof the Indian summer monsoon (Ganges and Brahmaputra)the contribution of glacier mass loss is higher than the sea-sonal contribution In other words for the first decade of the21st century the mass loss of glaciers is on average moreimportant for water availability in those two rivers than theirseasonal water storage effect The situation is different for theIndus Basin where the glacier melt season is also the dry sea-son which results in a relatively higher seasonally delayedglacier contribution to the river discharge There the season-ally delayed water release from the glaciers and the decadalglacier mass loss contribute on average equally to the riverdischarge Both contributions are strongly (and equally) de-pendent on the location of the discharge measurement andwill be significantly larger (in relative term) in more heav-ily glacierized and smaller mountain catchments located inthe upper part of these major river basins (Bookhagen andBurbank 2010)

6 Conclusion

In this study we assessed the spatial pattern of glaciermass balance over the PKH by comparing the SRTM andSPOT5 DEMs over 9 well-distributed study sites between1999 and 2011 We found slightly positive or balanced massbudgets in the western part (Pamir Karakoram) moder-ate mass loss in the east (Nepal Bhutan Hengduan Shan)

and the most negative mass balance in the monsoon-aridtransition zone (Spiti Lahaul in the western Himalaya)By extrapolating these values to climatically homogeneoussub-regions we estimate the PKH overall mass balanceto have beenminus014plusmn 008 m we yrminus1 between 1999 and2011 which corresponds to a contribution to sea level riseof 0028plusmn 0015 mm yrminus1 This is about 4 of the globalglacier mass loss estimated by Gardner et al (2013) thoughover a slightly shorter time period (2003ndash2009) The largestsource of uncertainties in those geodetic mass balance esti-mates is the poorly constrained penetration of the C-BandSRTM radar signal into snow and ice

Our technique of DEM differencing allows mapping in de-tail the spatial pattern of glacier elevation changes Thosemaps reveal many surge-type behaviors both in the Pamirand the Karakoram It also appears that the glacier-wide massbalance does not seem to be considerably affected by themass transfer from surges over thesim 10 yr time period stud-ied Similar lowering rates over debris-covered and clean iceare observed for four study sites despite the commonly as-sumed insulating effect of debris covers This suggests forthe scale of entire glacier tongues a spatial mixture of re-duced ablation under intact thick debris covers and enhancedablation by supraglacial lakes or ice cliffs or due to thin de-bris layer and very low glacier velocities in ablation areasthat reduce the ice advection downstream

The regionally heterogeneous pattern of ice wastage is inbroad agreement with contrasted trends in large-scale pre-cipitation and temperature However glaciological (eg ac-cumulation) and meteorological records are needed at highaltitude in the vicinity of glaciers to evaluate more closelythe local influence of atmospheric variables on glacier massbalance and fully understand the reasons behind those con-trasted mass budgets in the PKH

Supplementary material related to this article isavailable online at httpwwwthe-cryospherenet712632013tc-7-1263-2013-supplementpdf

Acknowledgements We thank G Cogley A Gardner T NuimuraT Bolch P Wagnon M Barandun M Hoelzle M Huss and ananonymous referee for their constructive comments that led to aconsiderably improved manuscript We thank the Water and PowerDevelopment Authority (WAPDA) for their hydrological data(processed by P Chevallier) and F Papa for sharing his updatedrun-off data ahead of publication We thank T Bolch for sharinghis glacier outlines from the Everest area We thank the USGSfor allowing free access to their Landsat archive CGIAR-SCI forSRTM C-band data DLR for SRTM X-band data and GLIMS forglacier inventories J Gardelle acknowledges a PhD fellowshipfrom the French Space Agency (CNES) and the French NationalResearch Center (CNRS) E Berthier acknowledges support fromCNES through the TOSCA and ISIS proposal no 397 and fromthe Programme National de Teacuteleacutedeacutetection Spatiale E Berthieracknowledges M Masiokas and R Villalba for welcoming him

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1283

at the IANIGLA (Mendoza Argentina) Y Arnaud acknowledgessupport from French National Research Agency through ANR-09-CEP-005-01PAPRIKA A Kaumlaumlb acknowledges support fromESA through Glaciers_cci (400010177810I-AM) and from theEuropean Research Council under the European Unionrsquos SeventhFramework Programme (FP2007-2013)ERC Grant Agreementn 320816

Edited by I M Howat

The publication of this articleis financed by CNRS-INSU

References

Adler R Huffman G Chang A Ferraro R Xie P JanowiakJ Rudolf B Schneider U Curtis S Bolvin D Gruber ASusskind J Arkin P and Nelkin E The Version-2 GlobalPrecipitation Climatology Project (GPCP) Monthly Precipita-tion Analysis (1979ndashPresent) J Hydrometeorol 4 1147ndash11672003

Archer D R and Fowler H J Spatial and temporal variationsin precipitation in the Upper Indus Basin global teleconnectionsand hydrological implications Hydrol Earth Syst Sci 8 47ndash61doi105194hess-8-47-2004 2004

Arendt A Bolch T Cogley J G Gardner A Hagen J-OHock R Kaser G Pfeffer W T Moholdt G Paul F RadicV Andreassen L Bajracharya S Beedle M Berthier EBhambri R Bliss A Brown I Burgess E Burgess DCawkwell F Chinn T Copland L Davies B de AngelisH Dolgova E Filbert K Forester R Fountain A Frey HGiffen B Glasser N Gurney S Hagg W Hall D Hari-tashya U K Hartmann G Helm C Herreid S Howat IKapustin G Khromova T Kienholz C Koenig M KohlerJ Kriegel D Kutuzov S Lavrentiev I LeBris R Lund JManley W Mayer C Miles E Li X Menounos B Mer-cer A Moelg N Mool P Nosenko G Negrete A NuthC Pettersson R Racoviteanu A Ranzi R Rastner P RauF Rich J Rott H Schneider C Seliverstov Y Sharp MSigurethsson O Stokes C Wheate R Winsvold S WolkenG Wyatt F and Zheltyhina N Randolph Glacier Inventory[v20] A Dataset of Global Glacier Outlines Global Land IceMeasurements from Space Boulder Colorado USA Digital Me-dia httpwwwglimsorgRGIrandolphhtml 2012

Azam M Wagnon P Ramanathan A Vincent C Sharma PArnaud Y Linda A Pottakkal J Chevallier P Singh V andBerthier E From balance to imbalance a shift in the dynamicbehaviour of Chhota Shigri glacier western Himalaya India JGlaciol 58 315ndash324 doi1031892012JoG11J123 2012

Azam M F Wagnon P Vincent C Ramanathan A Linda Aand Singh V B Reconstruction of the annual mass balanceof Chhota Shigri Glacier (Western Himalaya India) since 1969Ann Glaciol in revision 2013

Barrand N and Murray T Multivariate Controls on the In-cidence of Glacier Surging in the Karakoram Himalaya

Arct Antarct Alp Res 38 489ndash498 doi1016571523-0430(2006)38[489MCOTIO]20CO2 2006

Barundun M Huss M Azisov E Gafurov A HoelzleM Merkushkin A Salzmann N and Usubaliev R Re-establishing seasonal mass balance observation at AbramovGlacier Kyrgyzstan from 1968ndash2012 Abstract EGU2013-5668European Geophysical Union Vienna 2013

Benn D Bolch T Hands K Gulley J Luckman ANicholson L Quincey D Thompson S Toumi R andWiseman S Response of debris-covered glaciers in theMount Everest region to recent warming and implicationsfor outburst flood hazards Earth-Sci Rev 114 156ndash174doi101016jearscirev201203008 2012

Berthier E and Vincent C Relative contribution of surface mass-balance and ice-flux changes to the accelerated thinning of Merde Glace French Alps over 1979ndash2008 J Glaciol 58 501ndash512doi1031892012JoG11J083 2012

Berthier E Arnaud Y Baratoux D Vincent C and Reacutemy FRecent rapid thinning of the ldquo Mer de Glace rdquo glacier derivedfrom satellite optical images Geophys Res Lett 31 L17401doi1010292004GL020706 2004

Berthier E Arnaud Y Kumar R Ahmad S Wagnon P andChevallier P Remote sensing estimates of glacier mass balancesin the Himachal Pradesh (Western Himalaya India) RemoteSens Environ 108 327ndash338 doi101016jrse2006110172007

Bhambri R Bolch T Kawishwar P Dobhal D P SrivastavaD and Pratap B Heterogeneity in Glacier response from 1973to 2011 in the Shyok valley Karakoram India The CryosphereDiscuss 6 3049ndash3078 doi105194tcd-6-3049-2012 2012

Bhutiyani M Mass-balance studies on Siachen Glacier in theNubra valley Karakoram Himalaya India J Glaciol 45 112ndash118 1999

Bishop M P Schroder J F and Ward J L SPOT multispec-tral analysis for producing supraglacial debris-load estimates forBatura glacier Pakistan Geocarto Int 10 81ndash90 1995

Bolch T Pieczonka T and Benn D I Multi-decadal mass lossof glaciers in the Everest area (Nepal Himalaya) derived fromstereo imagery The Cryosphere 5 349ndash358 doi105194tc-5-349-2011 2011

Bolch T Kulkarni A Kaumlaumlb A Huggel C Paul F Cogley JFrey H Kargel J Fujita K Scheel M Bajracharya S andStoffel M The state and fate of Himalayan Glaciers Science336 310ndash314 doi101126science1215828 2012

Bollasina M Ming Y and Ramaswamy V Anthropogenicaerosols and the weakening of the South Asian summer mon-soon Science 334 502ndash505 doi101126science12049942011

Bookhagen B and Burbank D Toward a complete Himalayan hy-drological budget spatiotemporal distribution of snowmelt andrainfall and their impact on river discharge J Geophys Res115 F03019 doi1010292009JF001426 2010

Bretherton C Widmann M Dymnikov V Wallace J and BladeacuteI The effective number of spatial degrees of freedom of a time-varying field J Climate 12 1990ndash2009 1999

Copland L Pope S Bishop M Shroder J Clendon PBush A Kamp U Seong Y and Owen L Glacier veloc-ities across the central Karakoram Ann Glaciol 50 41ndash49doi103189172756409789624229 2009

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1267

Table 2 Elevation ranges equilibrium line altitudes (ELA) and accumulation area ratios (AAR) for the nine study sites The dates ofacquisition of Landsat images used for ELA determination are also given

Sub-region Glaciers elevation ELA (m) AAR Landsat imagesrange (m)

Min Max This study Scherler et Kaumlaumlb et This study Kaumlaumlb et PathRow Dateal (2011a) al (2012) al (2012)

Hengduan Shan 2630 6860 4970plusmn 320 NA NA 55 NA p135r39 23 Sep 1999Bhutan 4390 7480 5690plusmn 440

5700 555036

47p138r40 17 Nov 2000

Everest 4260 8380 5840plusmn 320 31 p140r41 31 oct 2009West Nepal 3950 6870 5590plusmn 138 37 p143r39 1 Aug 2009Spiti Lahaul 3780 6360 5390plusmn 140 5103 5500 34 35 p147r37 2 Aug 2002

p147r38Hindu Kush 2717 6385 5050plusmn 160 5138 30 p150r34 2 Sep 2000

p150r35Karakoram east 3230 7770

5030plusmn 280 4845 554076

47p148r35 7 Sep 1998

Karakoram west 2910 7920 66 p149r35 29 Aug 1998Pamir 2800 7090 4580plusmn 250 NA NA 52 NA p152r33 30 Jul 2000

growth of connected lakes and for glaciers that surged andthus advanced For the few glaciers that advanced their frontposition was updated manually based on the SPOT5 ortho-image

For each study site we favored our user-defined glacierinventories instead of the global Randolph Glacier Inventory(RGI v20 Arendt et al 2012) because the latter is hetero-geneous and for some study sites did not match our accu-racy requirements for the adjustments and corrections of theDEMs

We also estimated the altitude of the equilibrium line(ELA) as the snowline on thesim 2000 Landsat data with min-imal snow cover to separate ablation and accumulation ar-eas The ELA is assumed to be identical to the altitude ofthe snowline at the end of the ablation season (eg Raba-tel et al 2005) after the end of the monsoons or before thefirst snowfalls Therefore we manually digitized snowlinesfor more than 30 glaciers for each study site and computedtheir mean elevations Our ELAs are slightly different fromthose given by Scherler et al (2011a) and Kaumlaumlb et al (2012)probably due to the sensitivity of the ELA to the choice ofthe image (Table 2) Ideally the ELA should have been es-timated from images acquired at the end of the ablation sea-son during various years covering our whole study periodThis would be a highly time-consuming task that was not ac-complished here However we stress that in our study theELA is only used to compare elevation changes on the abla-tionaccumulation areas (Sect 41) and to estimate the sen-sitivity of our mass balances to the choice of the volume-to-mass conversion factor (Sect 35) Our preferred mass bal-ance estimate uses a unique volume-to-mass conversion fac-tor for the whole glacier and is thus independent from theELA

33 Adjustments and corrections of DEM biases

Different DEM processing steps are followed to extract un-biased glacier elevation changes for each study site (eg Nuthand Kaumlaumlb 2011 Gardelle et al 2012b)

331 Planimetric adjustment of the DEMs

A horizontal shift between two DEMs results in an aspect-dependent bias of elevation differences (Nuth and Kaumlaumlb2011) Here the horizontal shift is determined by minimizingthe root mean square error of elevation differences (SPOT5-SRTM) off glaciers where the terrain is assumed to be stableover the study period (Berthier et al 2007) The SRTM DEMis then resampled (cubic convolution) according to the shift

332 Alongacross track corrections (drift of thesatellite orbit)

For DEMs derived from stereo imagery the satellite acquisi-tion geometry can induce a bias along andor across the trackdirection (Berthier et al 2007) We estimate the azimuth ofeach SPOT5 ground track and use it to rotate the coordinatesystem accordingly (Nuth and Kaumlaumlb 2011) Then we com-pute the elevation differences along and across the satellitetrack on stable areas and when necessary correct the bias us-ing a 5th order polynomial fit

333 Curvature correction

As suggested by Paul (2008) and Gardelle et al (2012b) thedifference in original spatial resolutions of the two DEMs canlead to biases related to altitude in mountainous areas Whenthe curvature the first derivative of the slope is high (sharppeaks or ridges) the altitude tends to be underestimatedby the coarse DEM unable to reproduce high-frequency

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1268 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

slope variations Thus this apparent ldquoelevation biasrdquo is cor-rected using the relation between elevation differences andthe maximum terrain curvature estimated over stable areasoff glaciers (Gardelle et al 2012b) Note that the units inFig 1b in Gardelle et al (2012b) should be 10minus2 mminus1 in-stead of mminus1 (A Gardner personal communication 2012)However this does not impact the validity of the correction

334 Radar penetration correction

The SRTM DEM acquired in C-band potentially underesti-mates glacier elevations since at this wavelength (sim 56 cm)the penetration of the radar signal into snow and ice canreach several meters (eg Rignot et al 2001) Here we esti-mate this penetration for each study site by differencing theSRTM C-band (57 GHz) and X-band (97 GHz) DEMs Thelatter was acquired simultaneously with the SRTM X-bandat higher resolution (30 m) but with a narrower swath andthus has a coverage restricted to selected swaths only Sincethe penetration of the X-band is expected to be low com-pared to the C-band (an hypothesis that still needs to be con-firmed Ulaby et al 1986) we consider the elevation differ-ence (SRTMC-bandminus SRTMX-band) to be a first approxima-tion of C-band radar penetration over glaciers (Gardelle etal 2012b) The correction is calculated as a function of alti-tude and applied to each pixel separately

For the Hindu Kush Spiti Lahaul and West Nepal sitesdue to the lack of SRTMX-band data we were unable to de-termine the value of the penetration Thus we applied thecorrection computed over the nearest study site ie Karako-ram for Spiti Lahaul and Hindu Kush and Everest for WestNepal The mean values of the SRTMC-bandpenetrations overglaciers are given for each study site in Table 3

335 Seasonality correction

SPOT5 DEMs were acquired between late October and earlyJanuary depending on the study site Since the SRTM DEMis from mid-February we need to account for possible masschanges during this 1-to-5-month period in order to esti-mate glacier mass balance over an integer number of years(Gardelle et al 2012a) For the eastern sites (from WestNepal to Hengduan Shan) where glaciers are of summer-accumulation types (Fujita 2008) the correction was set to0 This choice is confirmed by recent field observations thatshow no winter accumulation on Mera and Pokalde glaciersin Nepal (Wagnon et al 2013) For the Karakoram studysites we used the winter accumulation rate+013 m we permonth measured on Biafo Glacier (Wake 1989) For theother western study sites (the Pamir Hindu Kush and SpitiLahaul sites) the correction is derived from the mean of win-ter mass balances of 35 glaciers all in the Northern Hemi-sphere+089 m we yrminus1 over 2000ndash2005 corresponding to+015 m we per winter month (Ohmura 2011) The lattervalue is slightly lower than the average winter mass balance

Table 3 Average SRTM C-band penetration estimates (in m)in February 2000 over glaciers in this study and from Kaumlaumlb etal (2012)

Sub-region This study Kaumlaumlb et al (2012)

Hengduan Shan 17 NABhutan 24

25plusmn 05Everest 14West Nepal NA

15plusmn 04Spiti Lahaul NAHindu Kush NA 24plusmn 04Karakoram East

34 24plusmn 03Karakoram WestPamir 18 NA

(+135plusmn 031 m we yrminus1) measured on Abramov Glacier inthe Pamir during 1968ndash1998 (WGMS 2012) and in reason-ably good agreement with the modeled mean winter massbalance (+108plusmn 034 m we yrminus1) of Chhota Shigri Glacierduring 1969ndash2012 in the western Himalaya (Azam et al2013)

336 Mean elevation changes and mass balancecalculation

Before averaging elevation changes we exclude all interpo-lated pixels in the SRTM or SPOT5 DEMs as well as unex-pected elevation changes exceedingplusmn150 m for study sites(Pamir and Karakoram) that include surge-type glaciers orplusmn80 m for other sites Those thresholds were chosen aftercareful inspection of the elevation difference maps and his-tograms in order to identify the largest realistic glacier eleva-tion changes Glaciers that are truncated at the edges of theDEM are also excluded from mass balance calculations asthey may bias the final results if for example only their accu-mulation or ablation area is sampled Although they are trun-cated Siachen (Karakoram East) and Fedtchenko (Pamir)glaciers were retained because of their wide coverage andbecause their accumulation and ablation areas were correctlysampled in the DEMs

Elevation changes are then analyzed for 100 m altitudebins Within each altitude band we average the elevationchanges after excluding pixels where absolute elevation dif-ferences differ by more than three standard deviations fromthe mean (Berthier et al 2004 Gardner et al 2012) Thisis an efficient way to exclude outliers based on the assump-tion that elevation changes should be similar at a given alti-tude within each study site This assumption does not holdfor regions containing actively surging glaciers Thus forthe Pamir and Karakoram study sites surge-type glaciersare treated separately ie elevation changes are averaged by100 m altitude bands for each surge-type glacier individually

Where no elevation change is available for a pixel weassign to it the value of the mean elevation change of the

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1269

altitude band it belongs to in order to assess the mass bal-ance over the whole glacier area

The conversion of elevation changes to mass balance re-quires knowledge of the density of the material that has beenlost or gained and its evolution during the study period Giventhe lack of repeat measurement of density profiles over theentire snowfirnice column in the PKH we applied a con-stant density conversion factor of 850 kg mminus3 for all ourstudy sites (Sapiano et al 1998 Huss 2013)

The mass balance of each surge-type glacier is computedseparately and added (area-weighted) to the mass balance ofthe rest of the glaciers to estimate the mass balance of thewhole study site

For study sites with a high concentration of growingpro-glacial lakes such as Bhutan Everest and West Nepal(Gardelle et al 2011) our mass losses are slightly underes-timated because they do not take into account the glacier icethat has been replaced by water during the expansion of thelake Subaqueous ice losses are only estimated for LhotseSharImja Glacier (Everest area Fujita et al 2009) for thesake of comparison to previous studious

The region-wide mass balance of PKH glaciers as definedin Sect 2 is computed by extrapolating the mass balancefrom each study site to a wider sub-region (Fig 1) that is as-sumed to experience the same glacier mass changes becauseof its similar climate (Bolch et al 2012 Kaumlaumlb et al 2012)The total glacier area for each sub-region is computed fromthe RGI v20 (Arendt et al 2012)

The periods over which we calculate mass balance areslightly different from one site to another (Table 1) depend-ing on the year of acquisition of the SPOT5 DEM (two in2008 two in 2010 and five in 2011) In the following we willassume that all mass balances are representative of the period1999ndash2011 in order to estimate a region-wide mass budgetof PKH glaciers neglecting thus part of the inter-annual vari-ability

34 Accuracy assessment

The errorE1hi to be expected for a pixel elevation change

1hi is assumed to equal the standard deviationσ1h of themean elevation change1h of the altitude band it belongs toThe value ofσ1h can range fromplusmn4 m toplusmn 20 m dependingon the study site and elevation This is rather conservative asthe value ofσ1h contains also the intrinsic natural variabilityof elevation changes within the altitude band (ie it containsboth noise and real geophysical signal)

The errorE1h of the mean elevation change1h in eachaltitude band is then calculated according to standard princi-ples of error propagation

E1h =E1hiradicNeff

(1)

Neff represents the number of independent measurements inthe altitude bandNeff will be lower thanNtot the total num-

ber of elevation change measurements1hi in the altitudeband since the latter are correlated spatially (Bretherton etal 1999)

Neff =Ntot middot PS

2d (2)

where PS is the pixel size (here 40 m)d is the distance ofspatial autocorrelation determined using Moranrsquos I autocor-relation index on elevation differences off glaciers On theaverage over the 9 study sitesd = 492plusmn 72 m

The error on the penetration correction was assumed to besystematic due to the not fully understood physics involvedand equal toplusmn 15 m This error was obtained by compar-ing the SRTM C-band penetration estimated using two inde-pendent methods (i) SRTMC-bandndash SRTMX-band and (ii) dif-ference between ICESat and SRTM when ICESat elevationchange trends during 2003ndash2008 are extrapolated to Febru-ary 2000 (Kaumlaumlb et al 2012) The maximum difference of thepenetration inferred from the two methods is 10 m (Table 3)A 50 additional error margin was added to account for thefact that this comparison could only be made on 2 of ourstudy sites

Given the slender observational support for the seasonal-ity correction (Section 33) we assumed its uncertainty tobe plusmn100 on the western sites ieplusmn015 m we per win-ter month The same absolute uncertainty (plusmn015 m we perwinter month) was used for the eastern sites where no sea-sonality correction is applied but ablation andor accumula-tion may occur in winter (Wagnon et al 2013)

All errors are summed quadratically within each altitudeband and then summed for all altitude bands assuming con-servatively that they are 100 dependent The error on theSRTM penetration correction clearly dominates our errorbudget for volume change

During the conversion from volume to mass we assumea plusmn60 kg mminus3 error on the density conversion factor (Huss2013) Thisplusmn7 error is similar to the one estimated byBolch et al (2011) In a sensitivity test the volume to massconversion is also performed using two alternative densityscenarios (Kaumlaumlb et al 2012) (i) a density of 900 kg mminus3 isassumed everywhere and (ii) 600 kg mminus3 in the accumula-tion area and 900 kg mminus3 in the ablation area These two sce-narios take among other things into account that elevationchanges could be due to changes in glacier dynamics or dueto surface mass balance We then calculate the maximum ab-solute difference between the mass balances calculated usingour preferred scenario (850plusmn 60 kg mminus3 everywhere) and themass balances calculated using the two others scenarios Forthe eastern sites (West Nepal to Hengduan Shan) the max-imum absolute difference is 003 m we yrminus1 For the west-ern sites (Pamir to Spiti Lahaul) it reaches 006 m we yrminus1Those uncertainties due to the choice of a given density sce-nario remain negligible compared to other sources of errors

We also include an error due to extrapolation of our massbalance estimate from our study site (the extent of each

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1270 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

SPOT5 DEM) to the unmeasured glaciers within each sub-region To estimate this regional extrapolation error we com-pare the ICESat elevation change trends (computed as inKaumlaumlb et al 2012) within 3 times 3 cells centered at the studysites (Table 5) with the ICESat-derived trends for the entiresub-regions The absolute difference between both trends isless than 002ndash003 m yrminus1 for Bhutan West Nepal Spiti La-haul and Karakoram ie negligible For Everest the eleva-tion trend for the entire sub-region is 012 m yrminus1 less nega-tive than for the 3 times 3 cells around the SPOT5 DEM cen-ter for Hindu Kush the trend for the entire sub-region is017 m yrminus1 more negative than for the 3 times 3 cell due to thevariability of elevation changes within this sub-region (Kaumlaumlbet al 2012) The regional extrapolation error is obtainedas the area-weighted mean absolute difference between themass balances derived from ICESat for those 3 times 3 cellsand the whole sub-region and equalsplusmn004 m we yrminus1

Finally for sub-regions and total PKH estimates of masschange we include an error on the glacier area computedfrom the RGI v20 (Arendt et al 2012) The inventory er-ror is specific to each study site and evaluated based on thecomparison of our user-defined inventory on each study sitewith the RGI v20 (Table 1) We note the remarkable accu-racy of the RGI for all but one of our study sites The rel-ative errors are generally of a few percents and up to 12 for West Nepal The differences between our and existing in-ventories are probably due to the difficulty of delimitatingdebris-covered glacier parts (eg Frey et al 2012 Paul etal 2013) and accumulation areas The Hengduan Shan studysite is an exception There the RGI is based on the ChineseGlacier Inventory (Shi et al 2009 Arendt et al 2012) andoverestimates the ice-covered area by 88 compared to ourLandsat-based inventory

Unless stated otherwise all error bars from this study andpreviously published studies correspond to one standard er-ror

4 Results

41 Overview of the elevation changes

The maps of glacier elevation changes are given in Figs 2ndash10for our nine study sites with as inset the distribution ofthe elevation differences off glaciers The full maps of ele-vation differences off glaciers are shown in the Supplement(Figs S2ndashS10) Those off glacier maps indicate a local ran-dom noise of aboutplusmn5ndash10 m However at length scales of afew kilometers the spatial variations in elevation differencesare small typically less than 1 m

For the four eastern study sites from the HengduanShan to the West Nepal (Fig 1) the elevation changesaveraged over the accumulation areas are small (between003plusmn 014 m yrminus1 to 013plusmn 015 m yrminus1) whereas clear sur-face lowering is observed for all four study sites in their

ablation areas ranging fromminus050plusmn 014 m yrminus1 (Bhutan)to minus081plusmn 013 m yrminus1 (Hengduan Shan) Over the BhutanEverest and West Nepal study sites glaciers in contact witha proglacial lake often experienced larger thinning rates

Further west the Spiti Lahaul is the only site wheresurface lowering is measured both in the accumula-tion area (minus034plusmn 016 m yrminus1) and the ablation area(minus070plusmn 014 m yrminus1) Clear surface lowering is also ob-served in the ablation areas of the Hindu Kush glaciers(minus050plusmn 018 m yrminus1)

In the Karakoram (east and west Figs 8 and 9) andin the Pamir (Fig 10) the pattern of elevation changesis highly heterogeneous due to the occurrence of numer-ous glacier surges or similar flow instabilities The eleva-tion changes in the accumulation areas of those three studysites are slightly positive (between+022plusmn 014 m yrminus1 and+030plusmn 018 m yrminus1) In the ablation areas small surfacelowering is observed for the two Karakoram study sites (av-erage of the two sitesminus033plusmn 016 m yrminus1) whereas no el-evation change is found in Pamir (minus002plusmn 013 m yrminus1)

Note that those mean rates of elevation changes are sensi-tive to the choice of the ELA assumed to equal the snowlinealtitude in Landsat images from around year 2000

42 Influence of a debris cover

To evaluate thinning rates over debris-covered and debris-free ice we perform a histogram adjustment in order tocompare data sets with similar altitude distributions This isneeded because debris-covered parts tend to be concentratedat lower elevations The adjustment consists in randomly ex-cluding pixels over debris-free ice from each elevation bandto match a scaled version of the debris-covered pixel distribu-tion (Kaumlaumlb et al 2012) Elevation changes with altitude canthen be compared for clean and debris-covered ice (Fig 11)

The thinning is higher for debris-covered ice in the Ever-est area and on the lowermost parts of glaciers in the PamirBut on average in the Pamir western Karakoram and SpitiLahaul the lowering of debris-free and debris-covered ice issimilar The thinning is stronger over clean ice in the Heng-duan Shan Bhutan West Nepal Hindu Kush and althoughto a lesser extent in the eastern Karakoram

43 Mass balance over the nine study sites

Glacier mass changes have been heterogeneous overthe PKH since 1999 Slight mass gains or balancedmass budgets are found in the north-west in thePamir (+014plusmn 013 m we yrminus1) and for the two studysites in the central Karakoram (+009plusmn 018 m we yrminus1

and +011plusmn 014 m we yrminus1) Mass loss is moderatein the adjacent Hindu Kush with a mass balance ofminus012plusmn 016 m we yrminus1 Glaciers in all our eastern studysites (two in Nepal Bhutan and Hengduan Shan) experi-enced homogeneous mass losses with mass balances ranging

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1271

Fig 2 Glacier elevation changes over the Hengduan Shan study site Inset distribution of the elevation differences off glacier The medianand the standard deviation of the elevation difference and the number of pixels off glaciers are given (by construction the mean elevationdifference is null)

Fig 3Same as Fig 2 but for the Bhutan study site

from minus033plusmn 014 m we yrminus1 to minus022plusmn 012 m we yrminus1The most negative mass balance is measured in Spiti Lahaul(western Himalaya) atminus045plusmn 013 m we yrminus1 (Table 4)

However within a study site mass balances can be highlyvariable from one glacier to another In the Everest area twolarge glaciers (71 km2 each) experienced distinctly differentmass balance withminus016plusmn 016 m we yrminus1 for the south-flowing Ngozumpa Glacier andminus059plusmn 016 m we yrminus1

for the north-flowing Rongbuk Glacier (Fig 4) In Bhutan(Fig 3) mass loss is higher south (minus025plusmn 013 m we yrminus1)

than north (minus014plusmn 012 m we yrminus1) of the main Hi-malayan ridge though not at a statistically significant level

The region-wide mass budget of PKH glaciers isnegative minus101plusmn 55 Gt yrminus1 (minus014plusmn 008 m we yrminus1)which corresponds to a sea level rise contribution of0028plusmn 0015 mm yrminus1 over 1999ndash2011

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1272 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 4Same as Fig 2 but for the Everest study site

Fig 5Same as Fig 2 but for the West Nepal study site

44 Glacier surges

In the Pamir and the Karakoram the spatial pattern of ele-vation changes of many glaciers is typical of surge events orflow instabilities They can be identified because of their veryhigh thinning and thickening rates that can both reach up to16 m yrminus1 (Figs 8ndash10)

Most of them have already been reported to be surge-type glaciers based on their velocity (Kotlyakov et al 2008Quincey et al 2011 Copland et al 2009 Heid and Kaumlaumlb2012) morphology (Copland et al 2011 Barrand and Mur-ray 2006 Hewitt 2007) or elevation changes (Gardelle etal 2012a) Here we only identify surge-type glaciers for thesake of mass balance calculation and we do not attempt toprovide an exhaustive up-to-date surge inventory This is why

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1273

Fig 6 Same as Fig 2 but for the Spiti Lahaul study site The map of elevation changes contains many voids due to the presence ofthin clouds during the acquisition of the SPOT5 DEM The mass balance of the Chhota Shigri and Bara Shigri glaciers are respectivelyminus039plusmn 018 m we yrminus1 andminus048plusmn 018 m we yrminus1

the published surge inventories in the Pamir and the Karako-ram do not necessarily match the surge-type glaciers labeledon the elevation change maps as these refer only to our ob-servation period (1999ndash2011)

The mass balance of surge-type glaciers (respectivelynon-surge-type glaciers) is +019plusmn 022 m we yrminus1

(+012plusmn 011 m we yrminus1) in the Pamir+009plusmn 030 m we yrminus1 (+009plusmn 015 m we yrminus1) inthe western Karakoram and+010plusmn 025 m we yrminus1

(+012plusmn 012 m we yrminus1) in the eastern Karakoram Theoverall similarity between the mass budgets for surge-typeand non-surge-type glaciers shows that those flow instabili-ties seem not to affect the glacier-wide mass balance at leastover short time periods here 10 yr

5 Discussion

51 SRTM penetration

Alternative estimates of SRTMC-bandpenetration for the PKHregion were obtained by Kaumlaumlb et al (2012) by comparing el-evations of the ICESat altimeter (Zwally et al 2002) with theSRTMC-bandDEM in PKH and extrapolating the 2003ndash2008trend of elevation changes back to the acquisition date ofSRTM The difference between east and west is more dis-tinct in our case with a mean penetration in the Karakoram of34 m (24plusmn 03 m in Kaumlaumlb et al 2012) 24 m in Bhutan and14 m around the Everest area (25plusmn 05 m for a wider areaincluding east Nepal and Bhutan in Kaumlaumlb et al 2012) Thiseastwest gradient is expected given that in February 2000when the SRTM mission was flown the snowfirn thicknesswas probably larger in the western PKH where glaciers areof winter-accumulation type than in the eastern PKH where

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1274 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 7Same as Fig 2 but for the Hindu Kush study site

glaciers are of summer-accumulation type Further studieswould be necessary to investigate the possible reasons be-hind the relatively low SRTMC-bandpenetration for the Pamir(18 m) that do not follow the eastwest gradient mentionedabove

Both our and Kaumlaumlb et alrsquos (2012) estimates agree withintheir error bars and seem thus to provide robust first-orderestimates for the SRTMC-bandradar penetration in the regionHowever it remains to be verified if a penetration depth com-puted at the regional scale is appropriate for a single smallglacier whose altitude distribution is different from the oneof whole study site (eg Abramov Glacier in the far north ofour Pamir study site)

52 Thinning over debris-covered ice

PKH glaciers are heavily debris-covered in their ablation ar-eas because of steep rock walls that surround them and in-tense avalanche activity Twelve percent (12 ) of the glacierarea in our study sites are covered with debris (up to 22 inthe Everest area Table 1) which is in agreement with pre-vious estimates for the Himalaya and the Karakoram 13 in Kaumlaumlb et al (2012) and 10 in Bolch et al (2012) Thedebris layer is expected to modify the ablation of the under-

lying ice It will increase ablation if its thickness is just a fewcentimeters (dominance of the albedo decrease) or decreaseablation it if the layer is thick enough to protect the ice fromsolar radiation (Mattson et al 1993 Mihalcea et al 2006Benn et al 2012 Lejeune et al 2013)

However recent studies have reported similar overall thin-ning rates over debris-covered and debris-free ice in thePKH (Kaumlaumlb et al 2012 Gardelle et al 2012a) or evenhigher lowering rates over debris-covered parts in the Ever-est area (Nuimura et al 2012) Our findings are consistentwith Nuimura et al (2012) for the Everest study site with astronger thinning (rate of elevation change ofminus097 m yrminus1)

in debris-covered areas than over clean ice (rate of elevationchange ofminus057 m yrminus1) In the Pamir Karakoram and SpitiLahaul study sites these rates are similar while in the HinduKush West Nepal Bhutan and Hengduan Shan study sitesthinning is greater over clean ice in agreement with the com-monly assumed protective effect of debris (Fig 4) Thosedifferences between our 9 study sites suggest a complex re-lationship between debris cover and glacier wastage

As local glacier thickness changes are a result of bothmass balance processes and a dynamic component theseunexpected high thinning rates over debris-covered parts

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1275

Fig 8Same as Fig 2 but for the Karakoram east study site Surge-type glaciers in an active surge phase (resp quiescent phase) are identifiedwith a solid triangle (resp solid circle)

are potentially the result of glacier-wide processes Over aglacier tongue the ablation can be enhanced by the pres-ence of steep exposed ice cliffs or supraglacial lakes andponds (Sakai et al 2000 2002) In addition it is also pos-sible that the glacier tongues are mostly covered by thindebris layers such that the albedo effect dominates the in-sulating effect resulting in an increased tongue-wide abla-tion The prevalence of thin debris was recently suggestedfor debris-covered glaciers around the Mount Gongga in thesouth-eastern Tibetan Plateau (Zhang et al 2013b) Finallyice-flow rates could be very low at debris-covered parts asshown by Quincey et al (2007) in the Everest area Kaumlaumlb(2005) for Bhutan and Scherler et al (2011b) in the HinduKush-Karakoram-Himalaya Thus all three factors are likelyto increase the overall thinning under debris despite the un-doubted protective effect of a continuous thick debris coverat a local scale Future investigations combining surface ve-locity fields and maps of elevation changes could allow eval-uating more closely the relative role of ablation and ice fluxesin thinning rates over PKH glacier tongues (eg Berthierand Vincent 2012 Nuth et al 2012) Measurements of the

spatial distribution of debris thickness over the PKH glaciertongues are also needed

Note that in the case of debris-covered tongues the ele-vation change measurement represents the glacier thicknesschange but also the possible debris thickness evolution Thelatter remains poorly known in the PKH or in any othermountain range But based on the debris discharges givenby Bishop et al (1995) for Batura Glacier (Pakistan) 48to 90times 103 m3 yrminus1 the thickening of the debris can be es-timated between 27 and 55 mm yrminus1 which is negligiblegiven the magnitude of the glacier elevation changes

53 Comparison to previous mass balance estimates

Throughout the PKH there is only a single peer-reviewedglaciological record (on Chhota Shigri Glacier) coveringmost of our study period (Wagnon et al 2007 Vincent etal 2013) Our geodetic mass balance estimate for the SpitiLahaul study site has been recently compared to those glacio-logical measurements on Chhota Shigri Glacier (Vincent etal 2013) Mass balance records reported in Yao et al (2012)for the Tibetan Plateau and the surroundings mountain ranges

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1276 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 9Same as Fig 8 but for the Karakoram west study site

only start during glaciological year 20052006 This is whyhere we do not compare our mass balance estimates toglaciological records and limit our comparison to others re-gional estimates obtained using the geodetic or gravimetricmethods

We stress that the comparison of our mass balance mea-surements to published estimates is complicated by thefact that generally they do not cover the same time pe-riod Little is known about the inter-annual variability inthe PKH during the 21st century due to the limited num-ber of long glaciological time series The 9 yr mass balance

record on Chhota Shigri Glacier reveals a relatively highinter-annual variability of the annual mass balance (stan-dard deviationplusmn 074 m we yrminus1 during 2003ndash2011 Vin-cent et al 2013) which is similar to the one obtained be-tween 1968 and 1998 on Abramov Glacier (standard devi-ation 069 m we yrminus1 WGMS 2012) Thus inter-annualvariability can explain part of the difference between two cu-mulative mass balance estimates over 5ndash10 yr even if they areoffset by only one or two years

Based on ICESat repeat track measurements and theSRTM DEM Kaumlaumlb et al (2012) measured a region-wide

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1277

Fig 10Same as Fig 8 but for the Pamir study site

mass balance ofminus021plusmn 005 m we yrminus1 between 2003and 2008 for Hindu Kush-Karakoram-Himalaya glaciersFor that same area (ie excluding the Pamir and theHengduan Shan study sites) we found a mass balance ofminus015plusmn 007 m we yrminus1 Thus our result agrees with Kaumlaumlbet al (2012) within the error bars although our study periodis longer (1999ndash2011) and our measurement principle andextrapolation approach are different In addition we havecalculated updated mass balances from Kaumlaumlb et al (2012)ie averages over 3times 3 degree cells centered over our studysites and find that they are consistent within error barswith our mass balance estimates (Table 5) Similar resultsare found when our mass balances are compared to a spa-tially extended ICESat analysis for 2003ndash2009 (Gardner etal 2013) Mass losses measured with ICESat (Kaumlaumlb et al2012 Gardner et al 2013) tend to be slightly larger thanours This could be due to the difference in time periodsAlso our smaller mass losses could be partly explained if wesystematically underestimated the poorly constrained pene-

tration of the C-band andor X-band radar signal into ice andsnow

At a more local scale Bolch et al (2011) reported a massbalance ofminus079plusmn 052 m yrminus1 between 2002 and 2007 byDEM differencing over a glacier area of 62 km2 includingten glaciers south and west of Mt Everest For these sameglaciers Nuimura et al (2012) found a mean mass balanceof minus045plusmn 060 m yrminus1 during 2000ndash2008 while they com-puted a mass balance ofminus040plusmn 025 m yrminus1 between 1992and 2008 over a glacier area of 183 km2 around Mt EverestIn the present study we found an average mass balance ofminus041plusmn 021 m yrminus1 over the ten glaciers mentioned above(Table 5 Fig S1) and a value ofminus026plusmn 013 m yrminus1 overthe whole Everest study site between 2000 and 2011 Ourvalues are comparable to those of Nuimura et al (2012)Our mass balance is less negative than the 2002ndash2007 massbalance of Bolch et al (2011) but not statistically differentwhen error bars are considered (Fig 12) Indeed Bolch etal (2011) acknowledged some high uncertainties for their

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1278 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Table 4Glacier mass budget of the nine sub-regions to which the results of the study sites (mass balances) have been extrapolated For eachof them the total glacierized area as well as the percentage of the glacier area over which the mass balance was computed (ie study sites)are given The total glacier area is derived from the RGI v20 (Arendt et al 2012) The error for the mass balance in each sub-region is theroot of sum of squares (RSS) of the mass balance error of each study site and the regional extrapolation error The error for the mass budgetin each sub-region includes additionally the error from the total glacier area in the RGI v20

Sub-region Total glacier Measured glacier Mass balance Mass budgetarea (km2) area () (m we yrminus1) (Gt yrminus1)

Hengduan Shan 11584 12 minus033plusmn 014 minus38plusmn 38Bhutan 4021 34 minus022plusmn 013 minus09plusmn 05Everest 6226 23 minus026plusmn 014 minus16plusmn 08West Nepal 6849 13 minus032plusmn 014 minus22plusmn 10Spiti Lahaul 9043 23 minus045plusmn 014 minus41plusmn 13Hindu Kush 6135 13 minus012plusmn 016 minus07plusmn 10Karakoram East

1902428 +011 plusmn 014 +19 plusmn 31

Karakoram West 30 +009 plusmn 018Pamir 9369 34 +014 plusmn 014 +13 plusmn 13

TotalArea- 72251 30 minus014plusmn 008 minus101plusmn 55weighted average

Table 5Comparison of geodetic mass balances between this study and previous published results with overlapping study periods and similargeographic locations Updated figures from Kaumlaumlb et al (2012) are averaged over 3 times 3 cells centered over the study sites of the presentstudy

GlacierSite name This study Bolch et al (2011) Nuimura et al (2012) Kaumlaumlb et al (2012 updated)2002ndash2007 2000ndash2008 2003ndash2008

Study Mass balance Area (km2)a Mass balance Area Mass balance Mass balanceperiod (m we yrminus1) (m we yrminus1) (km2) (m we yrminus1) (m we yrminus1)

Hengduan San 1999ndash2010 minus022 plusmn 014 1303 (55 )Bhutan 1999ndash2010 minus022 plusmn 012 1384 (64 ) minus052 plusmn 016b

Everest

1999ndash2011

minus026 plusmn 013 1461 (58 ) minus039 plusmn 011AX010 minus090plusmn 034 04Changri SharNup minus042plusmn 017 161 (79 ) minus029plusmn 052 130 minus055plusmn 038Khumbu minus051plusmn 019 203 (47 ) minus045plusmn 052 170 minus076plusmn 052Nuptse minus037plusmn 020 50 (74 ) minus040plusmn 053 40 minus034plusmn 027Lhotse Nup minus021plusmn 027 24 (70 ) minus103plusmn 051 19 minus022plusmn 047Lhotse minus043plusmn 018 85 (83 ) minus110plusmn 052 65 minus067plusmn 051Lhotse SharImja minus070plusmn 052 98 (44 ) minus145plusmn 052 107 minus093plusmn 060Amphu Laptse minus046plusmn 034 25 (38 ) minus077plusmn 052 15 minus018plusmn 094Chukhung +044 plusmn 024 42 (47 ) +004 plusmn 054 38 +043 plusmn 081Ama Dablam minus049plusmn 017 36 (54 ) minus056plusmn 052 22 minus056plusmn 073Duwo minus016plusmn 026 19 (18 ) minus196plusmn 053 10 minus068plusmn 074Total Khumbu minus041plusmn 021 744 minus079plusmn 052 617 minus045plusmn 060(10 Glaciers above)

West Nepal 1999ndash2011 minus032 plusmn 013 908 (40 ) minus032 plusmn 012Spiti Lahaul 1999ndash2011 minus045 plusmn 013 2110 (46 ) minus038 plusmn 006Karakoram East 1999ndash2010 +011 plusmn 014 5328 (42 ) minus004 plusmn 004Karakoram West 1999ndash2008 +009 plusmn 018 5434 (45 )Hindu Kush 1999ndash2008 minus012 plusmn 016 793 (80 ) minus020 plusmn 006Pamir 1999ndash2011 +014 plusmn 013 3178 (50 )

a The total glacier area is given in km2 and in parenthesis the of the glacier area actually covered with measurementsb For this cell the ICESat coverage is insufficient and does not sample all glacier elevations

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1279

Fig 11 Elevation changes (SPOT5-SRTM) in the ablation area for clean ice (blue circles) and debris-covered ice (black circles) for eachstudy site For the Karakoram (east and west) and the Pamir study sites only non-surge-type glaciers are considered Standard error of theelevation differences are typically less thanplusmn2 m with each 100 m elevation band (not shown for the sake of clarity) Note elevation changesover separate sections of a glacier cannot be treated as mass changes due to the disregard of glacier dynamics

estimate over this short time period The differences in sur-vey periods and in the glacier outlines may also explainpart of these discrepancies In particular the delimitation ofthe accumulation area is notoriously difficult and Bolch etal (2011) did not include some of the steep high altitudeslopes which we included due to their potential avalanchecontribution For the 10 glaciers mentioned above our massbalances would become 012 m we yrminus1 more negative ifwe used the exact same outlines as Bolch et al (2011)Our positive mass balance for the small Chukhung Glacier(+044plusmn 024 m we yrminus1) is in agreement with Nuimura etal (2012) but enigmatic in a context of regional glacier lossand deserves further investigation If the 2002ndash2007 massbalance estimate by Bolch et al (2011) is excluded all otherrecent mass balance estimates suggest no acceleration of themass loss in the Everest area when compared to the long-term(1970ndash2007) mass balance measured by Bolch et al (2011)in this regionminus032plusmn 008 m weyrminus1

In Bhutan our results indicate a stronger mass loss forsouthern glaciers than for northern ones Interestingly in thisarea Kaumlaumlb (2005) measured high speeds (up to 200 m yrminus1)

for glaciers north of the Himalayan main ridge and ratherlow velocities over southern glacier tongues by using repeat

ASTER data This is consistent with the more negative massbalance that we report for the southern glaciers in Bhutanie for the slow flowing presumably downwasting glaciers

Berthier et al (2007) reported a mass balance betweenminus069 andminus085 m we yrminus1 for an ice-covered area of915 km2 around Chhota Shigri Glacier between 1999 (SRTMDEM) and 2004 (SPOT5-HRG DEM) For the same areawe found a mass balance ofminus045plusmn 013 m we yrminus1 over1999ndash2011 The difference in the study period may explainpart of the differences Also the methodology by Berthieret al (2007) did not explicitly take into account the SRTMradar penetration over glaciers and also corrected an appar-ent elevation-dependent bias according to glacier elevationswhich is now known as a resolution issue and is best cor-rected according to the terrain maximum curvature (Gardelleet al 2012b) More details and discussions about these dif-ferences can be found in Vincent et al (2013)

Importantly the results of the eastern Karakoram studysite confirm the balanced or slightly positive mass budgetreported by Gardelle et al (2012a) in the western Karako-ram over the past decade The two Karakoram study sitesare adjacent with a small overlapping area (sim 1100 km2 ofglaciers) where elevation differences agree within 1 m Both

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1280 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 12 Comparison of geodetic mass balances for 10 glaciers inthe Mount Everest area 1999ndash2010 mass balances measured inthe present study are compared with published estimates during2002ndash2007 (Bolch et al 2011) and during 2000ndash2008 (Nuimuraet al 2012) The 1-to-1 line is drawn The mean difference (plusmn1standard deviation) between (i) our values and Bolch et al (2011)are (047plusmn 058 m we yrminus1) and (ii) our values and Nuimura etal (2012) are (012plusmn 021 m we yrminus1) Part of the differences be-tween estimates may be due to the different periods surveyed andthe different glacier outlines used

sites show similar mass balances over slightly different pe-riods +011plusmn 014 m yrminus1 over 1999ndash2010 in the east and+009plusmn 018 m yrminus1 over 1999ndash2008 in the west This con-firms a ldquoKarakoram anomalyrdquo that was first suggested basedon field observations (Hewitt 2005) We report a mass bal-ance of+010plusmn 013 m yrminus1 for Baltoro Glacier (see loca-tion in Fig 8 size= 304 km2) between 1999 and 2010which is consistent with its gradual speed-up reported since2003 (Quincey et al 2009) However given the completelydifferent methods used in our and the latter study we can-not conclude that this demonstrates a real shift from nega-tive to positive mass balance The mass budget of SiachenGlacier (1145 km2 sometimes referred to as the highestbattlefield in the world) is also balanced or slightly pos-itive (+014plusmn 014 m we yrminus1) Thus the alarming state-ment by Rasul (2012) that this glacier retreated by 59 kmand lost 17 of its mass during 1989ndash2009 is very likelyerroneous Again within error bars our regional mass bal-ance in the Karakoram agrees with the mass balance ofminus003plusmn 004 m yrminus1 found by Kaumlaumlb et al (2012) over 2003ndash2008 and with the mass balance ofminus010plusmn 018 m we yrminus1

(2-sigma error level) obtained by Gardner et al (2013) for aregion including both the Karakoram and the Hindu Kush

In the northernmost part of the Pamir in the Alay moun-tain range the mass balance of Abramov Glacier has beenmeasured in the field for 30 yr between 1968 and 1998(WGMS 2012) with a mean value ofminus046 m yrminus1 Herewe report a mass balance ofminus003plusmn 014 m yrminus1 for this

glacier over 1999ndash2011 Our near zero mass budget estimatefor this glacier disagrees with negative mass balances recon-structed for the same period with a model run using biascorrected NCEPNCAR re-analysis data and calibrated withfield data (Barundun et al 2013) An underestimate of ourSRTM C-Band penetration in the Pamir as a whole and inparticular for the small Abramov Glacier may contribute tothese differences

For the whole Pamir study site our mass budget is bal-anced or slightly positive (+014plusmn 013 m we yrminus1) In theeastern Pamir the mass balance of Muztag Ata Glacier was+025 m we yrminus1 between 2005 and 2010 (Yao et al 2012)Although Muztag Ata Glacier is located outside of our Pamirstudy site and his glaciological records only cover the sec-ond half of our study period this measurement is consistentwith our remotely sensed mass balance Thus the ldquoKarako-ram anomalyrdquo should probably be renamed the ldquoPamir-Karakoram anomalyrdquo

This slightly positive or balanced mass budget measuredin the Pamir seems to contradict previous findings by Heidand Kaumlaumlb (2012) They reported rapidly decreasing glacierspeeds in this region between two snapshots of annual ve-locities in 20002001 and 20092010 an observation thatwould fit with negative mass balances (Span and Kuhn 2003Azam et al 2012) We note though that the relationship be-tween glacier mass balance and ice fluxes is not straightfor-ward and might in particular involve a time delay betweena change in mass balance and the related dynamic response(Heid and Kaumlaumlb 2012) Thus the decrease in speed couldwell be due to negative mass balances before or partiallybefore 1999ndash2011 We acknowledge that this discrepancyneeds to be investigated further in particular by examiningcontinuous changes in annual glacier speed through the firstdecade of the 21st Century Indeed Heid and Kaumlaumlb (2012)measured differences between two annual periods which maynot represent mean decadal velocity changes

Jacob et al (2012) using satellite gravimetry (GRACE)measured a mass budget ofminus5plusmn 6 Gt yrminus1 over 2003ndash2010for the ldquoKarakoram-Himalayardquo their region 8c which cor-responds to our study area excluding the Pamir Karakoramand Hindu Kush sub-regions but including some glaciersnorth of the Himalayan ridge already on the Tibetan PlateauFor this subset of our PKH study area we measured amass budget ofminus126plusmn 42 Gt yrminus1 The two estimates over-lap within their error bars especially if our error is dou-bled to match the two standard error level used by Jacob etal (2012) The estimate of Jacob et al (2012) over our com-plete study region (PKH) ieminus6plusmn 8 Gt yrminus1 (sum of theirregions 8b and 8c) also includes mass changes for the KunlunShan sub-region for which we did not measure glacier massbalances Therefore the comparison with our PKH value(minus101plusmn 55 Gt yrminus1) is not straightforward but suggests areasonable agreement especially if we consider the slightmass gain (15plusmn 17 Gt yrminus1) measured with ICESat in theWest Kunlun (Gardner et al 2013)

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1281

Table 6 Annual average of the glacier seasonal and decadal mass loss contributions to Indus Ganges and Brahmaputra discharges Basinarea and seasonal contributions are given by Kaser et al (2010) glacier area in each basin is derived from the RGI v20 (Arendt et al 2012)Numbers in parentheses indicate the percentage of glacierized area within the river basin Seasonal contributions were modeled by Kaser etal (2010) based on the CRU 20 dataset over 1961ndash1990 assuming a balanced glacier mass budget Glacier decadal mass loss contributionsare given as annual average over 1999ndash2011

River Basin area Glacier area Glacier contribution Mean discharge(km2) (km2) (m3 sminus1) (m3 sminus1)

Seasonally delayed Decadal mass lossKaser et al (2010) (this study)

Indus 1 139 814 25 598 (2 ) 105 103plusmn 72 2146 (Chevallier personalcommunication 2012)

Ganges 1 023 609 11 168 (1 ) 47 103plusmn 49 11739 Papa et al (2012Brahmaputra 527 666 15 296 (3 ) 33 147plusmn 64 19710 updated)

54 Reasons for the heterogeneous pattern of massbalance

The PKH glacier mass balance is slightly negative(minus014plusmn 008 m we yrminus1) but changes are not homoge-neous from one sub-region to another and seem to coincidewith the climatic settings of the mountain range For the east-ern sites (from the Hengduan Shan to West Nepal) which areunder the influence of the Indian and south-east Asian sum-mer monsoons the mass balances are moderately negative(sim minus026 m we yrminus1) In the northwest of the PKH at thePamir and Karakoram sites where glacier climate is char-acterized by the westerlies in winter mass balances are bal-anced or slightly positive (sim +010 m we yrminus1) In betweenthese two influences the glaciers of the Spiti Lahaul sitein a monsoon-arid transition zone experienced the strongestmass loss (minus045plusmn 013 m we yrminus1)

This heterogeneous pattern of mass balances seems to berelated to trends in precipitation throughout the PKH Archerand Fowler (2004) reported an increase in winter precipi-tation over the Upper Indus Basin since 1961 a trend con-firmed by Yao et al (2012) over the Pamir and the Karako-ram based on the analysis of the Global Precipitation Cli-matology Project data set (GPCP Adler et al 2003) Thisprecipitation increase is a source of greater accumulation forglaciers and thus a possible explanation for their slightly pos-itive mass budgets In addition in the Upper Indus Basin themean summer temperature has been decreasing since 1961(Fowler and Archer 2006) which is consistent with the find-ings of Shekhar et al (2010) in the Karakoram who reporteda decrease in both maximum and minimum temperature since1988 These increases in winter precipitation and summertemperature likely translate in greater accumulation and re-duced ablation changes which are consistent with the bal-anced or slightly positive glacier mass budget

On the other hand in the east of the PKH the Indiansummer monsoon has been weakening since the 1950s (Bol-lasina et al 2011) which could have reduced the amount

of snowfall over the eastern study sites and thus resultedin negative mass balances In addition maximum and min-imum temperatures increased since 1988 in the western Hi-malaya (Shekhar et al 2010) and maximum temperaturesincreased in the central Himalaya (Nepal) since the mid-1970s (Shrestha et al 1999) Thus both precipitation andtemperature trends in the Himalaya likely contributed to thenegative mass balances measured over the correspondingstudy sites (from Spiti Lahaul to Hengduan Shan Fig 1)

However gridded data sets such as GPCP are not neces-sarily representative of the climate at high altitude IndeedHewitt (2005) reported a 5-to-10-fold increase in precipi-tation between the glacier front and the accumulation areain the Karakoram that is not captured by reanalysis dataas recently confirmed by Immerzeel et al (2012) Thus di-rect measurements of accumulation on High Mountain Asiaglaciers (eg Azam et al 2013) and high-altitude weatherstations are eagerly needed to validate the large scale grid-ded data set (eg reanalysis) test their ability to describe thespecific climate near to PKH glaciers and assess the relativerole played by temperature and precipitation changes

55 Contribution to water resources

Our glacier mass balance can be averaged over large riverbasins to estimate the mean contribution to river runoffdue to a decade of glacier mass loss We assume therebyonly direct river runoff without loss terms For the threerivers for which we cover the entire glacierized area of theircatchments within our sub-regions (Fig 1) these contribu-tions to the river discharge are equal to 103 m3 sminus1 (Indus)103 m3 sminus1 (Ganges) and 147 m3 sminus1 (Brahmaputra) for theperiod 1999ndash2011 (Table 6)

Rivers of PKH being key water resources measures oftheir discharges are highly sensitive and difficult to access forpolitical reasons Papa et al (2012) built recent time series ofdischarges for Ganges and Brahmaputra by using altimetrymeasurements The locations of the corresponding dischargeestimates in Bangladesh are given in Fig 1 We can therefore

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1282 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

evaluate the relative contribution of glacier decadal mass loss(Table 6) to annual river run-off at 09 for the Gangesat Hardinge (basin area ofsim 1 020 000 km2) and 07 forBrahmaputra at Bahadurabad (basin area ofsim 530 000 km2)between 1999 and 2011 Given the discharge measured at theGuddu station by the Surface Water Hydrology Project of theWater and Power Development Authority (Pakistan see lo-cation on Fig 1) between 1999 and 2003 we can estimatethe contribution of glacier mass loss at 48 for the IndusRiver at this station

The seasonal contributions of glaciers to river run-off (iethe part of the annual precipitation that experiences season-ally delayed release from the glaciers) directly results fromseasonal variations of meteorological quantities in contrastto the glacier mass loss contribution which results from thelong-term climatic change Both runoff components directlyaffect the amount of water available in a river at a given lo-cation and at a given time so that their comparison could beof interest Even if the seasonal timing of the glacier massloss contribution is unclear it could in general peak at asimilar time of the year than the seasonal glacier contribu-tion so that both components rather enlarge each other thancancel Seasonal glacier contributions have been modeled byKaser et al (2010) The latter study assumed glaciers to bein equilibrium with the present climate and thus prescribedglacier mass balances equal to 0 Thus their seasonally de-layed glacier contributions do not include the part due todecadal glacier mass loss and is complementary to our es-timates (Table 6) For the eastern basins under the influenceof the Indian summer monsoon (Ganges and Brahmaputra)the contribution of glacier mass loss is higher than the sea-sonal contribution In other words for the first decade of the21st century the mass loss of glaciers is on average moreimportant for water availability in those two rivers than theirseasonal water storage effect The situation is different for theIndus Basin where the glacier melt season is also the dry sea-son which results in a relatively higher seasonally delayedglacier contribution to the river discharge There the season-ally delayed water release from the glaciers and the decadalglacier mass loss contribute on average equally to the riverdischarge Both contributions are strongly (and equally) de-pendent on the location of the discharge measurement andwill be significantly larger (in relative term) in more heav-ily glacierized and smaller mountain catchments located inthe upper part of these major river basins (Bookhagen andBurbank 2010)

6 Conclusion

In this study we assessed the spatial pattern of glaciermass balance over the PKH by comparing the SRTM andSPOT5 DEMs over 9 well-distributed study sites between1999 and 2011 We found slightly positive or balanced massbudgets in the western part (Pamir Karakoram) moder-ate mass loss in the east (Nepal Bhutan Hengduan Shan)

and the most negative mass balance in the monsoon-aridtransition zone (Spiti Lahaul in the western Himalaya)By extrapolating these values to climatically homogeneoussub-regions we estimate the PKH overall mass balanceto have beenminus014plusmn 008 m we yrminus1 between 1999 and2011 which corresponds to a contribution to sea level riseof 0028plusmn 0015 mm yrminus1 This is about 4 of the globalglacier mass loss estimated by Gardner et al (2013) thoughover a slightly shorter time period (2003ndash2009) The largestsource of uncertainties in those geodetic mass balance esti-mates is the poorly constrained penetration of the C-BandSRTM radar signal into snow and ice

Our technique of DEM differencing allows mapping in de-tail the spatial pattern of glacier elevation changes Thosemaps reveal many surge-type behaviors both in the Pamirand the Karakoram It also appears that the glacier-wide massbalance does not seem to be considerably affected by themass transfer from surges over thesim 10 yr time period stud-ied Similar lowering rates over debris-covered and clean iceare observed for four study sites despite the commonly as-sumed insulating effect of debris covers This suggests forthe scale of entire glacier tongues a spatial mixture of re-duced ablation under intact thick debris covers and enhancedablation by supraglacial lakes or ice cliffs or due to thin de-bris layer and very low glacier velocities in ablation areasthat reduce the ice advection downstream

The regionally heterogeneous pattern of ice wastage is inbroad agreement with contrasted trends in large-scale pre-cipitation and temperature However glaciological (eg ac-cumulation) and meteorological records are needed at highaltitude in the vicinity of glaciers to evaluate more closelythe local influence of atmospheric variables on glacier massbalance and fully understand the reasons behind those con-trasted mass budgets in the PKH

Supplementary material related to this article isavailable online at httpwwwthe-cryospherenet712632013tc-7-1263-2013-supplementpdf

Acknowledgements We thank G Cogley A Gardner T NuimuraT Bolch P Wagnon M Barandun M Hoelzle M Huss and ananonymous referee for their constructive comments that led to aconsiderably improved manuscript We thank the Water and PowerDevelopment Authority (WAPDA) for their hydrological data(processed by P Chevallier) and F Papa for sharing his updatedrun-off data ahead of publication We thank T Bolch for sharinghis glacier outlines from the Everest area We thank the USGSfor allowing free access to their Landsat archive CGIAR-SCI forSRTM C-band data DLR for SRTM X-band data and GLIMS forglacier inventories J Gardelle acknowledges a PhD fellowshipfrom the French Space Agency (CNES) and the French NationalResearch Center (CNRS) E Berthier acknowledges support fromCNES through the TOSCA and ISIS proposal no 397 and fromthe Programme National de Teacuteleacutedeacutetection Spatiale E Berthieracknowledges M Masiokas and R Villalba for welcoming him

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1283

at the IANIGLA (Mendoza Argentina) Y Arnaud acknowledgessupport from French National Research Agency through ANR-09-CEP-005-01PAPRIKA A Kaumlaumlb acknowledges support fromESA through Glaciers_cci (400010177810I-AM) and from theEuropean Research Council under the European Unionrsquos SeventhFramework Programme (FP2007-2013)ERC Grant Agreementn 320816

Edited by I M Howat

The publication of this articleis financed by CNRS-INSU

References

Adler R Huffman G Chang A Ferraro R Xie P JanowiakJ Rudolf B Schneider U Curtis S Bolvin D Gruber ASusskind J Arkin P and Nelkin E The Version-2 GlobalPrecipitation Climatology Project (GPCP) Monthly Precipita-tion Analysis (1979ndashPresent) J Hydrometeorol 4 1147ndash11672003

Archer D R and Fowler H J Spatial and temporal variationsin precipitation in the Upper Indus Basin global teleconnectionsand hydrological implications Hydrol Earth Syst Sci 8 47ndash61doi105194hess-8-47-2004 2004

Arendt A Bolch T Cogley J G Gardner A Hagen J-OHock R Kaser G Pfeffer W T Moholdt G Paul F RadicV Andreassen L Bajracharya S Beedle M Berthier EBhambri R Bliss A Brown I Burgess E Burgess DCawkwell F Chinn T Copland L Davies B de AngelisH Dolgova E Filbert K Forester R Fountain A Frey HGiffen B Glasser N Gurney S Hagg W Hall D Hari-tashya U K Hartmann G Helm C Herreid S Howat IKapustin G Khromova T Kienholz C Koenig M KohlerJ Kriegel D Kutuzov S Lavrentiev I LeBris R Lund JManley W Mayer C Miles E Li X Menounos B Mer-cer A Moelg N Mool P Nosenko G Negrete A NuthC Pettersson R Racoviteanu A Ranzi R Rastner P RauF Rich J Rott H Schneider C Seliverstov Y Sharp MSigurethsson O Stokes C Wheate R Winsvold S WolkenG Wyatt F and Zheltyhina N Randolph Glacier Inventory[v20] A Dataset of Global Glacier Outlines Global Land IceMeasurements from Space Boulder Colorado USA Digital Me-dia httpwwwglimsorgRGIrandolphhtml 2012

Azam M Wagnon P Ramanathan A Vincent C Sharma PArnaud Y Linda A Pottakkal J Chevallier P Singh V andBerthier E From balance to imbalance a shift in the dynamicbehaviour of Chhota Shigri glacier western Himalaya India JGlaciol 58 315ndash324 doi1031892012JoG11J123 2012

Azam M F Wagnon P Vincent C Ramanathan A Linda Aand Singh V B Reconstruction of the annual mass balanceof Chhota Shigri Glacier (Western Himalaya India) since 1969Ann Glaciol in revision 2013

Barrand N and Murray T Multivariate Controls on the In-cidence of Glacier Surging in the Karakoram Himalaya

Arct Antarct Alp Res 38 489ndash498 doi1016571523-0430(2006)38[489MCOTIO]20CO2 2006

Barundun M Huss M Azisov E Gafurov A HoelzleM Merkushkin A Salzmann N and Usubaliev R Re-establishing seasonal mass balance observation at AbramovGlacier Kyrgyzstan from 1968ndash2012 Abstract EGU2013-5668European Geophysical Union Vienna 2013

Benn D Bolch T Hands K Gulley J Luckman ANicholson L Quincey D Thompson S Toumi R andWiseman S Response of debris-covered glaciers in theMount Everest region to recent warming and implicationsfor outburst flood hazards Earth-Sci Rev 114 156ndash174doi101016jearscirev201203008 2012

Berthier E and Vincent C Relative contribution of surface mass-balance and ice-flux changes to the accelerated thinning of Merde Glace French Alps over 1979ndash2008 J Glaciol 58 501ndash512doi1031892012JoG11J083 2012

Berthier E Arnaud Y Baratoux D Vincent C and Reacutemy FRecent rapid thinning of the ldquo Mer de Glace rdquo glacier derivedfrom satellite optical images Geophys Res Lett 31 L17401doi1010292004GL020706 2004

Berthier E Arnaud Y Kumar R Ahmad S Wagnon P andChevallier P Remote sensing estimates of glacier mass balancesin the Himachal Pradesh (Western Himalaya India) RemoteSens Environ 108 327ndash338 doi101016jrse2006110172007

Bhambri R Bolch T Kawishwar P Dobhal D P SrivastavaD and Pratap B Heterogeneity in Glacier response from 1973to 2011 in the Shyok valley Karakoram India The CryosphereDiscuss 6 3049ndash3078 doi105194tcd-6-3049-2012 2012

Bhutiyani M Mass-balance studies on Siachen Glacier in theNubra valley Karakoram Himalaya India J Glaciol 45 112ndash118 1999

Bishop M P Schroder J F and Ward J L SPOT multispec-tral analysis for producing supraglacial debris-load estimates forBatura glacier Pakistan Geocarto Int 10 81ndash90 1995

Bolch T Pieczonka T and Benn D I Multi-decadal mass lossof glaciers in the Everest area (Nepal Himalaya) derived fromstereo imagery The Cryosphere 5 349ndash358 doi105194tc-5-349-2011 2011

Bolch T Kulkarni A Kaumlaumlb A Huggel C Paul F Cogley JFrey H Kargel J Fujita K Scheel M Bajracharya S andStoffel M The state and fate of Himalayan Glaciers Science336 310ndash314 doi101126science1215828 2012

Bollasina M Ming Y and Ramaswamy V Anthropogenicaerosols and the weakening of the South Asian summer mon-soon Science 334 502ndash505 doi101126science12049942011

Bookhagen B and Burbank D Toward a complete Himalayan hy-drological budget spatiotemporal distribution of snowmelt andrainfall and their impact on river discharge J Geophys Res115 F03019 doi1010292009JF001426 2010

Bretherton C Widmann M Dymnikov V Wallace J and BladeacuteI The effective number of spatial degrees of freedom of a time-varying field J Climate 12 1990ndash2009 1999

Copland L Pope S Bishop M Shroder J Clendon PBush A Kamp U Seong Y and Owen L Glacier veloc-ities across the central Karakoram Ann Glaciol 50 41ndash49doi103189172756409789624229 2009

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

1268 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

slope variations Thus this apparent ldquoelevation biasrdquo is cor-rected using the relation between elevation differences andthe maximum terrain curvature estimated over stable areasoff glaciers (Gardelle et al 2012b) Note that the units inFig 1b in Gardelle et al (2012b) should be 10minus2 mminus1 in-stead of mminus1 (A Gardner personal communication 2012)However this does not impact the validity of the correction

334 Radar penetration correction

The SRTM DEM acquired in C-band potentially underesti-mates glacier elevations since at this wavelength (sim 56 cm)the penetration of the radar signal into snow and ice canreach several meters (eg Rignot et al 2001) Here we esti-mate this penetration for each study site by differencing theSRTM C-band (57 GHz) and X-band (97 GHz) DEMs Thelatter was acquired simultaneously with the SRTM X-bandat higher resolution (30 m) but with a narrower swath andthus has a coverage restricted to selected swaths only Sincethe penetration of the X-band is expected to be low com-pared to the C-band (an hypothesis that still needs to be con-firmed Ulaby et al 1986) we consider the elevation differ-ence (SRTMC-bandminus SRTMX-band) to be a first approxima-tion of C-band radar penetration over glaciers (Gardelle etal 2012b) The correction is calculated as a function of alti-tude and applied to each pixel separately

For the Hindu Kush Spiti Lahaul and West Nepal sitesdue to the lack of SRTMX-band data we were unable to de-termine the value of the penetration Thus we applied thecorrection computed over the nearest study site ie Karako-ram for Spiti Lahaul and Hindu Kush and Everest for WestNepal The mean values of the SRTMC-bandpenetrations overglaciers are given for each study site in Table 3

335 Seasonality correction

SPOT5 DEMs were acquired between late October and earlyJanuary depending on the study site Since the SRTM DEMis from mid-February we need to account for possible masschanges during this 1-to-5-month period in order to esti-mate glacier mass balance over an integer number of years(Gardelle et al 2012a) For the eastern sites (from WestNepal to Hengduan Shan) where glaciers are of summer-accumulation types (Fujita 2008) the correction was set to0 This choice is confirmed by recent field observations thatshow no winter accumulation on Mera and Pokalde glaciersin Nepal (Wagnon et al 2013) For the Karakoram studysites we used the winter accumulation rate+013 m we permonth measured on Biafo Glacier (Wake 1989) For theother western study sites (the Pamir Hindu Kush and SpitiLahaul sites) the correction is derived from the mean of win-ter mass balances of 35 glaciers all in the Northern Hemi-sphere+089 m we yrminus1 over 2000ndash2005 corresponding to+015 m we per winter month (Ohmura 2011) The lattervalue is slightly lower than the average winter mass balance

Table 3 Average SRTM C-band penetration estimates (in m)in February 2000 over glaciers in this study and from Kaumlaumlb etal (2012)

Sub-region This study Kaumlaumlb et al (2012)

Hengduan Shan 17 NABhutan 24

25plusmn 05Everest 14West Nepal NA

15plusmn 04Spiti Lahaul NAHindu Kush NA 24plusmn 04Karakoram East

34 24plusmn 03Karakoram WestPamir 18 NA

(+135plusmn 031 m we yrminus1) measured on Abramov Glacier inthe Pamir during 1968ndash1998 (WGMS 2012) and in reason-ably good agreement with the modeled mean winter massbalance (+108plusmn 034 m we yrminus1) of Chhota Shigri Glacierduring 1969ndash2012 in the western Himalaya (Azam et al2013)

336 Mean elevation changes and mass balancecalculation

Before averaging elevation changes we exclude all interpo-lated pixels in the SRTM or SPOT5 DEMs as well as unex-pected elevation changes exceedingplusmn150 m for study sites(Pamir and Karakoram) that include surge-type glaciers orplusmn80 m for other sites Those thresholds were chosen aftercareful inspection of the elevation difference maps and his-tograms in order to identify the largest realistic glacier eleva-tion changes Glaciers that are truncated at the edges of theDEM are also excluded from mass balance calculations asthey may bias the final results if for example only their accu-mulation or ablation area is sampled Although they are trun-cated Siachen (Karakoram East) and Fedtchenko (Pamir)glaciers were retained because of their wide coverage andbecause their accumulation and ablation areas were correctlysampled in the DEMs

Elevation changes are then analyzed for 100 m altitudebins Within each altitude band we average the elevationchanges after excluding pixels where absolute elevation dif-ferences differ by more than three standard deviations fromthe mean (Berthier et al 2004 Gardner et al 2012) Thisis an efficient way to exclude outliers based on the assump-tion that elevation changes should be similar at a given alti-tude within each study site This assumption does not holdfor regions containing actively surging glaciers Thus forthe Pamir and Karakoram study sites surge-type glaciersare treated separately ie elevation changes are averaged by100 m altitude bands for each surge-type glacier individually

Where no elevation change is available for a pixel weassign to it the value of the mean elevation change of the

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1269

altitude band it belongs to in order to assess the mass bal-ance over the whole glacier area

The conversion of elevation changes to mass balance re-quires knowledge of the density of the material that has beenlost or gained and its evolution during the study period Giventhe lack of repeat measurement of density profiles over theentire snowfirnice column in the PKH we applied a con-stant density conversion factor of 850 kg mminus3 for all ourstudy sites (Sapiano et al 1998 Huss 2013)

The mass balance of each surge-type glacier is computedseparately and added (area-weighted) to the mass balance ofthe rest of the glaciers to estimate the mass balance of thewhole study site

For study sites with a high concentration of growingpro-glacial lakes such as Bhutan Everest and West Nepal(Gardelle et al 2011) our mass losses are slightly underes-timated because they do not take into account the glacier icethat has been replaced by water during the expansion of thelake Subaqueous ice losses are only estimated for LhotseSharImja Glacier (Everest area Fujita et al 2009) for thesake of comparison to previous studious

The region-wide mass balance of PKH glaciers as definedin Sect 2 is computed by extrapolating the mass balancefrom each study site to a wider sub-region (Fig 1) that is as-sumed to experience the same glacier mass changes becauseof its similar climate (Bolch et al 2012 Kaumlaumlb et al 2012)The total glacier area for each sub-region is computed fromthe RGI v20 (Arendt et al 2012)

The periods over which we calculate mass balance areslightly different from one site to another (Table 1) depend-ing on the year of acquisition of the SPOT5 DEM (two in2008 two in 2010 and five in 2011) In the following we willassume that all mass balances are representative of the period1999ndash2011 in order to estimate a region-wide mass budgetof PKH glaciers neglecting thus part of the inter-annual vari-ability

34 Accuracy assessment

The errorE1hi to be expected for a pixel elevation change

1hi is assumed to equal the standard deviationσ1h of themean elevation change1h of the altitude band it belongs toThe value ofσ1h can range fromplusmn4 m toplusmn 20 m dependingon the study site and elevation This is rather conservative asthe value ofσ1h contains also the intrinsic natural variabilityof elevation changes within the altitude band (ie it containsboth noise and real geophysical signal)

The errorE1h of the mean elevation change1h in eachaltitude band is then calculated according to standard princi-ples of error propagation

E1h =E1hiradicNeff

(1)

Neff represents the number of independent measurements inthe altitude bandNeff will be lower thanNtot the total num-

ber of elevation change measurements1hi in the altitudeband since the latter are correlated spatially (Bretherton etal 1999)

Neff =Ntot middot PS

2d (2)

where PS is the pixel size (here 40 m)d is the distance ofspatial autocorrelation determined using Moranrsquos I autocor-relation index on elevation differences off glaciers On theaverage over the 9 study sitesd = 492plusmn 72 m

The error on the penetration correction was assumed to besystematic due to the not fully understood physics involvedand equal toplusmn 15 m This error was obtained by compar-ing the SRTM C-band penetration estimated using two inde-pendent methods (i) SRTMC-bandndash SRTMX-band and (ii) dif-ference between ICESat and SRTM when ICESat elevationchange trends during 2003ndash2008 are extrapolated to Febru-ary 2000 (Kaumlaumlb et al 2012) The maximum difference of thepenetration inferred from the two methods is 10 m (Table 3)A 50 additional error margin was added to account for thefact that this comparison could only be made on 2 of ourstudy sites

Given the slender observational support for the seasonal-ity correction (Section 33) we assumed its uncertainty tobe plusmn100 on the western sites ieplusmn015 m we per win-ter month The same absolute uncertainty (plusmn015 m we perwinter month) was used for the eastern sites where no sea-sonality correction is applied but ablation andor accumula-tion may occur in winter (Wagnon et al 2013)

All errors are summed quadratically within each altitudeband and then summed for all altitude bands assuming con-servatively that they are 100 dependent The error on theSRTM penetration correction clearly dominates our errorbudget for volume change

During the conversion from volume to mass we assumea plusmn60 kg mminus3 error on the density conversion factor (Huss2013) Thisplusmn7 error is similar to the one estimated byBolch et al (2011) In a sensitivity test the volume to massconversion is also performed using two alternative densityscenarios (Kaumlaumlb et al 2012) (i) a density of 900 kg mminus3 isassumed everywhere and (ii) 600 kg mminus3 in the accumula-tion area and 900 kg mminus3 in the ablation area These two sce-narios take among other things into account that elevationchanges could be due to changes in glacier dynamics or dueto surface mass balance We then calculate the maximum ab-solute difference between the mass balances calculated usingour preferred scenario (850plusmn 60 kg mminus3 everywhere) and themass balances calculated using the two others scenarios Forthe eastern sites (West Nepal to Hengduan Shan) the max-imum absolute difference is 003 m we yrminus1 For the west-ern sites (Pamir to Spiti Lahaul) it reaches 006 m we yrminus1Those uncertainties due to the choice of a given density sce-nario remain negligible compared to other sources of errors

We also include an error due to extrapolation of our massbalance estimate from our study site (the extent of each

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1270 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

SPOT5 DEM) to the unmeasured glaciers within each sub-region To estimate this regional extrapolation error we com-pare the ICESat elevation change trends (computed as inKaumlaumlb et al 2012) within 3 times 3 cells centered at the studysites (Table 5) with the ICESat-derived trends for the entiresub-regions The absolute difference between both trends isless than 002ndash003 m yrminus1 for Bhutan West Nepal Spiti La-haul and Karakoram ie negligible For Everest the eleva-tion trend for the entire sub-region is 012 m yrminus1 less nega-tive than for the 3 times 3 cells around the SPOT5 DEM cen-ter for Hindu Kush the trend for the entire sub-region is017 m yrminus1 more negative than for the 3 times 3 cell due to thevariability of elevation changes within this sub-region (Kaumlaumlbet al 2012) The regional extrapolation error is obtainedas the area-weighted mean absolute difference between themass balances derived from ICESat for those 3 times 3 cellsand the whole sub-region and equalsplusmn004 m we yrminus1

Finally for sub-regions and total PKH estimates of masschange we include an error on the glacier area computedfrom the RGI v20 (Arendt et al 2012) The inventory er-ror is specific to each study site and evaluated based on thecomparison of our user-defined inventory on each study sitewith the RGI v20 (Table 1) We note the remarkable accu-racy of the RGI for all but one of our study sites The rel-ative errors are generally of a few percents and up to 12 for West Nepal The differences between our and existing in-ventories are probably due to the difficulty of delimitatingdebris-covered glacier parts (eg Frey et al 2012 Paul etal 2013) and accumulation areas The Hengduan Shan studysite is an exception There the RGI is based on the ChineseGlacier Inventory (Shi et al 2009 Arendt et al 2012) andoverestimates the ice-covered area by 88 compared to ourLandsat-based inventory

Unless stated otherwise all error bars from this study andpreviously published studies correspond to one standard er-ror

4 Results

41 Overview of the elevation changes

The maps of glacier elevation changes are given in Figs 2ndash10for our nine study sites with as inset the distribution ofthe elevation differences off glaciers The full maps of ele-vation differences off glaciers are shown in the Supplement(Figs S2ndashS10) Those off glacier maps indicate a local ran-dom noise of aboutplusmn5ndash10 m However at length scales of afew kilometers the spatial variations in elevation differencesare small typically less than 1 m

For the four eastern study sites from the HengduanShan to the West Nepal (Fig 1) the elevation changesaveraged over the accumulation areas are small (between003plusmn 014 m yrminus1 to 013plusmn 015 m yrminus1) whereas clear sur-face lowering is observed for all four study sites in their

ablation areas ranging fromminus050plusmn 014 m yrminus1 (Bhutan)to minus081plusmn 013 m yrminus1 (Hengduan Shan) Over the BhutanEverest and West Nepal study sites glaciers in contact witha proglacial lake often experienced larger thinning rates

Further west the Spiti Lahaul is the only site wheresurface lowering is measured both in the accumula-tion area (minus034plusmn 016 m yrminus1) and the ablation area(minus070plusmn 014 m yrminus1) Clear surface lowering is also ob-served in the ablation areas of the Hindu Kush glaciers(minus050plusmn 018 m yrminus1)

In the Karakoram (east and west Figs 8 and 9) andin the Pamir (Fig 10) the pattern of elevation changesis highly heterogeneous due to the occurrence of numer-ous glacier surges or similar flow instabilities The eleva-tion changes in the accumulation areas of those three studysites are slightly positive (between+022plusmn 014 m yrminus1 and+030plusmn 018 m yrminus1) In the ablation areas small surfacelowering is observed for the two Karakoram study sites (av-erage of the two sitesminus033plusmn 016 m yrminus1) whereas no el-evation change is found in Pamir (minus002plusmn 013 m yrminus1)

Note that those mean rates of elevation changes are sensi-tive to the choice of the ELA assumed to equal the snowlinealtitude in Landsat images from around year 2000

42 Influence of a debris cover

To evaluate thinning rates over debris-covered and debris-free ice we perform a histogram adjustment in order tocompare data sets with similar altitude distributions This isneeded because debris-covered parts tend to be concentratedat lower elevations The adjustment consists in randomly ex-cluding pixels over debris-free ice from each elevation bandto match a scaled version of the debris-covered pixel distribu-tion (Kaumlaumlb et al 2012) Elevation changes with altitude canthen be compared for clean and debris-covered ice (Fig 11)

The thinning is higher for debris-covered ice in the Ever-est area and on the lowermost parts of glaciers in the PamirBut on average in the Pamir western Karakoram and SpitiLahaul the lowering of debris-free and debris-covered ice issimilar The thinning is stronger over clean ice in the Heng-duan Shan Bhutan West Nepal Hindu Kush and althoughto a lesser extent in the eastern Karakoram

43 Mass balance over the nine study sites

Glacier mass changes have been heterogeneous overthe PKH since 1999 Slight mass gains or balancedmass budgets are found in the north-west in thePamir (+014plusmn 013 m we yrminus1) and for the two studysites in the central Karakoram (+009plusmn 018 m we yrminus1

and +011plusmn 014 m we yrminus1) Mass loss is moderatein the adjacent Hindu Kush with a mass balance ofminus012plusmn 016 m we yrminus1 Glaciers in all our eastern studysites (two in Nepal Bhutan and Hengduan Shan) experi-enced homogeneous mass losses with mass balances ranging

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1271

Fig 2 Glacier elevation changes over the Hengduan Shan study site Inset distribution of the elevation differences off glacier The medianand the standard deviation of the elevation difference and the number of pixels off glaciers are given (by construction the mean elevationdifference is null)

Fig 3Same as Fig 2 but for the Bhutan study site

from minus033plusmn 014 m we yrminus1 to minus022plusmn 012 m we yrminus1The most negative mass balance is measured in Spiti Lahaul(western Himalaya) atminus045plusmn 013 m we yrminus1 (Table 4)

However within a study site mass balances can be highlyvariable from one glacier to another In the Everest area twolarge glaciers (71 km2 each) experienced distinctly differentmass balance withminus016plusmn 016 m we yrminus1 for the south-flowing Ngozumpa Glacier andminus059plusmn 016 m we yrminus1

for the north-flowing Rongbuk Glacier (Fig 4) In Bhutan(Fig 3) mass loss is higher south (minus025plusmn 013 m we yrminus1)

than north (minus014plusmn 012 m we yrminus1) of the main Hi-malayan ridge though not at a statistically significant level

The region-wide mass budget of PKH glaciers isnegative minus101plusmn 55 Gt yrminus1 (minus014plusmn 008 m we yrminus1)which corresponds to a sea level rise contribution of0028plusmn 0015 mm yrminus1 over 1999ndash2011

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1272 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 4Same as Fig 2 but for the Everest study site

Fig 5Same as Fig 2 but for the West Nepal study site

44 Glacier surges

In the Pamir and the Karakoram the spatial pattern of ele-vation changes of many glaciers is typical of surge events orflow instabilities They can be identified because of their veryhigh thinning and thickening rates that can both reach up to16 m yrminus1 (Figs 8ndash10)

Most of them have already been reported to be surge-type glaciers based on their velocity (Kotlyakov et al 2008Quincey et al 2011 Copland et al 2009 Heid and Kaumlaumlb2012) morphology (Copland et al 2011 Barrand and Mur-ray 2006 Hewitt 2007) or elevation changes (Gardelle etal 2012a) Here we only identify surge-type glaciers for thesake of mass balance calculation and we do not attempt toprovide an exhaustive up-to-date surge inventory This is why

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1273

Fig 6 Same as Fig 2 but for the Spiti Lahaul study site The map of elevation changes contains many voids due to the presence ofthin clouds during the acquisition of the SPOT5 DEM The mass balance of the Chhota Shigri and Bara Shigri glaciers are respectivelyminus039plusmn 018 m we yrminus1 andminus048plusmn 018 m we yrminus1

the published surge inventories in the Pamir and the Karako-ram do not necessarily match the surge-type glaciers labeledon the elevation change maps as these refer only to our ob-servation period (1999ndash2011)

The mass balance of surge-type glaciers (respectivelynon-surge-type glaciers) is +019plusmn 022 m we yrminus1

(+012plusmn 011 m we yrminus1) in the Pamir+009plusmn 030 m we yrminus1 (+009plusmn 015 m we yrminus1) inthe western Karakoram and+010plusmn 025 m we yrminus1

(+012plusmn 012 m we yrminus1) in the eastern Karakoram Theoverall similarity between the mass budgets for surge-typeand non-surge-type glaciers shows that those flow instabili-ties seem not to affect the glacier-wide mass balance at leastover short time periods here 10 yr

5 Discussion

51 SRTM penetration

Alternative estimates of SRTMC-bandpenetration for the PKHregion were obtained by Kaumlaumlb et al (2012) by comparing el-evations of the ICESat altimeter (Zwally et al 2002) with theSRTMC-bandDEM in PKH and extrapolating the 2003ndash2008trend of elevation changes back to the acquisition date ofSRTM The difference between east and west is more dis-tinct in our case with a mean penetration in the Karakoram of34 m (24plusmn 03 m in Kaumlaumlb et al 2012) 24 m in Bhutan and14 m around the Everest area (25plusmn 05 m for a wider areaincluding east Nepal and Bhutan in Kaumlaumlb et al 2012) Thiseastwest gradient is expected given that in February 2000when the SRTM mission was flown the snowfirn thicknesswas probably larger in the western PKH where glaciers areof winter-accumulation type than in the eastern PKH where

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1274 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 7Same as Fig 2 but for the Hindu Kush study site

glaciers are of summer-accumulation type Further studieswould be necessary to investigate the possible reasons be-hind the relatively low SRTMC-bandpenetration for the Pamir(18 m) that do not follow the eastwest gradient mentionedabove

Both our and Kaumlaumlb et alrsquos (2012) estimates agree withintheir error bars and seem thus to provide robust first-orderestimates for the SRTMC-bandradar penetration in the regionHowever it remains to be verified if a penetration depth com-puted at the regional scale is appropriate for a single smallglacier whose altitude distribution is different from the oneof whole study site (eg Abramov Glacier in the far north ofour Pamir study site)

52 Thinning over debris-covered ice

PKH glaciers are heavily debris-covered in their ablation ar-eas because of steep rock walls that surround them and in-tense avalanche activity Twelve percent (12 ) of the glacierarea in our study sites are covered with debris (up to 22 inthe Everest area Table 1) which is in agreement with pre-vious estimates for the Himalaya and the Karakoram 13 in Kaumlaumlb et al (2012) and 10 in Bolch et al (2012) Thedebris layer is expected to modify the ablation of the under-

lying ice It will increase ablation if its thickness is just a fewcentimeters (dominance of the albedo decrease) or decreaseablation it if the layer is thick enough to protect the ice fromsolar radiation (Mattson et al 1993 Mihalcea et al 2006Benn et al 2012 Lejeune et al 2013)

However recent studies have reported similar overall thin-ning rates over debris-covered and debris-free ice in thePKH (Kaumlaumlb et al 2012 Gardelle et al 2012a) or evenhigher lowering rates over debris-covered parts in the Ever-est area (Nuimura et al 2012) Our findings are consistentwith Nuimura et al (2012) for the Everest study site with astronger thinning (rate of elevation change ofminus097 m yrminus1)

in debris-covered areas than over clean ice (rate of elevationchange ofminus057 m yrminus1) In the Pamir Karakoram and SpitiLahaul study sites these rates are similar while in the HinduKush West Nepal Bhutan and Hengduan Shan study sitesthinning is greater over clean ice in agreement with the com-monly assumed protective effect of debris (Fig 4) Thosedifferences between our 9 study sites suggest a complex re-lationship between debris cover and glacier wastage

As local glacier thickness changes are a result of bothmass balance processes and a dynamic component theseunexpected high thinning rates over debris-covered parts

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1275

Fig 8Same as Fig 2 but for the Karakoram east study site Surge-type glaciers in an active surge phase (resp quiescent phase) are identifiedwith a solid triangle (resp solid circle)

are potentially the result of glacier-wide processes Over aglacier tongue the ablation can be enhanced by the pres-ence of steep exposed ice cliffs or supraglacial lakes andponds (Sakai et al 2000 2002) In addition it is also pos-sible that the glacier tongues are mostly covered by thindebris layers such that the albedo effect dominates the in-sulating effect resulting in an increased tongue-wide abla-tion The prevalence of thin debris was recently suggestedfor debris-covered glaciers around the Mount Gongga in thesouth-eastern Tibetan Plateau (Zhang et al 2013b) Finallyice-flow rates could be very low at debris-covered parts asshown by Quincey et al (2007) in the Everest area Kaumlaumlb(2005) for Bhutan and Scherler et al (2011b) in the HinduKush-Karakoram-Himalaya Thus all three factors are likelyto increase the overall thinning under debris despite the un-doubted protective effect of a continuous thick debris coverat a local scale Future investigations combining surface ve-locity fields and maps of elevation changes could allow eval-uating more closely the relative role of ablation and ice fluxesin thinning rates over PKH glacier tongues (eg Berthierand Vincent 2012 Nuth et al 2012) Measurements of the

spatial distribution of debris thickness over the PKH glaciertongues are also needed

Note that in the case of debris-covered tongues the ele-vation change measurement represents the glacier thicknesschange but also the possible debris thickness evolution Thelatter remains poorly known in the PKH or in any othermountain range But based on the debris discharges givenby Bishop et al (1995) for Batura Glacier (Pakistan) 48to 90times 103 m3 yrminus1 the thickening of the debris can be es-timated between 27 and 55 mm yrminus1 which is negligiblegiven the magnitude of the glacier elevation changes

53 Comparison to previous mass balance estimates

Throughout the PKH there is only a single peer-reviewedglaciological record (on Chhota Shigri Glacier) coveringmost of our study period (Wagnon et al 2007 Vincent etal 2013) Our geodetic mass balance estimate for the SpitiLahaul study site has been recently compared to those glacio-logical measurements on Chhota Shigri Glacier (Vincent etal 2013) Mass balance records reported in Yao et al (2012)for the Tibetan Plateau and the surroundings mountain ranges

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1276 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 9Same as Fig 8 but for the Karakoram west study site

only start during glaciological year 20052006 This is whyhere we do not compare our mass balance estimates toglaciological records and limit our comparison to others re-gional estimates obtained using the geodetic or gravimetricmethods

We stress that the comparison of our mass balance mea-surements to published estimates is complicated by thefact that generally they do not cover the same time pe-riod Little is known about the inter-annual variability inthe PKH during the 21st century due to the limited num-ber of long glaciological time series The 9 yr mass balance

record on Chhota Shigri Glacier reveals a relatively highinter-annual variability of the annual mass balance (stan-dard deviationplusmn 074 m we yrminus1 during 2003ndash2011 Vin-cent et al 2013) which is similar to the one obtained be-tween 1968 and 1998 on Abramov Glacier (standard devi-ation 069 m we yrminus1 WGMS 2012) Thus inter-annualvariability can explain part of the difference between two cu-mulative mass balance estimates over 5ndash10 yr even if they areoffset by only one or two years

Based on ICESat repeat track measurements and theSRTM DEM Kaumlaumlb et al (2012) measured a region-wide

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1277

Fig 10Same as Fig 8 but for the Pamir study site

mass balance ofminus021plusmn 005 m we yrminus1 between 2003and 2008 for Hindu Kush-Karakoram-Himalaya glaciersFor that same area (ie excluding the Pamir and theHengduan Shan study sites) we found a mass balance ofminus015plusmn 007 m we yrminus1 Thus our result agrees with Kaumlaumlbet al (2012) within the error bars although our study periodis longer (1999ndash2011) and our measurement principle andextrapolation approach are different In addition we havecalculated updated mass balances from Kaumlaumlb et al (2012)ie averages over 3times 3 degree cells centered over our studysites and find that they are consistent within error barswith our mass balance estimates (Table 5) Similar resultsare found when our mass balances are compared to a spa-tially extended ICESat analysis for 2003ndash2009 (Gardner etal 2013) Mass losses measured with ICESat (Kaumlaumlb et al2012 Gardner et al 2013) tend to be slightly larger thanours This could be due to the difference in time periodsAlso our smaller mass losses could be partly explained if wesystematically underestimated the poorly constrained pene-

tration of the C-band andor X-band radar signal into ice andsnow

At a more local scale Bolch et al (2011) reported a massbalance ofminus079plusmn 052 m yrminus1 between 2002 and 2007 byDEM differencing over a glacier area of 62 km2 includingten glaciers south and west of Mt Everest For these sameglaciers Nuimura et al (2012) found a mean mass balanceof minus045plusmn 060 m yrminus1 during 2000ndash2008 while they com-puted a mass balance ofminus040plusmn 025 m yrminus1 between 1992and 2008 over a glacier area of 183 km2 around Mt EverestIn the present study we found an average mass balance ofminus041plusmn 021 m yrminus1 over the ten glaciers mentioned above(Table 5 Fig S1) and a value ofminus026plusmn 013 m yrminus1 overthe whole Everest study site between 2000 and 2011 Ourvalues are comparable to those of Nuimura et al (2012)Our mass balance is less negative than the 2002ndash2007 massbalance of Bolch et al (2011) but not statistically differentwhen error bars are considered (Fig 12) Indeed Bolch etal (2011) acknowledged some high uncertainties for their

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1278 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Table 4Glacier mass budget of the nine sub-regions to which the results of the study sites (mass balances) have been extrapolated For eachof them the total glacierized area as well as the percentage of the glacier area over which the mass balance was computed (ie study sites)are given The total glacier area is derived from the RGI v20 (Arendt et al 2012) The error for the mass balance in each sub-region is theroot of sum of squares (RSS) of the mass balance error of each study site and the regional extrapolation error The error for the mass budgetin each sub-region includes additionally the error from the total glacier area in the RGI v20

Sub-region Total glacier Measured glacier Mass balance Mass budgetarea (km2) area () (m we yrminus1) (Gt yrminus1)

Hengduan Shan 11584 12 minus033plusmn 014 minus38plusmn 38Bhutan 4021 34 minus022plusmn 013 minus09plusmn 05Everest 6226 23 minus026plusmn 014 minus16plusmn 08West Nepal 6849 13 minus032plusmn 014 minus22plusmn 10Spiti Lahaul 9043 23 minus045plusmn 014 minus41plusmn 13Hindu Kush 6135 13 minus012plusmn 016 minus07plusmn 10Karakoram East

1902428 +011 plusmn 014 +19 plusmn 31

Karakoram West 30 +009 plusmn 018Pamir 9369 34 +014 plusmn 014 +13 plusmn 13

TotalArea- 72251 30 minus014plusmn 008 minus101plusmn 55weighted average

Table 5Comparison of geodetic mass balances between this study and previous published results with overlapping study periods and similargeographic locations Updated figures from Kaumlaumlb et al (2012) are averaged over 3 times 3 cells centered over the study sites of the presentstudy

GlacierSite name This study Bolch et al (2011) Nuimura et al (2012) Kaumlaumlb et al (2012 updated)2002ndash2007 2000ndash2008 2003ndash2008

Study Mass balance Area (km2)a Mass balance Area Mass balance Mass balanceperiod (m we yrminus1) (m we yrminus1) (km2) (m we yrminus1) (m we yrminus1)

Hengduan San 1999ndash2010 minus022 plusmn 014 1303 (55 )Bhutan 1999ndash2010 minus022 plusmn 012 1384 (64 ) minus052 plusmn 016b

Everest

1999ndash2011

minus026 plusmn 013 1461 (58 ) minus039 plusmn 011AX010 minus090plusmn 034 04Changri SharNup minus042plusmn 017 161 (79 ) minus029plusmn 052 130 minus055plusmn 038Khumbu minus051plusmn 019 203 (47 ) minus045plusmn 052 170 minus076plusmn 052Nuptse minus037plusmn 020 50 (74 ) minus040plusmn 053 40 minus034plusmn 027Lhotse Nup minus021plusmn 027 24 (70 ) minus103plusmn 051 19 minus022plusmn 047Lhotse minus043plusmn 018 85 (83 ) minus110plusmn 052 65 minus067plusmn 051Lhotse SharImja minus070plusmn 052 98 (44 ) minus145plusmn 052 107 minus093plusmn 060Amphu Laptse minus046plusmn 034 25 (38 ) minus077plusmn 052 15 minus018plusmn 094Chukhung +044 plusmn 024 42 (47 ) +004 plusmn 054 38 +043 plusmn 081Ama Dablam minus049plusmn 017 36 (54 ) minus056plusmn 052 22 minus056plusmn 073Duwo minus016plusmn 026 19 (18 ) minus196plusmn 053 10 minus068plusmn 074Total Khumbu minus041plusmn 021 744 minus079plusmn 052 617 minus045plusmn 060(10 Glaciers above)

West Nepal 1999ndash2011 minus032 plusmn 013 908 (40 ) minus032 plusmn 012Spiti Lahaul 1999ndash2011 minus045 plusmn 013 2110 (46 ) minus038 plusmn 006Karakoram East 1999ndash2010 +011 plusmn 014 5328 (42 ) minus004 plusmn 004Karakoram West 1999ndash2008 +009 plusmn 018 5434 (45 )Hindu Kush 1999ndash2008 minus012 plusmn 016 793 (80 ) minus020 plusmn 006Pamir 1999ndash2011 +014 plusmn 013 3178 (50 )

a The total glacier area is given in km2 and in parenthesis the of the glacier area actually covered with measurementsb For this cell the ICESat coverage is insufficient and does not sample all glacier elevations

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1279

Fig 11 Elevation changes (SPOT5-SRTM) in the ablation area for clean ice (blue circles) and debris-covered ice (black circles) for eachstudy site For the Karakoram (east and west) and the Pamir study sites only non-surge-type glaciers are considered Standard error of theelevation differences are typically less thanplusmn2 m with each 100 m elevation band (not shown for the sake of clarity) Note elevation changesover separate sections of a glacier cannot be treated as mass changes due to the disregard of glacier dynamics

estimate over this short time period The differences in sur-vey periods and in the glacier outlines may also explainpart of these discrepancies In particular the delimitation ofthe accumulation area is notoriously difficult and Bolch etal (2011) did not include some of the steep high altitudeslopes which we included due to their potential avalanchecontribution For the 10 glaciers mentioned above our massbalances would become 012 m we yrminus1 more negative ifwe used the exact same outlines as Bolch et al (2011)Our positive mass balance for the small Chukhung Glacier(+044plusmn 024 m we yrminus1) is in agreement with Nuimura etal (2012) but enigmatic in a context of regional glacier lossand deserves further investigation If the 2002ndash2007 massbalance estimate by Bolch et al (2011) is excluded all otherrecent mass balance estimates suggest no acceleration of themass loss in the Everest area when compared to the long-term(1970ndash2007) mass balance measured by Bolch et al (2011)in this regionminus032plusmn 008 m weyrminus1

In Bhutan our results indicate a stronger mass loss forsouthern glaciers than for northern ones Interestingly in thisarea Kaumlaumlb (2005) measured high speeds (up to 200 m yrminus1)

for glaciers north of the Himalayan main ridge and ratherlow velocities over southern glacier tongues by using repeat

ASTER data This is consistent with the more negative massbalance that we report for the southern glaciers in Bhutanie for the slow flowing presumably downwasting glaciers

Berthier et al (2007) reported a mass balance betweenminus069 andminus085 m we yrminus1 for an ice-covered area of915 km2 around Chhota Shigri Glacier between 1999 (SRTMDEM) and 2004 (SPOT5-HRG DEM) For the same areawe found a mass balance ofminus045plusmn 013 m we yrminus1 over1999ndash2011 The difference in the study period may explainpart of the differences Also the methodology by Berthieret al (2007) did not explicitly take into account the SRTMradar penetration over glaciers and also corrected an appar-ent elevation-dependent bias according to glacier elevationswhich is now known as a resolution issue and is best cor-rected according to the terrain maximum curvature (Gardelleet al 2012b) More details and discussions about these dif-ferences can be found in Vincent et al (2013)

Importantly the results of the eastern Karakoram studysite confirm the balanced or slightly positive mass budgetreported by Gardelle et al (2012a) in the western Karako-ram over the past decade The two Karakoram study sitesare adjacent with a small overlapping area (sim 1100 km2 ofglaciers) where elevation differences agree within 1 m Both

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1280 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 12 Comparison of geodetic mass balances for 10 glaciers inthe Mount Everest area 1999ndash2010 mass balances measured inthe present study are compared with published estimates during2002ndash2007 (Bolch et al 2011) and during 2000ndash2008 (Nuimuraet al 2012) The 1-to-1 line is drawn The mean difference (plusmn1standard deviation) between (i) our values and Bolch et al (2011)are (047plusmn 058 m we yrminus1) and (ii) our values and Nuimura etal (2012) are (012plusmn 021 m we yrminus1) Part of the differences be-tween estimates may be due to the different periods surveyed andthe different glacier outlines used

sites show similar mass balances over slightly different pe-riods +011plusmn 014 m yrminus1 over 1999ndash2010 in the east and+009plusmn 018 m yrminus1 over 1999ndash2008 in the west This con-firms a ldquoKarakoram anomalyrdquo that was first suggested basedon field observations (Hewitt 2005) We report a mass bal-ance of+010plusmn 013 m yrminus1 for Baltoro Glacier (see loca-tion in Fig 8 size= 304 km2) between 1999 and 2010which is consistent with its gradual speed-up reported since2003 (Quincey et al 2009) However given the completelydifferent methods used in our and the latter study we can-not conclude that this demonstrates a real shift from nega-tive to positive mass balance The mass budget of SiachenGlacier (1145 km2 sometimes referred to as the highestbattlefield in the world) is also balanced or slightly pos-itive (+014plusmn 014 m we yrminus1) Thus the alarming state-ment by Rasul (2012) that this glacier retreated by 59 kmand lost 17 of its mass during 1989ndash2009 is very likelyerroneous Again within error bars our regional mass bal-ance in the Karakoram agrees with the mass balance ofminus003plusmn 004 m yrminus1 found by Kaumlaumlb et al (2012) over 2003ndash2008 and with the mass balance ofminus010plusmn 018 m we yrminus1

(2-sigma error level) obtained by Gardner et al (2013) for aregion including both the Karakoram and the Hindu Kush

In the northernmost part of the Pamir in the Alay moun-tain range the mass balance of Abramov Glacier has beenmeasured in the field for 30 yr between 1968 and 1998(WGMS 2012) with a mean value ofminus046 m yrminus1 Herewe report a mass balance ofminus003plusmn 014 m yrminus1 for this

glacier over 1999ndash2011 Our near zero mass budget estimatefor this glacier disagrees with negative mass balances recon-structed for the same period with a model run using biascorrected NCEPNCAR re-analysis data and calibrated withfield data (Barundun et al 2013) An underestimate of ourSRTM C-Band penetration in the Pamir as a whole and inparticular for the small Abramov Glacier may contribute tothese differences

For the whole Pamir study site our mass budget is bal-anced or slightly positive (+014plusmn 013 m we yrminus1) In theeastern Pamir the mass balance of Muztag Ata Glacier was+025 m we yrminus1 between 2005 and 2010 (Yao et al 2012)Although Muztag Ata Glacier is located outside of our Pamirstudy site and his glaciological records only cover the sec-ond half of our study period this measurement is consistentwith our remotely sensed mass balance Thus the ldquoKarako-ram anomalyrdquo should probably be renamed the ldquoPamir-Karakoram anomalyrdquo

This slightly positive or balanced mass budget measuredin the Pamir seems to contradict previous findings by Heidand Kaumlaumlb (2012) They reported rapidly decreasing glacierspeeds in this region between two snapshots of annual ve-locities in 20002001 and 20092010 an observation thatwould fit with negative mass balances (Span and Kuhn 2003Azam et al 2012) We note though that the relationship be-tween glacier mass balance and ice fluxes is not straightfor-ward and might in particular involve a time delay betweena change in mass balance and the related dynamic response(Heid and Kaumlaumlb 2012) Thus the decrease in speed couldwell be due to negative mass balances before or partiallybefore 1999ndash2011 We acknowledge that this discrepancyneeds to be investigated further in particular by examiningcontinuous changes in annual glacier speed through the firstdecade of the 21st Century Indeed Heid and Kaumlaumlb (2012)measured differences between two annual periods which maynot represent mean decadal velocity changes

Jacob et al (2012) using satellite gravimetry (GRACE)measured a mass budget ofminus5plusmn 6 Gt yrminus1 over 2003ndash2010for the ldquoKarakoram-Himalayardquo their region 8c which cor-responds to our study area excluding the Pamir Karakoramand Hindu Kush sub-regions but including some glaciersnorth of the Himalayan ridge already on the Tibetan PlateauFor this subset of our PKH study area we measured amass budget ofminus126plusmn 42 Gt yrminus1 The two estimates over-lap within their error bars especially if our error is dou-bled to match the two standard error level used by Jacob etal (2012) The estimate of Jacob et al (2012) over our com-plete study region (PKH) ieminus6plusmn 8 Gt yrminus1 (sum of theirregions 8b and 8c) also includes mass changes for the KunlunShan sub-region for which we did not measure glacier massbalances Therefore the comparison with our PKH value(minus101plusmn 55 Gt yrminus1) is not straightforward but suggests areasonable agreement especially if we consider the slightmass gain (15plusmn 17 Gt yrminus1) measured with ICESat in theWest Kunlun (Gardner et al 2013)

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1281

Table 6 Annual average of the glacier seasonal and decadal mass loss contributions to Indus Ganges and Brahmaputra discharges Basinarea and seasonal contributions are given by Kaser et al (2010) glacier area in each basin is derived from the RGI v20 (Arendt et al 2012)Numbers in parentheses indicate the percentage of glacierized area within the river basin Seasonal contributions were modeled by Kaser etal (2010) based on the CRU 20 dataset over 1961ndash1990 assuming a balanced glacier mass budget Glacier decadal mass loss contributionsare given as annual average over 1999ndash2011

River Basin area Glacier area Glacier contribution Mean discharge(km2) (km2) (m3 sminus1) (m3 sminus1)

Seasonally delayed Decadal mass lossKaser et al (2010) (this study)

Indus 1 139 814 25 598 (2 ) 105 103plusmn 72 2146 (Chevallier personalcommunication 2012)

Ganges 1 023 609 11 168 (1 ) 47 103plusmn 49 11739 Papa et al (2012Brahmaputra 527 666 15 296 (3 ) 33 147plusmn 64 19710 updated)

54 Reasons for the heterogeneous pattern of massbalance

The PKH glacier mass balance is slightly negative(minus014plusmn 008 m we yrminus1) but changes are not homoge-neous from one sub-region to another and seem to coincidewith the climatic settings of the mountain range For the east-ern sites (from the Hengduan Shan to West Nepal) which areunder the influence of the Indian and south-east Asian sum-mer monsoons the mass balances are moderately negative(sim minus026 m we yrminus1) In the northwest of the PKH at thePamir and Karakoram sites where glacier climate is char-acterized by the westerlies in winter mass balances are bal-anced or slightly positive (sim +010 m we yrminus1) In betweenthese two influences the glaciers of the Spiti Lahaul sitein a monsoon-arid transition zone experienced the strongestmass loss (minus045plusmn 013 m we yrminus1)

This heterogeneous pattern of mass balances seems to berelated to trends in precipitation throughout the PKH Archerand Fowler (2004) reported an increase in winter precipi-tation over the Upper Indus Basin since 1961 a trend con-firmed by Yao et al (2012) over the Pamir and the Karako-ram based on the analysis of the Global Precipitation Cli-matology Project data set (GPCP Adler et al 2003) Thisprecipitation increase is a source of greater accumulation forglaciers and thus a possible explanation for their slightly pos-itive mass budgets In addition in the Upper Indus Basin themean summer temperature has been decreasing since 1961(Fowler and Archer 2006) which is consistent with the find-ings of Shekhar et al (2010) in the Karakoram who reporteda decrease in both maximum and minimum temperature since1988 These increases in winter precipitation and summertemperature likely translate in greater accumulation and re-duced ablation changes which are consistent with the bal-anced or slightly positive glacier mass budget

On the other hand in the east of the PKH the Indiansummer monsoon has been weakening since the 1950s (Bol-lasina et al 2011) which could have reduced the amount

of snowfall over the eastern study sites and thus resultedin negative mass balances In addition maximum and min-imum temperatures increased since 1988 in the western Hi-malaya (Shekhar et al 2010) and maximum temperaturesincreased in the central Himalaya (Nepal) since the mid-1970s (Shrestha et al 1999) Thus both precipitation andtemperature trends in the Himalaya likely contributed to thenegative mass balances measured over the correspondingstudy sites (from Spiti Lahaul to Hengduan Shan Fig 1)

However gridded data sets such as GPCP are not neces-sarily representative of the climate at high altitude IndeedHewitt (2005) reported a 5-to-10-fold increase in precipi-tation between the glacier front and the accumulation areain the Karakoram that is not captured by reanalysis dataas recently confirmed by Immerzeel et al (2012) Thus di-rect measurements of accumulation on High Mountain Asiaglaciers (eg Azam et al 2013) and high-altitude weatherstations are eagerly needed to validate the large scale grid-ded data set (eg reanalysis) test their ability to describe thespecific climate near to PKH glaciers and assess the relativerole played by temperature and precipitation changes

55 Contribution to water resources

Our glacier mass balance can be averaged over large riverbasins to estimate the mean contribution to river runoffdue to a decade of glacier mass loss We assume therebyonly direct river runoff without loss terms For the threerivers for which we cover the entire glacierized area of theircatchments within our sub-regions (Fig 1) these contribu-tions to the river discharge are equal to 103 m3 sminus1 (Indus)103 m3 sminus1 (Ganges) and 147 m3 sminus1 (Brahmaputra) for theperiod 1999ndash2011 (Table 6)

Rivers of PKH being key water resources measures oftheir discharges are highly sensitive and difficult to access forpolitical reasons Papa et al (2012) built recent time series ofdischarges for Ganges and Brahmaputra by using altimetrymeasurements The locations of the corresponding dischargeestimates in Bangladesh are given in Fig 1 We can therefore

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1282 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

evaluate the relative contribution of glacier decadal mass loss(Table 6) to annual river run-off at 09 for the Gangesat Hardinge (basin area ofsim 1 020 000 km2) and 07 forBrahmaputra at Bahadurabad (basin area ofsim 530 000 km2)between 1999 and 2011 Given the discharge measured at theGuddu station by the Surface Water Hydrology Project of theWater and Power Development Authority (Pakistan see lo-cation on Fig 1) between 1999 and 2003 we can estimatethe contribution of glacier mass loss at 48 for the IndusRiver at this station

The seasonal contributions of glaciers to river run-off (iethe part of the annual precipitation that experiences season-ally delayed release from the glaciers) directly results fromseasonal variations of meteorological quantities in contrastto the glacier mass loss contribution which results from thelong-term climatic change Both runoff components directlyaffect the amount of water available in a river at a given lo-cation and at a given time so that their comparison could beof interest Even if the seasonal timing of the glacier massloss contribution is unclear it could in general peak at asimilar time of the year than the seasonal glacier contribu-tion so that both components rather enlarge each other thancancel Seasonal glacier contributions have been modeled byKaser et al (2010) The latter study assumed glaciers to bein equilibrium with the present climate and thus prescribedglacier mass balances equal to 0 Thus their seasonally de-layed glacier contributions do not include the part due todecadal glacier mass loss and is complementary to our es-timates (Table 6) For the eastern basins under the influenceof the Indian summer monsoon (Ganges and Brahmaputra)the contribution of glacier mass loss is higher than the sea-sonal contribution In other words for the first decade of the21st century the mass loss of glaciers is on average moreimportant for water availability in those two rivers than theirseasonal water storage effect The situation is different for theIndus Basin where the glacier melt season is also the dry sea-son which results in a relatively higher seasonally delayedglacier contribution to the river discharge There the season-ally delayed water release from the glaciers and the decadalglacier mass loss contribute on average equally to the riverdischarge Both contributions are strongly (and equally) de-pendent on the location of the discharge measurement andwill be significantly larger (in relative term) in more heav-ily glacierized and smaller mountain catchments located inthe upper part of these major river basins (Bookhagen andBurbank 2010)

6 Conclusion

In this study we assessed the spatial pattern of glaciermass balance over the PKH by comparing the SRTM andSPOT5 DEMs over 9 well-distributed study sites between1999 and 2011 We found slightly positive or balanced massbudgets in the western part (Pamir Karakoram) moder-ate mass loss in the east (Nepal Bhutan Hengduan Shan)

and the most negative mass balance in the monsoon-aridtransition zone (Spiti Lahaul in the western Himalaya)By extrapolating these values to climatically homogeneoussub-regions we estimate the PKH overall mass balanceto have beenminus014plusmn 008 m we yrminus1 between 1999 and2011 which corresponds to a contribution to sea level riseof 0028plusmn 0015 mm yrminus1 This is about 4 of the globalglacier mass loss estimated by Gardner et al (2013) thoughover a slightly shorter time period (2003ndash2009) The largestsource of uncertainties in those geodetic mass balance esti-mates is the poorly constrained penetration of the C-BandSRTM radar signal into snow and ice

Our technique of DEM differencing allows mapping in de-tail the spatial pattern of glacier elevation changes Thosemaps reveal many surge-type behaviors both in the Pamirand the Karakoram It also appears that the glacier-wide massbalance does not seem to be considerably affected by themass transfer from surges over thesim 10 yr time period stud-ied Similar lowering rates over debris-covered and clean iceare observed for four study sites despite the commonly as-sumed insulating effect of debris covers This suggests forthe scale of entire glacier tongues a spatial mixture of re-duced ablation under intact thick debris covers and enhancedablation by supraglacial lakes or ice cliffs or due to thin de-bris layer and very low glacier velocities in ablation areasthat reduce the ice advection downstream

The regionally heterogeneous pattern of ice wastage is inbroad agreement with contrasted trends in large-scale pre-cipitation and temperature However glaciological (eg ac-cumulation) and meteorological records are needed at highaltitude in the vicinity of glaciers to evaluate more closelythe local influence of atmospheric variables on glacier massbalance and fully understand the reasons behind those con-trasted mass budgets in the PKH

Supplementary material related to this article isavailable online at httpwwwthe-cryospherenet712632013tc-7-1263-2013-supplementpdf

Acknowledgements We thank G Cogley A Gardner T NuimuraT Bolch P Wagnon M Barandun M Hoelzle M Huss and ananonymous referee for their constructive comments that led to aconsiderably improved manuscript We thank the Water and PowerDevelopment Authority (WAPDA) for their hydrological data(processed by P Chevallier) and F Papa for sharing his updatedrun-off data ahead of publication We thank T Bolch for sharinghis glacier outlines from the Everest area We thank the USGSfor allowing free access to their Landsat archive CGIAR-SCI forSRTM C-band data DLR for SRTM X-band data and GLIMS forglacier inventories J Gardelle acknowledges a PhD fellowshipfrom the French Space Agency (CNES) and the French NationalResearch Center (CNRS) E Berthier acknowledges support fromCNES through the TOSCA and ISIS proposal no 397 and fromthe Programme National de Teacuteleacutedeacutetection Spatiale E Berthieracknowledges M Masiokas and R Villalba for welcoming him

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1283

at the IANIGLA (Mendoza Argentina) Y Arnaud acknowledgessupport from French National Research Agency through ANR-09-CEP-005-01PAPRIKA A Kaumlaumlb acknowledges support fromESA through Glaciers_cci (400010177810I-AM) and from theEuropean Research Council under the European Unionrsquos SeventhFramework Programme (FP2007-2013)ERC Grant Agreementn 320816

Edited by I M Howat

The publication of this articleis financed by CNRS-INSU

References

Adler R Huffman G Chang A Ferraro R Xie P JanowiakJ Rudolf B Schneider U Curtis S Bolvin D Gruber ASusskind J Arkin P and Nelkin E The Version-2 GlobalPrecipitation Climatology Project (GPCP) Monthly Precipita-tion Analysis (1979ndashPresent) J Hydrometeorol 4 1147ndash11672003

Archer D R and Fowler H J Spatial and temporal variationsin precipitation in the Upper Indus Basin global teleconnectionsand hydrological implications Hydrol Earth Syst Sci 8 47ndash61doi105194hess-8-47-2004 2004

Arendt A Bolch T Cogley J G Gardner A Hagen J-OHock R Kaser G Pfeffer W T Moholdt G Paul F RadicV Andreassen L Bajracharya S Beedle M Berthier EBhambri R Bliss A Brown I Burgess E Burgess DCawkwell F Chinn T Copland L Davies B de AngelisH Dolgova E Filbert K Forester R Fountain A Frey HGiffen B Glasser N Gurney S Hagg W Hall D Hari-tashya U K Hartmann G Helm C Herreid S Howat IKapustin G Khromova T Kienholz C Koenig M KohlerJ Kriegel D Kutuzov S Lavrentiev I LeBris R Lund JManley W Mayer C Miles E Li X Menounos B Mer-cer A Moelg N Mool P Nosenko G Negrete A NuthC Pettersson R Racoviteanu A Ranzi R Rastner P RauF Rich J Rott H Schneider C Seliverstov Y Sharp MSigurethsson O Stokes C Wheate R Winsvold S WolkenG Wyatt F and Zheltyhina N Randolph Glacier Inventory[v20] A Dataset of Global Glacier Outlines Global Land IceMeasurements from Space Boulder Colorado USA Digital Me-dia httpwwwglimsorgRGIrandolphhtml 2012

Azam M Wagnon P Ramanathan A Vincent C Sharma PArnaud Y Linda A Pottakkal J Chevallier P Singh V andBerthier E From balance to imbalance a shift in the dynamicbehaviour of Chhota Shigri glacier western Himalaya India JGlaciol 58 315ndash324 doi1031892012JoG11J123 2012

Azam M F Wagnon P Vincent C Ramanathan A Linda Aand Singh V B Reconstruction of the annual mass balanceof Chhota Shigri Glacier (Western Himalaya India) since 1969Ann Glaciol in revision 2013

Barrand N and Murray T Multivariate Controls on the In-cidence of Glacier Surging in the Karakoram Himalaya

Arct Antarct Alp Res 38 489ndash498 doi1016571523-0430(2006)38[489MCOTIO]20CO2 2006

Barundun M Huss M Azisov E Gafurov A HoelzleM Merkushkin A Salzmann N and Usubaliev R Re-establishing seasonal mass balance observation at AbramovGlacier Kyrgyzstan from 1968ndash2012 Abstract EGU2013-5668European Geophysical Union Vienna 2013

Benn D Bolch T Hands K Gulley J Luckman ANicholson L Quincey D Thompson S Toumi R andWiseman S Response of debris-covered glaciers in theMount Everest region to recent warming and implicationsfor outburst flood hazards Earth-Sci Rev 114 156ndash174doi101016jearscirev201203008 2012

Berthier E and Vincent C Relative contribution of surface mass-balance and ice-flux changes to the accelerated thinning of Merde Glace French Alps over 1979ndash2008 J Glaciol 58 501ndash512doi1031892012JoG11J083 2012

Berthier E Arnaud Y Baratoux D Vincent C and Reacutemy FRecent rapid thinning of the ldquo Mer de Glace rdquo glacier derivedfrom satellite optical images Geophys Res Lett 31 L17401doi1010292004GL020706 2004

Berthier E Arnaud Y Kumar R Ahmad S Wagnon P andChevallier P Remote sensing estimates of glacier mass balancesin the Himachal Pradesh (Western Himalaya India) RemoteSens Environ 108 327ndash338 doi101016jrse2006110172007

Bhambri R Bolch T Kawishwar P Dobhal D P SrivastavaD and Pratap B Heterogeneity in Glacier response from 1973to 2011 in the Shyok valley Karakoram India The CryosphereDiscuss 6 3049ndash3078 doi105194tcd-6-3049-2012 2012

Bhutiyani M Mass-balance studies on Siachen Glacier in theNubra valley Karakoram Himalaya India J Glaciol 45 112ndash118 1999

Bishop M P Schroder J F and Ward J L SPOT multispec-tral analysis for producing supraglacial debris-load estimates forBatura glacier Pakistan Geocarto Int 10 81ndash90 1995

Bolch T Pieczonka T and Benn D I Multi-decadal mass lossof glaciers in the Everest area (Nepal Himalaya) derived fromstereo imagery The Cryosphere 5 349ndash358 doi105194tc-5-349-2011 2011

Bolch T Kulkarni A Kaumlaumlb A Huggel C Paul F Cogley JFrey H Kargel J Fujita K Scheel M Bajracharya S andStoffel M The state and fate of Himalayan Glaciers Science336 310ndash314 doi101126science1215828 2012

Bollasina M Ming Y and Ramaswamy V Anthropogenicaerosols and the weakening of the South Asian summer mon-soon Science 334 502ndash505 doi101126science12049942011

Bookhagen B and Burbank D Toward a complete Himalayan hy-drological budget spatiotemporal distribution of snowmelt andrainfall and their impact on river discharge J Geophys Res115 F03019 doi1010292009JF001426 2010

Bretherton C Widmann M Dymnikov V Wallace J and BladeacuteI The effective number of spatial degrees of freedom of a time-varying field J Climate 12 1990ndash2009 1999

Copland L Pope S Bishop M Shroder J Clendon PBush A Kamp U Seong Y and Owen L Glacier veloc-ities across the central Karakoram Ann Glaciol 50 41ndash49doi103189172756409789624229 2009

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1269

altitude band it belongs to in order to assess the mass bal-ance over the whole glacier area

The conversion of elevation changes to mass balance re-quires knowledge of the density of the material that has beenlost or gained and its evolution during the study period Giventhe lack of repeat measurement of density profiles over theentire snowfirnice column in the PKH we applied a con-stant density conversion factor of 850 kg mminus3 for all ourstudy sites (Sapiano et al 1998 Huss 2013)

The mass balance of each surge-type glacier is computedseparately and added (area-weighted) to the mass balance ofthe rest of the glaciers to estimate the mass balance of thewhole study site

For study sites with a high concentration of growingpro-glacial lakes such as Bhutan Everest and West Nepal(Gardelle et al 2011) our mass losses are slightly underes-timated because they do not take into account the glacier icethat has been replaced by water during the expansion of thelake Subaqueous ice losses are only estimated for LhotseSharImja Glacier (Everest area Fujita et al 2009) for thesake of comparison to previous studious

The region-wide mass balance of PKH glaciers as definedin Sect 2 is computed by extrapolating the mass balancefrom each study site to a wider sub-region (Fig 1) that is as-sumed to experience the same glacier mass changes becauseof its similar climate (Bolch et al 2012 Kaumlaumlb et al 2012)The total glacier area for each sub-region is computed fromthe RGI v20 (Arendt et al 2012)

The periods over which we calculate mass balance areslightly different from one site to another (Table 1) depend-ing on the year of acquisition of the SPOT5 DEM (two in2008 two in 2010 and five in 2011) In the following we willassume that all mass balances are representative of the period1999ndash2011 in order to estimate a region-wide mass budgetof PKH glaciers neglecting thus part of the inter-annual vari-ability

34 Accuracy assessment

The errorE1hi to be expected for a pixel elevation change

1hi is assumed to equal the standard deviationσ1h of themean elevation change1h of the altitude band it belongs toThe value ofσ1h can range fromplusmn4 m toplusmn 20 m dependingon the study site and elevation This is rather conservative asthe value ofσ1h contains also the intrinsic natural variabilityof elevation changes within the altitude band (ie it containsboth noise and real geophysical signal)

The errorE1h of the mean elevation change1h in eachaltitude band is then calculated according to standard princi-ples of error propagation

E1h =E1hiradicNeff

(1)

Neff represents the number of independent measurements inthe altitude bandNeff will be lower thanNtot the total num-

ber of elevation change measurements1hi in the altitudeband since the latter are correlated spatially (Bretherton etal 1999)

Neff =Ntot middot PS

2d (2)

where PS is the pixel size (here 40 m)d is the distance ofspatial autocorrelation determined using Moranrsquos I autocor-relation index on elevation differences off glaciers On theaverage over the 9 study sitesd = 492plusmn 72 m

The error on the penetration correction was assumed to besystematic due to the not fully understood physics involvedand equal toplusmn 15 m This error was obtained by compar-ing the SRTM C-band penetration estimated using two inde-pendent methods (i) SRTMC-bandndash SRTMX-band and (ii) dif-ference between ICESat and SRTM when ICESat elevationchange trends during 2003ndash2008 are extrapolated to Febru-ary 2000 (Kaumlaumlb et al 2012) The maximum difference of thepenetration inferred from the two methods is 10 m (Table 3)A 50 additional error margin was added to account for thefact that this comparison could only be made on 2 of ourstudy sites

Given the slender observational support for the seasonal-ity correction (Section 33) we assumed its uncertainty tobe plusmn100 on the western sites ieplusmn015 m we per win-ter month The same absolute uncertainty (plusmn015 m we perwinter month) was used for the eastern sites where no sea-sonality correction is applied but ablation andor accumula-tion may occur in winter (Wagnon et al 2013)

All errors are summed quadratically within each altitudeband and then summed for all altitude bands assuming con-servatively that they are 100 dependent The error on theSRTM penetration correction clearly dominates our errorbudget for volume change

During the conversion from volume to mass we assumea plusmn60 kg mminus3 error on the density conversion factor (Huss2013) Thisplusmn7 error is similar to the one estimated byBolch et al (2011) In a sensitivity test the volume to massconversion is also performed using two alternative densityscenarios (Kaumlaumlb et al 2012) (i) a density of 900 kg mminus3 isassumed everywhere and (ii) 600 kg mminus3 in the accumula-tion area and 900 kg mminus3 in the ablation area These two sce-narios take among other things into account that elevationchanges could be due to changes in glacier dynamics or dueto surface mass balance We then calculate the maximum ab-solute difference between the mass balances calculated usingour preferred scenario (850plusmn 60 kg mminus3 everywhere) and themass balances calculated using the two others scenarios Forthe eastern sites (West Nepal to Hengduan Shan) the max-imum absolute difference is 003 m we yrminus1 For the west-ern sites (Pamir to Spiti Lahaul) it reaches 006 m we yrminus1Those uncertainties due to the choice of a given density sce-nario remain negligible compared to other sources of errors

We also include an error due to extrapolation of our massbalance estimate from our study site (the extent of each

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1270 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

SPOT5 DEM) to the unmeasured glaciers within each sub-region To estimate this regional extrapolation error we com-pare the ICESat elevation change trends (computed as inKaumlaumlb et al 2012) within 3 times 3 cells centered at the studysites (Table 5) with the ICESat-derived trends for the entiresub-regions The absolute difference between both trends isless than 002ndash003 m yrminus1 for Bhutan West Nepal Spiti La-haul and Karakoram ie negligible For Everest the eleva-tion trend for the entire sub-region is 012 m yrminus1 less nega-tive than for the 3 times 3 cells around the SPOT5 DEM cen-ter for Hindu Kush the trend for the entire sub-region is017 m yrminus1 more negative than for the 3 times 3 cell due to thevariability of elevation changes within this sub-region (Kaumlaumlbet al 2012) The regional extrapolation error is obtainedas the area-weighted mean absolute difference between themass balances derived from ICESat for those 3 times 3 cellsand the whole sub-region and equalsplusmn004 m we yrminus1

Finally for sub-regions and total PKH estimates of masschange we include an error on the glacier area computedfrom the RGI v20 (Arendt et al 2012) The inventory er-ror is specific to each study site and evaluated based on thecomparison of our user-defined inventory on each study sitewith the RGI v20 (Table 1) We note the remarkable accu-racy of the RGI for all but one of our study sites The rel-ative errors are generally of a few percents and up to 12 for West Nepal The differences between our and existing in-ventories are probably due to the difficulty of delimitatingdebris-covered glacier parts (eg Frey et al 2012 Paul etal 2013) and accumulation areas The Hengduan Shan studysite is an exception There the RGI is based on the ChineseGlacier Inventory (Shi et al 2009 Arendt et al 2012) andoverestimates the ice-covered area by 88 compared to ourLandsat-based inventory

Unless stated otherwise all error bars from this study andpreviously published studies correspond to one standard er-ror

4 Results

41 Overview of the elevation changes

The maps of glacier elevation changes are given in Figs 2ndash10for our nine study sites with as inset the distribution ofthe elevation differences off glaciers The full maps of ele-vation differences off glaciers are shown in the Supplement(Figs S2ndashS10) Those off glacier maps indicate a local ran-dom noise of aboutplusmn5ndash10 m However at length scales of afew kilometers the spatial variations in elevation differencesare small typically less than 1 m

For the four eastern study sites from the HengduanShan to the West Nepal (Fig 1) the elevation changesaveraged over the accumulation areas are small (between003plusmn 014 m yrminus1 to 013plusmn 015 m yrminus1) whereas clear sur-face lowering is observed for all four study sites in their

ablation areas ranging fromminus050plusmn 014 m yrminus1 (Bhutan)to minus081plusmn 013 m yrminus1 (Hengduan Shan) Over the BhutanEverest and West Nepal study sites glaciers in contact witha proglacial lake often experienced larger thinning rates

Further west the Spiti Lahaul is the only site wheresurface lowering is measured both in the accumula-tion area (minus034plusmn 016 m yrminus1) and the ablation area(minus070plusmn 014 m yrminus1) Clear surface lowering is also ob-served in the ablation areas of the Hindu Kush glaciers(minus050plusmn 018 m yrminus1)

In the Karakoram (east and west Figs 8 and 9) andin the Pamir (Fig 10) the pattern of elevation changesis highly heterogeneous due to the occurrence of numer-ous glacier surges or similar flow instabilities The eleva-tion changes in the accumulation areas of those three studysites are slightly positive (between+022plusmn 014 m yrminus1 and+030plusmn 018 m yrminus1) In the ablation areas small surfacelowering is observed for the two Karakoram study sites (av-erage of the two sitesminus033plusmn 016 m yrminus1) whereas no el-evation change is found in Pamir (minus002plusmn 013 m yrminus1)

Note that those mean rates of elevation changes are sensi-tive to the choice of the ELA assumed to equal the snowlinealtitude in Landsat images from around year 2000

42 Influence of a debris cover

To evaluate thinning rates over debris-covered and debris-free ice we perform a histogram adjustment in order tocompare data sets with similar altitude distributions This isneeded because debris-covered parts tend to be concentratedat lower elevations The adjustment consists in randomly ex-cluding pixels over debris-free ice from each elevation bandto match a scaled version of the debris-covered pixel distribu-tion (Kaumlaumlb et al 2012) Elevation changes with altitude canthen be compared for clean and debris-covered ice (Fig 11)

The thinning is higher for debris-covered ice in the Ever-est area and on the lowermost parts of glaciers in the PamirBut on average in the Pamir western Karakoram and SpitiLahaul the lowering of debris-free and debris-covered ice issimilar The thinning is stronger over clean ice in the Heng-duan Shan Bhutan West Nepal Hindu Kush and althoughto a lesser extent in the eastern Karakoram

43 Mass balance over the nine study sites

Glacier mass changes have been heterogeneous overthe PKH since 1999 Slight mass gains or balancedmass budgets are found in the north-west in thePamir (+014plusmn 013 m we yrminus1) and for the two studysites in the central Karakoram (+009plusmn 018 m we yrminus1

and +011plusmn 014 m we yrminus1) Mass loss is moderatein the adjacent Hindu Kush with a mass balance ofminus012plusmn 016 m we yrminus1 Glaciers in all our eastern studysites (two in Nepal Bhutan and Hengduan Shan) experi-enced homogeneous mass losses with mass balances ranging

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1271

Fig 2 Glacier elevation changes over the Hengduan Shan study site Inset distribution of the elevation differences off glacier The medianand the standard deviation of the elevation difference and the number of pixels off glaciers are given (by construction the mean elevationdifference is null)

Fig 3Same as Fig 2 but for the Bhutan study site

from minus033plusmn 014 m we yrminus1 to minus022plusmn 012 m we yrminus1The most negative mass balance is measured in Spiti Lahaul(western Himalaya) atminus045plusmn 013 m we yrminus1 (Table 4)

However within a study site mass balances can be highlyvariable from one glacier to another In the Everest area twolarge glaciers (71 km2 each) experienced distinctly differentmass balance withminus016plusmn 016 m we yrminus1 for the south-flowing Ngozumpa Glacier andminus059plusmn 016 m we yrminus1

for the north-flowing Rongbuk Glacier (Fig 4) In Bhutan(Fig 3) mass loss is higher south (minus025plusmn 013 m we yrminus1)

than north (minus014plusmn 012 m we yrminus1) of the main Hi-malayan ridge though not at a statistically significant level

The region-wide mass budget of PKH glaciers isnegative minus101plusmn 55 Gt yrminus1 (minus014plusmn 008 m we yrminus1)which corresponds to a sea level rise contribution of0028plusmn 0015 mm yrminus1 over 1999ndash2011

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1272 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 4Same as Fig 2 but for the Everest study site

Fig 5Same as Fig 2 but for the West Nepal study site

44 Glacier surges

In the Pamir and the Karakoram the spatial pattern of ele-vation changes of many glaciers is typical of surge events orflow instabilities They can be identified because of their veryhigh thinning and thickening rates that can both reach up to16 m yrminus1 (Figs 8ndash10)

Most of them have already been reported to be surge-type glaciers based on their velocity (Kotlyakov et al 2008Quincey et al 2011 Copland et al 2009 Heid and Kaumlaumlb2012) morphology (Copland et al 2011 Barrand and Mur-ray 2006 Hewitt 2007) or elevation changes (Gardelle etal 2012a) Here we only identify surge-type glaciers for thesake of mass balance calculation and we do not attempt toprovide an exhaustive up-to-date surge inventory This is why

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1273

Fig 6 Same as Fig 2 but for the Spiti Lahaul study site The map of elevation changes contains many voids due to the presence ofthin clouds during the acquisition of the SPOT5 DEM The mass balance of the Chhota Shigri and Bara Shigri glaciers are respectivelyminus039plusmn 018 m we yrminus1 andminus048plusmn 018 m we yrminus1

the published surge inventories in the Pamir and the Karako-ram do not necessarily match the surge-type glaciers labeledon the elevation change maps as these refer only to our ob-servation period (1999ndash2011)

The mass balance of surge-type glaciers (respectivelynon-surge-type glaciers) is +019plusmn 022 m we yrminus1

(+012plusmn 011 m we yrminus1) in the Pamir+009plusmn 030 m we yrminus1 (+009plusmn 015 m we yrminus1) inthe western Karakoram and+010plusmn 025 m we yrminus1

(+012plusmn 012 m we yrminus1) in the eastern Karakoram Theoverall similarity between the mass budgets for surge-typeand non-surge-type glaciers shows that those flow instabili-ties seem not to affect the glacier-wide mass balance at leastover short time periods here 10 yr

5 Discussion

51 SRTM penetration

Alternative estimates of SRTMC-bandpenetration for the PKHregion were obtained by Kaumlaumlb et al (2012) by comparing el-evations of the ICESat altimeter (Zwally et al 2002) with theSRTMC-bandDEM in PKH and extrapolating the 2003ndash2008trend of elevation changes back to the acquisition date ofSRTM The difference between east and west is more dis-tinct in our case with a mean penetration in the Karakoram of34 m (24plusmn 03 m in Kaumlaumlb et al 2012) 24 m in Bhutan and14 m around the Everest area (25plusmn 05 m for a wider areaincluding east Nepal and Bhutan in Kaumlaumlb et al 2012) Thiseastwest gradient is expected given that in February 2000when the SRTM mission was flown the snowfirn thicknesswas probably larger in the western PKH where glaciers areof winter-accumulation type than in the eastern PKH where

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1274 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 7Same as Fig 2 but for the Hindu Kush study site

glaciers are of summer-accumulation type Further studieswould be necessary to investigate the possible reasons be-hind the relatively low SRTMC-bandpenetration for the Pamir(18 m) that do not follow the eastwest gradient mentionedabove

Both our and Kaumlaumlb et alrsquos (2012) estimates agree withintheir error bars and seem thus to provide robust first-orderestimates for the SRTMC-bandradar penetration in the regionHowever it remains to be verified if a penetration depth com-puted at the regional scale is appropriate for a single smallglacier whose altitude distribution is different from the oneof whole study site (eg Abramov Glacier in the far north ofour Pamir study site)

52 Thinning over debris-covered ice

PKH glaciers are heavily debris-covered in their ablation ar-eas because of steep rock walls that surround them and in-tense avalanche activity Twelve percent (12 ) of the glacierarea in our study sites are covered with debris (up to 22 inthe Everest area Table 1) which is in agreement with pre-vious estimates for the Himalaya and the Karakoram 13 in Kaumlaumlb et al (2012) and 10 in Bolch et al (2012) Thedebris layer is expected to modify the ablation of the under-

lying ice It will increase ablation if its thickness is just a fewcentimeters (dominance of the albedo decrease) or decreaseablation it if the layer is thick enough to protect the ice fromsolar radiation (Mattson et al 1993 Mihalcea et al 2006Benn et al 2012 Lejeune et al 2013)

However recent studies have reported similar overall thin-ning rates over debris-covered and debris-free ice in thePKH (Kaumlaumlb et al 2012 Gardelle et al 2012a) or evenhigher lowering rates over debris-covered parts in the Ever-est area (Nuimura et al 2012) Our findings are consistentwith Nuimura et al (2012) for the Everest study site with astronger thinning (rate of elevation change ofminus097 m yrminus1)

in debris-covered areas than over clean ice (rate of elevationchange ofminus057 m yrminus1) In the Pamir Karakoram and SpitiLahaul study sites these rates are similar while in the HinduKush West Nepal Bhutan and Hengduan Shan study sitesthinning is greater over clean ice in agreement with the com-monly assumed protective effect of debris (Fig 4) Thosedifferences between our 9 study sites suggest a complex re-lationship between debris cover and glacier wastage

As local glacier thickness changes are a result of bothmass balance processes and a dynamic component theseunexpected high thinning rates over debris-covered parts

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1275

Fig 8Same as Fig 2 but for the Karakoram east study site Surge-type glaciers in an active surge phase (resp quiescent phase) are identifiedwith a solid triangle (resp solid circle)

are potentially the result of glacier-wide processes Over aglacier tongue the ablation can be enhanced by the pres-ence of steep exposed ice cliffs or supraglacial lakes andponds (Sakai et al 2000 2002) In addition it is also pos-sible that the glacier tongues are mostly covered by thindebris layers such that the albedo effect dominates the in-sulating effect resulting in an increased tongue-wide abla-tion The prevalence of thin debris was recently suggestedfor debris-covered glaciers around the Mount Gongga in thesouth-eastern Tibetan Plateau (Zhang et al 2013b) Finallyice-flow rates could be very low at debris-covered parts asshown by Quincey et al (2007) in the Everest area Kaumlaumlb(2005) for Bhutan and Scherler et al (2011b) in the HinduKush-Karakoram-Himalaya Thus all three factors are likelyto increase the overall thinning under debris despite the un-doubted protective effect of a continuous thick debris coverat a local scale Future investigations combining surface ve-locity fields and maps of elevation changes could allow eval-uating more closely the relative role of ablation and ice fluxesin thinning rates over PKH glacier tongues (eg Berthierand Vincent 2012 Nuth et al 2012) Measurements of the

spatial distribution of debris thickness over the PKH glaciertongues are also needed

Note that in the case of debris-covered tongues the ele-vation change measurement represents the glacier thicknesschange but also the possible debris thickness evolution Thelatter remains poorly known in the PKH or in any othermountain range But based on the debris discharges givenby Bishop et al (1995) for Batura Glacier (Pakistan) 48to 90times 103 m3 yrminus1 the thickening of the debris can be es-timated between 27 and 55 mm yrminus1 which is negligiblegiven the magnitude of the glacier elevation changes

53 Comparison to previous mass balance estimates

Throughout the PKH there is only a single peer-reviewedglaciological record (on Chhota Shigri Glacier) coveringmost of our study period (Wagnon et al 2007 Vincent etal 2013) Our geodetic mass balance estimate for the SpitiLahaul study site has been recently compared to those glacio-logical measurements on Chhota Shigri Glacier (Vincent etal 2013) Mass balance records reported in Yao et al (2012)for the Tibetan Plateau and the surroundings mountain ranges

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1276 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 9Same as Fig 8 but for the Karakoram west study site

only start during glaciological year 20052006 This is whyhere we do not compare our mass balance estimates toglaciological records and limit our comparison to others re-gional estimates obtained using the geodetic or gravimetricmethods

We stress that the comparison of our mass balance mea-surements to published estimates is complicated by thefact that generally they do not cover the same time pe-riod Little is known about the inter-annual variability inthe PKH during the 21st century due to the limited num-ber of long glaciological time series The 9 yr mass balance

record on Chhota Shigri Glacier reveals a relatively highinter-annual variability of the annual mass balance (stan-dard deviationplusmn 074 m we yrminus1 during 2003ndash2011 Vin-cent et al 2013) which is similar to the one obtained be-tween 1968 and 1998 on Abramov Glacier (standard devi-ation 069 m we yrminus1 WGMS 2012) Thus inter-annualvariability can explain part of the difference between two cu-mulative mass balance estimates over 5ndash10 yr even if they areoffset by only one or two years

Based on ICESat repeat track measurements and theSRTM DEM Kaumlaumlb et al (2012) measured a region-wide

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1277

Fig 10Same as Fig 8 but for the Pamir study site

mass balance ofminus021plusmn 005 m we yrminus1 between 2003and 2008 for Hindu Kush-Karakoram-Himalaya glaciersFor that same area (ie excluding the Pamir and theHengduan Shan study sites) we found a mass balance ofminus015plusmn 007 m we yrminus1 Thus our result agrees with Kaumlaumlbet al (2012) within the error bars although our study periodis longer (1999ndash2011) and our measurement principle andextrapolation approach are different In addition we havecalculated updated mass balances from Kaumlaumlb et al (2012)ie averages over 3times 3 degree cells centered over our studysites and find that they are consistent within error barswith our mass balance estimates (Table 5) Similar resultsare found when our mass balances are compared to a spa-tially extended ICESat analysis for 2003ndash2009 (Gardner etal 2013) Mass losses measured with ICESat (Kaumlaumlb et al2012 Gardner et al 2013) tend to be slightly larger thanours This could be due to the difference in time periodsAlso our smaller mass losses could be partly explained if wesystematically underestimated the poorly constrained pene-

tration of the C-band andor X-band radar signal into ice andsnow

At a more local scale Bolch et al (2011) reported a massbalance ofminus079plusmn 052 m yrminus1 between 2002 and 2007 byDEM differencing over a glacier area of 62 km2 includingten glaciers south and west of Mt Everest For these sameglaciers Nuimura et al (2012) found a mean mass balanceof minus045plusmn 060 m yrminus1 during 2000ndash2008 while they com-puted a mass balance ofminus040plusmn 025 m yrminus1 between 1992and 2008 over a glacier area of 183 km2 around Mt EverestIn the present study we found an average mass balance ofminus041plusmn 021 m yrminus1 over the ten glaciers mentioned above(Table 5 Fig S1) and a value ofminus026plusmn 013 m yrminus1 overthe whole Everest study site between 2000 and 2011 Ourvalues are comparable to those of Nuimura et al (2012)Our mass balance is less negative than the 2002ndash2007 massbalance of Bolch et al (2011) but not statistically differentwhen error bars are considered (Fig 12) Indeed Bolch etal (2011) acknowledged some high uncertainties for their

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1278 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Table 4Glacier mass budget of the nine sub-regions to which the results of the study sites (mass balances) have been extrapolated For eachof them the total glacierized area as well as the percentage of the glacier area over which the mass balance was computed (ie study sites)are given The total glacier area is derived from the RGI v20 (Arendt et al 2012) The error for the mass balance in each sub-region is theroot of sum of squares (RSS) of the mass balance error of each study site and the regional extrapolation error The error for the mass budgetin each sub-region includes additionally the error from the total glacier area in the RGI v20

Sub-region Total glacier Measured glacier Mass balance Mass budgetarea (km2) area () (m we yrminus1) (Gt yrminus1)

Hengduan Shan 11584 12 minus033plusmn 014 minus38plusmn 38Bhutan 4021 34 minus022plusmn 013 minus09plusmn 05Everest 6226 23 minus026plusmn 014 minus16plusmn 08West Nepal 6849 13 minus032plusmn 014 minus22plusmn 10Spiti Lahaul 9043 23 minus045plusmn 014 minus41plusmn 13Hindu Kush 6135 13 minus012plusmn 016 minus07plusmn 10Karakoram East

1902428 +011 plusmn 014 +19 plusmn 31

Karakoram West 30 +009 plusmn 018Pamir 9369 34 +014 plusmn 014 +13 plusmn 13

TotalArea- 72251 30 minus014plusmn 008 minus101plusmn 55weighted average

Table 5Comparison of geodetic mass balances between this study and previous published results with overlapping study periods and similargeographic locations Updated figures from Kaumlaumlb et al (2012) are averaged over 3 times 3 cells centered over the study sites of the presentstudy

GlacierSite name This study Bolch et al (2011) Nuimura et al (2012) Kaumlaumlb et al (2012 updated)2002ndash2007 2000ndash2008 2003ndash2008

Study Mass balance Area (km2)a Mass balance Area Mass balance Mass balanceperiod (m we yrminus1) (m we yrminus1) (km2) (m we yrminus1) (m we yrminus1)

Hengduan San 1999ndash2010 minus022 plusmn 014 1303 (55 )Bhutan 1999ndash2010 minus022 plusmn 012 1384 (64 ) minus052 plusmn 016b

Everest

1999ndash2011

minus026 plusmn 013 1461 (58 ) minus039 plusmn 011AX010 minus090plusmn 034 04Changri SharNup minus042plusmn 017 161 (79 ) minus029plusmn 052 130 minus055plusmn 038Khumbu minus051plusmn 019 203 (47 ) minus045plusmn 052 170 minus076plusmn 052Nuptse minus037plusmn 020 50 (74 ) minus040plusmn 053 40 minus034plusmn 027Lhotse Nup minus021plusmn 027 24 (70 ) minus103plusmn 051 19 minus022plusmn 047Lhotse minus043plusmn 018 85 (83 ) minus110plusmn 052 65 minus067plusmn 051Lhotse SharImja minus070plusmn 052 98 (44 ) minus145plusmn 052 107 minus093plusmn 060Amphu Laptse minus046plusmn 034 25 (38 ) minus077plusmn 052 15 minus018plusmn 094Chukhung +044 plusmn 024 42 (47 ) +004 plusmn 054 38 +043 plusmn 081Ama Dablam minus049plusmn 017 36 (54 ) minus056plusmn 052 22 minus056plusmn 073Duwo minus016plusmn 026 19 (18 ) minus196plusmn 053 10 minus068plusmn 074Total Khumbu minus041plusmn 021 744 minus079plusmn 052 617 minus045plusmn 060(10 Glaciers above)

West Nepal 1999ndash2011 minus032 plusmn 013 908 (40 ) minus032 plusmn 012Spiti Lahaul 1999ndash2011 minus045 plusmn 013 2110 (46 ) minus038 plusmn 006Karakoram East 1999ndash2010 +011 plusmn 014 5328 (42 ) minus004 plusmn 004Karakoram West 1999ndash2008 +009 plusmn 018 5434 (45 )Hindu Kush 1999ndash2008 minus012 plusmn 016 793 (80 ) minus020 plusmn 006Pamir 1999ndash2011 +014 plusmn 013 3178 (50 )

a The total glacier area is given in km2 and in parenthesis the of the glacier area actually covered with measurementsb For this cell the ICESat coverage is insufficient and does not sample all glacier elevations

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1279

Fig 11 Elevation changes (SPOT5-SRTM) in the ablation area for clean ice (blue circles) and debris-covered ice (black circles) for eachstudy site For the Karakoram (east and west) and the Pamir study sites only non-surge-type glaciers are considered Standard error of theelevation differences are typically less thanplusmn2 m with each 100 m elevation band (not shown for the sake of clarity) Note elevation changesover separate sections of a glacier cannot be treated as mass changes due to the disregard of glacier dynamics

estimate over this short time period The differences in sur-vey periods and in the glacier outlines may also explainpart of these discrepancies In particular the delimitation ofthe accumulation area is notoriously difficult and Bolch etal (2011) did not include some of the steep high altitudeslopes which we included due to their potential avalanchecontribution For the 10 glaciers mentioned above our massbalances would become 012 m we yrminus1 more negative ifwe used the exact same outlines as Bolch et al (2011)Our positive mass balance for the small Chukhung Glacier(+044plusmn 024 m we yrminus1) is in agreement with Nuimura etal (2012) but enigmatic in a context of regional glacier lossand deserves further investigation If the 2002ndash2007 massbalance estimate by Bolch et al (2011) is excluded all otherrecent mass balance estimates suggest no acceleration of themass loss in the Everest area when compared to the long-term(1970ndash2007) mass balance measured by Bolch et al (2011)in this regionminus032plusmn 008 m weyrminus1

In Bhutan our results indicate a stronger mass loss forsouthern glaciers than for northern ones Interestingly in thisarea Kaumlaumlb (2005) measured high speeds (up to 200 m yrminus1)

for glaciers north of the Himalayan main ridge and ratherlow velocities over southern glacier tongues by using repeat

ASTER data This is consistent with the more negative massbalance that we report for the southern glaciers in Bhutanie for the slow flowing presumably downwasting glaciers

Berthier et al (2007) reported a mass balance betweenminus069 andminus085 m we yrminus1 for an ice-covered area of915 km2 around Chhota Shigri Glacier between 1999 (SRTMDEM) and 2004 (SPOT5-HRG DEM) For the same areawe found a mass balance ofminus045plusmn 013 m we yrminus1 over1999ndash2011 The difference in the study period may explainpart of the differences Also the methodology by Berthieret al (2007) did not explicitly take into account the SRTMradar penetration over glaciers and also corrected an appar-ent elevation-dependent bias according to glacier elevationswhich is now known as a resolution issue and is best cor-rected according to the terrain maximum curvature (Gardelleet al 2012b) More details and discussions about these dif-ferences can be found in Vincent et al (2013)

Importantly the results of the eastern Karakoram studysite confirm the balanced or slightly positive mass budgetreported by Gardelle et al (2012a) in the western Karako-ram over the past decade The two Karakoram study sitesare adjacent with a small overlapping area (sim 1100 km2 ofglaciers) where elevation differences agree within 1 m Both

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1280 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 12 Comparison of geodetic mass balances for 10 glaciers inthe Mount Everest area 1999ndash2010 mass balances measured inthe present study are compared with published estimates during2002ndash2007 (Bolch et al 2011) and during 2000ndash2008 (Nuimuraet al 2012) The 1-to-1 line is drawn The mean difference (plusmn1standard deviation) between (i) our values and Bolch et al (2011)are (047plusmn 058 m we yrminus1) and (ii) our values and Nuimura etal (2012) are (012plusmn 021 m we yrminus1) Part of the differences be-tween estimates may be due to the different periods surveyed andthe different glacier outlines used

sites show similar mass balances over slightly different pe-riods +011plusmn 014 m yrminus1 over 1999ndash2010 in the east and+009plusmn 018 m yrminus1 over 1999ndash2008 in the west This con-firms a ldquoKarakoram anomalyrdquo that was first suggested basedon field observations (Hewitt 2005) We report a mass bal-ance of+010plusmn 013 m yrminus1 for Baltoro Glacier (see loca-tion in Fig 8 size= 304 km2) between 1999 and 2010which is consistent with its gradual speed-up reported since2003 (Quincey et al 2009) However given the completelydifferent methods used in our and the latter study we can-not conclude that this demonstrates a real shift from nega-tive to positive mass balance The mass budget of SiachenGlacier (1145 km2 sometimes referred to as the highestbattlefield in the world) is also balanced or slightly pos-itive (+014plusmn 014 m we yrminus1) Thus the alarming state-ment by Rasul (2012) that this glacier retreated by 59 kmand lost 17 of its mass during 1989ndash2009 is very likelyerroneous Again within error bars our regional mass bal-ance in the Karakoram agrees with the mass balance ofminus003plusmn 004 m yrminus1 found by Kaumlaumlb et al (2012) over 2003ndash2008 and with the mass balance ofminus010plusmn 018 m we yrminus1

(2-sigma error level) obtained by Gardner et al (2013) for aregion including both the Karakoram and the Hindu Kush

In the northernmost part of the Pamir in the Alay moun-tain range the mass balance of Abramov Glacier has beenmeasured in the field for 30 yr between 1968 and 1998(WGMS 2012) with a mean value ofminus046 m yrminus1 Herewe report a mass balance ofminus003plusmn 014 m yrminus1 for this

glacier over 1999ndash2011 Our near zero mass budget estimatefor this glacier disagrees with negative mass balances recon-structed for the same period with a model run using biascorrected NCEPNCAR re-analysis data and calibrated withfield data (Barundun et al 2013) An underestimate of ourSRTM C-Band penetration in the Pamir as a whole and inparticular for the small Abramov Glacier may contribute tothese differences

For the whole Pamir study site our mass budget is bal-anced or slightly positive (+014plusmn 013 m we yrminus1) In theeastern Pamir the mass balance of Muztag Ata Glacier was+025 m we yrminus1 between 2005 and 2010 (Yao et al 2012)Although Muztag Ata Glacier is located outside of our Pamirstudy site and his glaciological records only cover the sec-ond half of our study period this measurement is consistentwith our remotely sensed mass balance Thus the ldquoKarako-ram anomalyrdquo should probably be renamed the ldquoPamir-Karakoram anomalyrdquo

This slightly positive or balanced mass budget measuredin the Pamir seems to contradict previous findings by Heidand Kaumlaumlb (2012) They reported rapidly decreasing glacierspeeds in this region between two snapshots of annual ve-locities in 20002001 and 20092010 an observation thatwould fit with negative mass balances (Span and Kuhn 2003Azam et al 2012) We note though that the relationship be-tween glacier mass balance and ice fluxes is not straightfor-ward and might in particular involve a time delay betweena change in mass balance and the related dynamic response(Heid and Kaumlaumlb 2012) Thus the decrease in speed couldwell be due to negative mass balances before or partiallybefore 1999ndash2011 We acknowledge that this discrepancyneeds to be investigated further in particular by examiningcontinuous changes in annual glacier speed through the firstdecade of the 21st Century Indeed Heid and Kaumlaumlb (2012)measured differences between two annual periods which maynot represent mean decadal velocity changes

Jacob et al (2012) using satellite gravimetry (GRACE)measured a mass budget ofminus5plusmn 6 Gt yrminus1 over 2003ndash2010for the ldquoKarakoram-Himalayardquo their region 8c which cor-responds to our study area excluding the Pamir Karakoramand Hindu Kush sub-regions but including some glaciersnorth of the Himalayan ridge already on the Tibetan PlateauFor this subset of our PKH study area we measured amass budget ofminus126plusmn 42 Gt yrminus1 The two estimates over-lap within their error bars especially if our error is dou-bled to match the two standard error level used by Jacob etal (2012) The estimate of Jacob et al (2012) over our com-plete study region (PKH) ieminus6plusmn 8 Gt yrminus1 (sum of theirregions 8b and 8c) also includes mass changes for the KunlunShan sub-region for which we did not measure glacier massbalances Therefore the comparison with our PKH value(minus101plusmn 55 Gt yrminus1) is not straightforward but suggests areasonable agreement especially if we consider the slightmass gain (15plusmn 17 Gt yrminus1) measured with ICESat in theWest Kunlun (Gardner et al 2013)

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1281

Table 6 Annual average of the glacier seasonal and decadal mass loss contributions to Indus Ganges and Brahmaputra discharges Basinarea and seasonal contributions are given by Kaser et al (2010) glacier area in each basin is derived from the RGI v20 (Arendt et al 2012)Numbers in parentheses indicate the percentage of glacierized area within the river basin Seasonal contributions were modeled by Kaser etal (2010) based on the CRU 20 dataset over 1961ndash1990 assuming a balanced glacier mass budget Glacier decadal mass loss contributionsare given as annual average over 1999ndash2011

River Basin area Glacier area Glacier contribution Mean discharge(km2) (km2) (m3 sminus1) (m3 sminus1)

Seasonally delayed Decadal mass lossKaser et al (2010) (this study)

Indus 1 139 814 25 598 (2 ) 105 103plusmn 72 2146 (Chevallier personalcommunication 2012)

Ganges 1 023 609 11 168 (1 ) 47 103plusmn 49 11739 Papa et al (2012Brahmaputra 527 666 15 296 (3 ) 33 147plusmn 64 19710 updated)

54 Reasons for the heterogeneous pattern of massbalance

The PKH glacier mass balance is slightly negative(minus014plusmn 008 m we yrminus1) but changes are not homoge-neous from one sub-region to another and seem to coincidewith the climatic settings of the mountain range For the east-ern sites (from the Hengduan Shan to West Nepal) which areunder the influence of the Indian and south-east Asian sum-mer monsoons the mass balances are moderately negative(sim minus026 m we yrminus1) In the northwest of the PKH at thePamir and Karakoram sites where glacier climate is char-acterized by the westerlies in winter mass balances are bal-anced or slightly positive (sim +010 m we yrminus1) In betweenthese two influences the glaciers of the Spiti Lahaul sitein a monsoon-arid transition zone experienced the strongestmass loss (minus045plusmn 013 m we yrminus1)

This heterogeneous pattern of mass balances seems to berelated to trends in precipitation throughout the PKH Archerand Fowler (2004) reported an increase in winter precipi-tation over the Upper Indus Basin since 1961 a trend con-firmed by Yao et al (2012) over the Pamir and the Karako-ram based on the analysis of the Global Precipitation Cli-matology Project data set (GPCP Adler et al 2003) Thisprecipitation increase is a source of greater accumulation forglaciers and thus a possible explanation for their slightly pos-itive mass budgets In addition in the Upper Indus Basin themean summer temperature has been decreasing since 1961(Fowler and Archer 2006) which is consistent with the find-ings of Shekhar et al (2010) in the Karakoram who reporteda decrease in both maximum and minimum temperature since1988 These increases in winter precipitation and summertemperature likely translate in greater accumulation and re-duced ablation changes which are consistent with the bal-anced or slightly positive glacier mass budget

On the other hand in the east of the PKH the Indiansummer monsoon has been weakening since the 1950s (Bol-lasina et al 2011) which could have reduced the amount

of snowfall over the eastern study sites and thus resultedin negative mass balances In addition maximum and min-imum temperatures increased since 1988 in the western Hi-malaya (Shekhar et al 2010) and maximum temperaturesincreased in the central Himalaya (Nepal) since the mid-1970s (Shrestha et al 1999) Thus both precipitation andtemperature trends in the Himalaya likely contributed to thenegative mass balances measured over the correspondingstudy sites (from Spiti Lahaul to Hengduan Shan Fig 1)

However gridded data sets such as GPCP are not neces-sarily representative of the climate at high altitude IndeedHewitt (2005) reported a 5-to-10-fold increase in precipi-tation between the glacier front and the accumulation areain the Karakoram that is not captured by reanalysis dataas recently confirmed by Immerzeel et al (2012) Thus di-rect measurements of accumulation on High Mountain Asiaglaciers (eg Azam et al 2013) and high-altitude weatherstations are eagerly needed to validate the large scale grid-ded data set (eg reanalysis) test their ability to describe thespecific climate near to PKH glaciers and assess the relativerole played by temperature and precipitation changes

55 Contribution to water resources

Our glacier mass balance can be averaged over large riverbasins to estimate the mean contribution to river runoffdue to a decade of glacier mass loss We assume therebyonly direct river runoff without loss terms For the threerivers for which we cover the entire glacierized area of theircatchments within our sub-regions (Fig 1) these contribu-tions to the river discharge are equal to 103 m3 sminus1 (Indus)103 m3 sminus1 (Ganges) and 147 m3 sminus1 (Brahmaputra) for theperiod 1999ndash2011 (Table 6)

Rivers of PKH being key water resources measures oftheir discharges are highly sensitive and difficult to access forpolitical reasons Papa et al (2012) built recent time series ofdischarges for Ganges and Brahmaputra by using altimetrymeasurements The locations of the corresponding dischargeestimates in Bangladesh are given in Fig 1 We can therefore

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1282 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

evaluate the relative contribution of glacier decadal mass loss(Table 6) to annual river run-off at 09 for the Gangesat Hardinge (basin area ofsim 1 020 000 km2) and 07 forBrahmaputra at Bahadurabad (basin area ofsim 530 000 km2)between 1999 and 2011 Given the discharge measured at theGuddu station by the Surface Water Hydrology Project of theWater and Power Development Authority (Pakistan see lo-cation on Fig 1) between 1999 and 2003 we can estimatethe contribution of glacier mass loss at 48 for the IndusRiver at this station

The seasonal contributions of glaciers to river run-off (iethe part of the annual precipitation that experiences season-ally delayed release from the glaciers) directly results fromseasonal variations of meteorological quantities in contrastto the glacier mass loss contribution which results from thelong-term climatic change Both runoff components directlyaffect the amount of water available in a river at a given lo-cation and at a given time so that their comparison could beof interest Even if the seasonal timing of the glacier massloss contribution is unclear it could in general peak at asimilar time of the year than the seasonal glacier contribu-tion so that both components rather enlarge each other thancancel Seasonal glacier contributions have been modeled byKaser et al (2010) The latter study assumed glaciers to bein equilibrium with the present climate and thus prescribedglacier mass balances equal to 0 Thus their seasonally de-layed glacier contributions do not include the part due todecadal glacier mass loss and is complementary to our es-timates (Table 6) For the eastern basins under the influenceof the Indian summer monsoon (Ganges and Brahmaputra)the contribution of glacier mass loss is higher than the sea-sonal contribution In other words for the first decade of the21st century the mass loss of glaciers is on average moreimportant for water availability in those two rivers than theirseasonal water storage effect The situation is different for theIndus Basin where the glacier melt season is also the dry sea-son which results in a relatively higher seasonally delayedglacier contribution to the river discharge There the season-ally delayed water release from the glaciers and the decadalglacier mass loss contribute on average equally to the riverdischarge Both contributions are strongly (and equally) de-pendent on the location of the discharge measurement andwill be significantly larger (in relative term) in more heav-ily glacierized and smaller mountain catchments located inthe upper part of these major river basins (Bookhagen andBurbank 2010)

6 Conclusion

In this study we assessed the spatial pattern of glaciermass balance over the PKH by comparing the SRTM andSPOT5 DEMs over 9 well-distributed study sites between1999 and 2011 We found slightly positive or balanced massbudgets in the western part (Pamir Karakoram) moder-ate mass loss in the east (Nepal Bhutan Hengduan Shan)

and the most negative mass balance in the monsoon-aridtransition zone (Spiti Lahaul in the western Himalaya)By extrapolating these values to climatically homogeneoussub-regions we estimate the PKH overall mass balanceto have beenminus014plusmn 008 m we yrminus1 between 1999 and2011 which corresponds to a contribution to sea level riseof 0028plusmn 0015 mm yrminus1 This is about 4 of the globalglacier mass loss estimated by Gardner et al (2013) thoughover a slightly shorter time period (2003ndash2009) The largestsource of uncertainties in those geodetic mass balance esti-mates is the poorly constrained penetration of the C-BandSRTM radar signal into snow and ice

Our technique of DEM differencing allows mapping in de-tail the spatial pattern of glacier elevation changes Thosemaps reveal many surge-type behaviors both in the Pamirand the Karakoram It also appears that the glacier-wide massbalance does not seem to be considerably affected by themass transfer from surges over thesim 10 yr time period stud-ied Similar lowering rates over debris-covered and clean iceare observed for four study sites despite the commonly as-sumed insulating effect of debris covers This suggests forthe scale of entire glacier tongues a spatial mixture of re-duced ablation under intact thick debris covers and enhancedablation by supraglacial lakes or ice cliffs or due to thin de-bris layer and very low glacier velocities in ablation areasthat reduce the ice advection downstream

The regionally heterogeneous pattern of ice wastage is inbroad agreement with contrasted trends in large-scale pre-cipitation and temperature However glaciological (eg ac-cumulation) and meteorological records are needed at highaltitude in the vicinity of glaciers to evaluate more closelythe local influence of atmospheric variables on glacier massbalance and fully understand the reasons behind those con-trasted mass budgets in the PKH

Supplementary material related to this article isavailable online at httpwwwthe-cryospherenet712632013tc-7-1263-2013-supplementpdf

Acknowledgements We thank G Cogley A Gardner T NuimuraT Bolch P Wagnon M Barandun M Hoelzle M Huss and ananonymous referee for their constructive comments that led to aconsiderably improved manuscript We thank the Water and PowerDevelopment Authority (WAPDA) for their hydrological data(processed by P Chevallier) and F Papa for sharing his updatedrun-off data ahead of publication We thank T Bolch for sharinghis glacier outlines from the Everest area We thank the USGSfor allowing free access to their Landsat archive CGIAR-SCI forSRTM C-band data DLR for SRTM X-band data and GLIMS forglacier inventories J Gardelle acknowledges a PhD fellowshipfrom the French Space Agency (CNES) and the French NationalResearch Center (CNRS) E Berthier acknowledges support fromCNES through the TOSCA and ISIS proposal no 397 and fromthe Programme National de Teacuteleacutedeacutetection Spatiale E Berthieracknowledges M Masiokas and R Villalba for welcoming him

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1283

at the IANIGLA (Mendoza Argentina) Y Arnaud acknowledgessupport from French National Research Agency through ANR-09-CEP-005-01PAPRIKA A Kaumlaumlb acknowledges support fromESA through Glaciers_cci (400010177810I-AM) and from theEuropean Research Council under the European Unionrsquos SeventhFramework Programme (FP2007-2013)ERC Grant Agreementn 320816

Edited by I M Howat

The publication of this articleis financed by CNRS-INSU

References

Adler R Huffman G Chang A Ferraro R Xie P JanowiakJ Rudolf B Schneider U Curtis S Bolvin D Gruber ASusskind J Arkin P and Nelkin E The Version-2 GlobalPrecipitation Climatology Project (GPCP) Monthly Precipita-tion Analysis (1979ndashPresent) J Hydrometeorol 4 1147ndash11672003

Archer D R and Fowler H J Spatial and temporal variationsin precipitation in the Upper Indus Basin global teleconnectionsand hydrological implications Hydrol Earth Syst Sci 8 47ndash61doi105194hess-8-47-2004 2004

Arendt A Bolch T Cogley J G Gardner A Hagen J-OHock R Kaser G Pfeffer W T Moholdt G Paul F RadicV Andreassen L Bajracharya S Beedle M Berthier EBhambri R Bliss A Brown I Burgess E Burgess DCawkwell F Chinn T Copland L Davies B de AngelisH Dolgova E Filbert K Forester R Fountain A Frey HGiffen B Glasser N Gurney S Hagg W Hall D Hari-tashya U K Hartmann G Helm C Herreid S Howat IKapustin G Khromova T Kienholz C Koenig M KohlerJ Kriegel D Kutuzov S Lavrentiev I LeBris R Lund JManley W Mayer C Miles E Li X Menounos B Mer-cer A Moelg N Mool P Nosenko G Negrete A NuthC Pettersson R Racoviteanu A Ranzi R Rastner P RauF Rich J Rott H Schneider C Seliverstov Y Sharp MSigurethsson O Stokes C Wheate R Winsvold S WolkenG Wyatt F and Zheltyhina N Randolph Glacier Inventory[v20] A Dataset of Global Glacier Outlines Global Land IceMeasurements from Space Boulder Colorado USA Digital Me-dia httpwwwglimsorgRGIrandolphhtml 2012

Azam M Wagnon P Ramanathan A Vincent C Sharma PArnaud Y Linda A Pottakkal J Chevallier P Singh V andBerthier E From balance to imbalance a shift in the dynamicbehaviour of Chhota Shigri glacier western Himalaya India JGlaciol 58 315ndash324 doi1031892012JoG11J123 2012

Azam M F Wagnon P Vincent C Ramanathan A Linda Aand Singh V B Reconstruction of the annual mass balanceof Chhota Shigri Glacier (Western Himalaya India) since 1969Ann Glaciol in revision 2013

Barrand N and Murray T Multivariate Controls on the In-cidence of Glacier Surging in the Karakoram Himalaya

Arct Antarct Alp Res 38 489ndash498 doi1016571523-0430(2006)38[489MCOTIO]20CO2 2006

Barundun M Huss M Azisov E Gafurov A HoelzleM Merkushkin A Salzmann N and Usubaliev R Re-establishing seasonal mass balance observation at AbramovGlacier Kyrgyzstan from 1968ndash2012 Abstract EGU2013-5668European Geophysical Union Vienna 2013

Benn D Bolch T Hands K Gulley J Luckman ANicholson L Quincey D Thompson S Toumi R andWiseman S Response of debris-covered glaciers in theMount Everest region to recent warming and implicationsfor outburst flood hazards Earth-Sci Rev 114 156ndash174doi101016jearscirev201203008 2012

Berthier E and Vincent C Relative contribution of surface mass-balance and ice-flux changes to the accelerated thinning of Merde Glace French Alps over 1979ndash2008 J Glaciol 58 501ndash512doi1031892012JoG11J083 2012

Berthier E Arnaud Y Baratoux D Vincent C and Reacutemy FRecent rapid thinning of the ldquo Mer de Glace rdquo glacier derivedfrom satellite optical images Geophys Res Lett 31 L17401doi1010292004GL020706 2004

Berthier E Arnaud Y Kumar R Ahmad S Wagnon P andChevallier P Remote sensing estimates of glacier mass balancesin the Himachal Pradesh (Western Himalaya India) RemoteSens Environ 108 327ndash338 doi101016jrse2006110172007

Bhambri R Bolch T Kawishwar P Dobhal D P SrivastavaD and Pratap B Heterogeneity in Glacier response from 1973to 2011 in the Shyok valley Karakoram India The CryosphereDiscuss 6 3049ndash3078 doi105194tcd-6-3049-2012 2012

Bhutiyani M Mass-balance studies on Siachen Glacier in theNubra valley Karakoram Himalaya India J Glaciol 45 112ndash118 1999

Bishop M P Schroder J F and Ward J L SPOT multispec-tral analysis for producing supraglacial debris-load estimates forBatura glacier Pakistan Geocarto Int 10 81ndash90 1995

Bolch T Pieczonka T and Benn D I Multi-decadal mass lossof glaciers in the Everest area (Nepal Himalaya) derived fromstereo imagery The Cryosphere 5 349ndash358 doi105194tc-5-349-2011 2011

Bolch T Kulkarni A Kaumlaumlb A Huggel C Paul F Cogley JFrey H Kargel J Fujita K Scheel M Bajracharya S andStoffel M The state and fate of Himalayan Glaciers Science336 310ndash314 doi101126science1215828 2012

Bollasina M Ming Y and Ramaswamy V Anthropogenicaerosols and the weakening of the South Asian summer mon-soon Science 334 502ndash505 doi101126science12049942011

Bookhagen B and Burbank D Toward a complete Himalayan hy-drological budget spatiotemporal distribution of snowmelt andrainfall and their impact on river discharge J Geophys Res115 F03019 doi1010292009JF001426 2010

Bretherton C Widmann M Dymnikov V Wallace J and BladeacuteI The effective number of spatial degrees of freedom of a time-varying field J Climate 12 1990ndash2009 1999

Copland L Pope S Bishop M Shroder J Clendon PBush A Kamp U Seong Y and Owen L Glacier veloc-ities across the central Karakoram Ann Glaciol 50 41ndash49doi103189172756409789624229 2009

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

1270 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

SPOT5 DEM) to the unmeasured glaciers within each sub-region To estimate this regional extrapolation error we com-pare the ICESat elevation change trends (computed as inKaumlaumlb et al 2012) within 3 times 3 cells centered at the studysites (Table 5) with the ICESat-derived trends for the entiresub-regions The absolute difference between both trends isless than 002ndash003 m yrminus1 for Bhutan West Nepal Spiti La-haul and Karakoram ie negligible For Everest the eleva-tion trend for the entire sub-region is 012 m yrminus1 less nega-tive than for the 3 times 3 cells around the SPOT5 DEM cen-ter for Hindu Kush the trend for the entire sub-region is017 m yrminus1 more negative than for the 3 times 3 cell due to thevariability of elevation changes within this sub-region (Kaumlaumlbet al 2012) The regional extrapolation error is obtainedas the area-weighted mean absolute difference between themass balances derived from ICESat for those 3 times 3 cellsand the whole sub-region and equalsplusmn004 m we yrminus1

Finally for sub-regions and total PKH estimates of masschange we include an error on the glacier area computedfrom the RGI v20 (Arendt et al 2012) The inventory er-ror is specific to each study site and evaluated based on thecomparison of our user-defined inventory on each study sitewith the RGI v20 (Table 1) We note the remarkable accu-racy of the RGI for all but one of our study sites The rel-ative errors are generally of a few percents and up to 12 for West Nepal The differences between our and existing in-ventories are probably due to the difficulty of delimitatingdebris-covered glacier parts (eg Frey et al 2012 Paul etal 2013) and accumulation areas The Hengduan Shan studysite is an exception There the RGI is based on the ChineseGlacier Inventory (Shi et al 2009 Arendt et al 2012) andoverestimates the ice-covered area by 88 compared to ourLandsat-based inventory

Unless stated otherwise all error bars from this study andpreviously published studies correspond to one standard er-ror

4 Results

41 Overview of the elevation changes

The maps of glacier elevation changes are given in Figs 2ndash10for our nine study sites with as inset the distribution ofthe elevation differences off glaciers The full maps of ele-vation differences off glaciers are shown in the Supplement(Figs S2ndashS10) Those off glacier maps indicate a local ran-dom noise of aboutplusmn5ndash10 m However at length scales of afew kilometers the spatial variations in elevation differencesare small typically less than 1 m

For the four eastern study sites from the HengduanShan to the West Nepal (Fig 1) the elevation changesaveraged over the accumulation areas are small (between003plusmn 014 m yrminus1 to 013plusmn 015 m yrminus1) whereas clear sur-face lowering is observed for all four study sites in their

ablation areas ranging fromminus050plusmn 014 m yrminus1 (Bhutan)to minus081plusmn 013 m yrminus1 (Hengduan Shan) Over the BhutanEverest and West Nepal study sites glaciers in contact witha proglacial lake often experienced larger thinning rates

Further west the Spiti Lahaul is the only site wheresurface lowering is measured both in the accumula-tion area (minus034plusmn 016 m yrminus1) and the ablation area(minus070plusmn 014 m yrminus1) Clear surface lowering is also ob-served in the ablation areas of the Hindu Kush glaciers(minus050plusmn 018 m yrminus1)

In the Karakoram (east and west Figs 8 and 9) andin the Pamir (Fig 10) the pattern of elevation changesis highly heterogeneous due to the occurrence of numer-ous glacier surges or similar flow instabilities The eleva-tion changes in the accumulation areas of those three studysites are slightly positive (between+022plusmn 014 m yrminus1 and+030plusmn 018 m yrminus1) In the ablation areas small surfacelowering is observed for the two Karakoram study sites (av-erage of the two sitesminus033plusmn 016 m yrminus1) whereas no el-evation change is found in Pamir (minus002plusmn 013 m yrminus1)

Note that those mean rates of elevation changes are sensi-tive to the choice of the ELA assumed to equal the snowlinealtitude in Landsat images from around year 2000

42 Influence of a debris cover

To evaluate thinning rates over debris-covered and debris-free ice we perform a histogram adjustment in order tocompare data sets with similar altitude distributions This isneeded because debris-covered parts tend to be concentratedat lower elevations The adjustment consists in randomly ex-cluding pixels over debris-free ice from each elevation bandto match a scaled version of the debris-covered pixel distribu-tion (Kaumlaumlb et al 2012) Elevation changes with altitude canthen be compared for clean and debris-covered ice (Fig 11)

The thinning is higher for debris-covered ice in the Ever-est area and on the lowermost parts of glaciers in the PamirBut on average in the Pamir western Karakoram and SpitiLahaul the lowering of debris-free and debris-covered ice issimilar The thinning is stronger over clean ice in the Heng-duan Shan Bhutan West Nepal Hindu Kush and althoughto a lesser extent in the eastern Karakoram

43 Mass balance over the nine study sites

Glacier mass changes have been heterogeneous overthe PKH since 1999 Slight mass gains or balancedmass budgets are found in the north-west in thePamir (+014plusmn 013 m we yrminus1) and for the two studysites in the central Karakoram (+009plusmn 018 m we yrminus1

and +011plusmn 014 m we yrminus1) Mass loss is moderatein the adjacent Hindu Kush with a mass balance ofminus012plusmn 016 m we yrminus1 Glaciers in all our eastern studysites (two in Nepal Bhutan and Hengduan Shan) experi-enced homogeneous mass losses with mass balances ranging

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1271

Fig 2 Glacier elevation changes over the Hengduan Shan study site Inset distribution of the elevation differences off glacier The medianand the standard deviation of the elevation difference and the number of pixels off glaciers are given (by construction the mean elevationdifference is null)

Fig 3Same as Fig 2 but for the Bhutan study site

from minus033plusmn 014 m we yrminus1 to minus022plusmn 012 m we yrminus1The most negative mass balance is measured in Spiti Lahaul(western Himalaya) atminus045plusmn 013 m we yrminus1 (Table 4)

However within a study site mass balances can be highlyvariable from one glacier to another In the Everest area twolarge glaciers (71 km2 each) experienced distinctly differentmass balance withminus016plusmn 016 m we yrminus1 for the south-flowing Ngozumpa Glacier andminus059plusmn 016 m we yrminus1

for the north-flowing Rongbuk Glacier (Fig 4) In Bhutan(Fig 3) mass loss is higher south (minus025plusmn 013 m we yrminus1)

than north (minus014plusmn 012 m we yrminus1) of the main Hi-malayan ridge though not at a statistically significant level

The region-wide mass budget of PKH glaciers isnegative minus101plusmn 55 Gt yrminus1 (minus014plusmn 008 m we yrminus1)which corresponds to a sea level rise contribution of0028plusmn 0015 mm yrminus1 over 1999ndash2011

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1272 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 4Same as Fig 2 but for the Everest study site

Fig 5Same as Fig 2 but for the West Nepal study site

44 Glacier surges

In the Pamir and the Karakoram the spatial pattern of ele-vation changes of many glaciers is typical of surge events orflow instabilities They can be identified because of their veryhigh thinning and thickening rates that can both reach up to16 m yrminus1 (Figs 8ndash10)

Most of them have already been reported to be surge-type glaciers based on their velocity (Kotlyakov et al 2008Quincey et al 2011 Copland et al 2009 Heid and Kaumlaumlb2012) morphology (Copland et al 2011 Barrand and Mur-ray 2006 Hewitt 2007) or elevation changes (Gardelle etal 2012a) Here we only identify surge-type glaciers for thesake of mass balance calculation and we do not attempt toprovide an exhaustive up-to-date surge inventory This is why

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1273

Fig 6 Same as Fig 2 but for the Spiti Lahaul study site The map of elevation changes contains many voids due to the presence ofthin clouds during the acquisition of the SPOT5 DEM The mass balance of the Chhota Shigri and Bara Shigri glaciers are respectivelyminus039plusmn 018 m we yrminus1 andminus048plusmn 018 m we yrminus1

the published surge inventories in the Pamir and the Karako-ram do not necessarily match the surge-type glaciers labeledon the elevation change maps as these refer only to our ob-servation period (1999ndash2011)

The mass balance of surge-type glaciers (respectivelynon-surge-type glaciers) is +019plusmn 022 m we yrminus1

(+012plusmn 011 m we yrminus1) in the Pamir+009plusmn 030 m we yrminus1 (+009plusmn 015 m we yrminus1) inthe western Karakoram and+010plusmn 025 m we yrminus1

(+012plusmn 012 m we yrminus1) in the eastern Karakoram Theoverall similarity between the mass budgets for surge-typeand non-surge-type glaciers shows that those flow instabili-ties seem not to affect the glacier-wide mass balance at leastover short time periods here 10 yr

5 Discussion

51 SRTM penetration

Alternative estimates of SRTMC-bandpenetration for the PKHregion were obtained by Kaumlaumlb et al (2012) by comparing el-evations of the ICESat altimeter (Zwally et al 2002) with theSRTMC-bandDEM in PKH and extrapolating the 2003ndash2008trend of elevation changes back to the acquisition date ofSRTM The difference between east and west is more dis-tinct in our case with a mean penetration in the Karakoram of34 m (24plusmn 03 m in Kaumlaumlb et al 2012) 24 m in Bhutan and14 m around the Everest area (25plusmn 05 m for a wider areaincluding east Nepal and Bhutan in Kaumlaumlb et al 2012) Thiseastwest gradient is expected given that in February 2000when the SRTM mission was flown the snowfirn thicknesswas probably larger in the western PKH where glaciers areof winter-accumulation type than in the eastern PKH where

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1274 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 7Same as Fig 2 but for the Hindu Kush study site

glaciers are of summer-accumulation type Further studieswould be necessary to investigate the possible reasons be-hind the relatively low SRTMC-bandpenetration for the Pamir(18 m) that do not follow the eastwest gradient mentionedabove

Both our and Kaumlaumlb et alrsquos (2012) estimates agree withintheir error bars and seem thus to provide robust first-orderestimates for the SRTMC-bandradar penetration in the regionHowever it remains to be verified if a penetration depth com-puted at the regional scale is appropriate for a single smallglacier whose altitude distribution is different from the oneof whole study site (eg Abramov Glacier in the far north ofour Pamir study site)

52 Thinning over debris-covered ice

PKH glaciers are heavily debris-covered in their ablation ar-eas because of steep rock walls that surround them and in-tense avalanche activity Twelve percent (12 ) of the glacierarea in our study sites are covered with debris (up to 22 inthe Everest area Table 1) which is in agreement with pre-vious estimates for the Himalaya and the Karakoram 13 in Kaumlaumlb et al (2012) and 10 in Bolch et al (2012) Thedebris layer is expected to modify the ablation of the under-

lying ice It will increase ablation if its thickness is just a fewcentimeters (dominance of the albedo decrease) or decreaseablation it if the layer is thick enough to protect the ice fromsolar radiation (Mattson et al 1993 Mihalcea et al 2006Benn et al 2012 Lejeune et al 2013)

However recent studies have reported similar overall thin-ning rates over debris-covered and debris-free ice in thePKH (Kaumlaumlb et al 2012 Gardelle et al 2012a) or evenhigher lowering rates over debris-covered parts in the Ever-est area (Nuimura et al 2012) Our findings are consistentwith Nuimura et al (2012) for the Everest study site with astronger thinning (rate of elevation change ofminus097 m yrminus1)

in debris-covered areas than over clean ice (rate of elevationchange ofminus057 m yrminus1) In the Pamir Karakoram and SpitiLahaul study sites these rates are similar while in the HinduKush West Nepal Bhutan and Hengduan Shan study sitesthinning is greater over clean ice in agreement with the com-monly assumed protective effect of debris (Fig 4) Thosedifferences between our 9 study sites suggest a complex re-lationship between debris cover and glacier wastage

As local glacier thickness changes are a result of bothmass balance processes and a dynamic component theseunexpected high thinning rates over debris-covered parts

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1275

Fig 8Same as Fig 2 but for the Karakoram east study site Surge-type glaciers in an active surge phase (resp quiescent phase) are identifiedwith a solid triangle (resp solid circle)

are potentially the result of glacier-wide processes Over aglacier tongue the ablation can be enhanced by the pres-ence of steep exposed ice cliffs or supraglacial lakes andponds (Sakai et al 2000 2002) In addition it is also pos-sible that the glacier tongues are mostly covered by thindebris layers such that the albedo effect dominates the in-sulating effect resulting in an increased tongue-wide abla-tion The prevalence of thin debris was recently suggestedfor debris-covered glaciers around the Mount Gongga in thesouth-eastern Tibetan Plateau (Zhang et al 2013b) Finallyice-flow rates could be very low at debris-covered parts asshown by Quincey et al (2007) in the Everest area Kaumlaumlb(2005) for Bhutan and Scherler et al (2011b) in the HinduKush-Karakoram-Himalaya Thus all three factors are likelyto increase the overall thinning under debris despite the un-doubted protective effect of a continuous thick debris coverat a local scale Future investigations combining surface ve-locity fields and maps of elevation changes could allow eval-uating more closely the relative role of ablation and ice fluxesin thinning rates over PKH glacier tongues (eg Berthierand Vincent 2012 Nuth et al 2012) Measurements of the

spatial distribution of debris thickness over the PKH glaciertongues are also needed

Note that in the case of debris-covered tongues the ele-vation change measurement represents the glacier thicknesschange but also the possible debris thickness evolution Thelatter remains poorly known in the PKH or in any othermountain range But based on the debris discharges givenby Bishop et al (1995) for Batura Glacier (Pakistan) 48to 90times 103 m3 yrminus1 the thickening of the debris can be es-timated between 27 and 55 mm yrminus1 which is negligiblegiven the magnitude of the glacier elevation changes

53 Comparison to previous mass balance estimates

Throughout the PKH there is only a single peer-reviewedglaciological record (on Chhota Shigri Glacier) coveringmost of our study period (Wagnon et al 2007 Vincent etal 2013) Our geodetic mass balance estimate for the SpitiLahaul study site has been recently compared to those glacio-logical measurements on Chhota Shigri Glacier (Vincent etal 2013) Mass balance records reported in Yao et al (2012)for the Tibetan Plateau and the surroundings mountain ranges

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1276 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 9Same as Fig 8 but for the Karakoram west study site

only start during glaciological year 20052006 This is whyhere we do not compare our mass balance estimates toglaciological records and limit our comparison to others re-gional estimates obtained using the geodetic or gravimetricmethods

We stress that the comparison of our mass balance mea-surements to published estimates is complicated by thefact that generally they do not cover the same time pe-riod Little is known about the inter-annual variability inthe PKH during the 21st century due to the limited num-ber of long glaciological time series The 9 yr mass balance

record on Chhota Shigri Glacier reveals a relatively highinter-annual variability of the annual mass balance (stan-dard deviationplusmn 074 m we yrminus1 during 2003ndash2011 Vin-cent et al 2013) which is similar to the one obtained be-tween 1968 and 1998 on Abramov Glacier (standard devi-ation 069 m we yrminus1 WGMS 2012) Thus inter-annualvariability can explain part of the difference between two cu-mulative mass balance estimates over 5ndash10 yr even if they areoffset by only one or two years

Based on ICESat repeat track measurements and theSRTM DEM Kaumlaumlb et al (2012) measured a region-wide

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1277

Fig 10Same as Fig 8 but for the Pamir study site

mass balance ofminus021plusmn 005 m we yrminus1 between 2003and 2008 for Hindu Kush-Karakoram-Himalaya glaciersFor that same area (ie excluding the Pamir and theHengduan Shan study sites) we found a mass balance ofminus015plusmn 007 m we yrminus1 Thus our result agrees with Kaumlaumlbet al (2012) within the error bars although our study periodis longer (1999ndash2011) and our measurement principle andextrapolation approach are different In addition we havecalculated updated mass balances from Kaumlaumlb et al (2012)ie averages over 3times 3 degree cells centered over our studysites and find that they are consistent within error barswith our mass balance estimates (Table 5) Similar resultsare found when our mass balances are compared to a spa-tially extended ICESat analysis for 2003ndash2009 (Gardner etal 2013) Mass losses measured with ICESat (Kaumlaumlb et al2012 Gardner et al 2013) tend to be slightly larger thanours This could be due to the difference in time periodsAlso our smaller mass losses could be partly explained if wesystematically underestimated the poorly constrained pene-

tration of the C-band andor X-band radar signal into ice andsnow

At a more local scale Bolch et al (2011) reported a massbalance ofminus079plusmn 052 m yrminus1 between 2002 and 2007 byDEM differencing over a glacier area of 62 km2 includingten glaciers south and west of Mt Everest For these sameglaciers Nuimura et al (2012) found a mean mass balanceof minus045plusmn 060 m yrminus1 during 2000ndash2008 while they com-puted a mass balance ofminus040plusmn 025 m yrminus1 between 1992and 2008 over a glacier area of 183 km2 around Mt EverestIn the present study we found an average mass balance ofminus041plusmn 021 m yrminus1 over the ten glaciers mentioned above(Table 5 Fig S1) and a value ofminus026plusmn 013 m yrminus1 overthe whole Everest study site between 2000 and 2011 Ourvalues are comparable to those of Nuimura et al (2012)Our mass balance is less negative than the 2002ndash2007 massbalance of Bolch et al (2011) but not statistically differentwhen error bars are considered (Fig 12) Indeed Bolch etal (2011) acknowledged some high uncertainties for their

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1278 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Table 4Glacier mass budget of the nine sub-regions to which the results of the study sites (mass balances) have been extrapolated For eachof them the total glacierized area as well as the percentage of the glacier area over which the mass balance was computed (ie study sites)are given The total glacier area is derived from the RGI v20 (Arendt et al 2012) The error for the mass balance in each sub-region is theroot of sum of squares (RSS) of the mass balance error of each study site and the regional extrapolation error The error for the mass budgetin each sub-region includes additionally the error from the total glacier area in the RGI v20

Sub-region Total glacier Measured glacier Mass balance Mass budgetarea (km2) area () (m we yrminus1) (Gt yrminus1)

Hengduan Shan 11584 12 minus033plusmn 014 minus38plusmn 38Bhutan 4021 34 minus022plusmn 013 minus09plusmn 05Everest 6226 23 minus026plusmn 014 minus16plusmn 08West Nepal 6849 13 minus032plusmn 014 minus22plusmn 10Spiti Lahaul 9043 23 minus045plusmn 014 minus41plusmn 13Hindu Kush 6135 13 minus012plusmn 016 minus07plusmn 10Karakoram East

1902428 +011 plusmn 014 +19 plusmn 31

Karakoram West 30 +009 plusmn 018Pamir 9369 34 +014 plusmn 014 +13 plusmn 13

TotalArea- 72251 30 minus014plusmn 008 minus101plusmn 55weighted average

Table 5Comparison of geodetic mass balances between this study and previous published results with overlapping study periods and similargeographic locations Updated figures from Kaumlaumlb et al (2012) are averaged over 3 times 3 cells centered over the study sites of the presentstudy

GlacierSite name This study Bolch et al (2011) Nuimura et al (2012) Kaumlaumlb et al (2012 updated)2002ndash2007 2000ndash2008 2003ndash2008

Study Mass balance Area (km2)a Mass balance Area Mass balance Mass balanceperiod (m we yrminus1) (m we yrminus1) (km2) (m we yrminus1) (m we yrminus1)

Hengduan San 1999ndash2010 minus022 plusmn 014 1303 (55 )Bhutan 1999ndash2010 minus022 plusmn 012 1384 (64 ) minus052 plusmn 016b

Everest

1999ndash2011

minus026 plusmn 013 1461 (58 ) minus039 plusmn 011AX010 minus090plusmn 034 04Changri SharNup minus042plusmn 017 161 (79 ) minus029plusmn 052 130 minus055plusmn 038Khumbu minus051plusmn 019 203 (47 ) minus045plusmn 052 170 minus076plusmn 052Nuptse minus037plusmn 020 50 (74 ) minus040plusmn 053 40 minus034plusmn 027Lhotse Nup minus021plusmn 027 24 (70 ) minus103plusmn 051 19 minus022plusmn 047Lhotse minus043plusmn 018 85 (83 ) minus110plusmn 052 65 minus067plusmn 051Lhotse SharImja minus070plusmn 052 98 (44 ) minus145plusmn 052 107 minus093plusmn 060Amphu Laptse minus046plusmn 034 25 (38 ) minus077plusmn 052 15 minus018plusmn 094Chukhung +044 plusmn 024 42 (47 ) +004 plusmn 054 38 +043 plusmn 081Ama Dablam minus049plusmn 017 36 (54 ) minus056plusmn 052 22 minus056plusmn 073Duwo minus016plusmn 026 19 (18 ) minus196plusmn 053 10 minus068plusmn 074Total Khumbu minus041plusmn 021 744 minus079plusmn 052 617 minus045plusmn 060(10 Glaciers above)

West Nepal 1999ndash2011 minus032 plusmn 013 908 (40 ) minus032 plusmn 012Spiti Lahaul 1999ndash2011 minus045 plusmn 013 2110 (46 ) minus038 plusmn 006Karakoram East 1999ndash2010 +011 plusmn 014 5328 (42 ) minus004 plusmn 004Karakoram West 1999ndash2008 +009 plusmn 018 5434 (45 )Hindu Kush 1999ndash2008 minus012 plusmn 016 793 (80 ) minus020 plusmn 006Pamir 1999ndash2011 +014 plusmn 013 3178 (50 )

a The total glacier area is given in km2 and in parenthesis the of the glacier area actually covered with measurementsb For this cell the ICESat coverage is insufficient and does not sample all glacier elevations

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1279

Fig 11 Elevation changes (SPOT5-SRTM) in the ablation area for clean ice (blue circles) and debris-covered ice (black circles) for eachstudy site For the Karakoram (east and west) and the Pamir study sites only non-surge-type glaciers are considered Standard error of theelevation differences are typically less thanplusmn2 m with each 100 m elevation band (not shown for the sake of clarity) Note elevation changesover separate sections of a glacier cannot be treated as mass changes due to the disregard of glacier dynamics

estimate over this short time period The differences in sur-vey periods and in the glacier outlines may also explainpart of these discrepancies In particular the delimitation ofthe accumulation area is notoriously difficult and Bolch etal (2011) did not include some of the steep high altitudeslopes which we included due to their potential avalanchecontribution For the 10 glaciers mentioned above our massbalances would become 012 m we yrminus1 more negative ifwe used the exact same outlines as Bolch et al (2011)Our positive mass balance for the small Chukhung Glacier(+044plusmn 024 m we yrminus1) is in agreement with Nuimura etal (2012) but enigmatic in a context of regional glacier lossand deserves further investigation If the 2002ndash2007 massbalance estimate by Bolch et al (2011) is excluded all otherrecent mass balance estimates suggest no acceleration of themass loss in the Everest area when compared to the long-term(1970ndash2007) mass balance measured by Bolch et al (2011)in this regionminus032plusmn 008 m weyrminus1

In Bhutan our results indicate a stronger mass loss forsouthern glaciers than for northern ones Interestingly in thisarea Kaumlaumlb (2005) measured high speeds (up to 200 m yrminus1)

for glaciers north of the Himalayan main ridge and ratherlow velocities over southern glacier tongues by using repeat

ASTER data This is consistent with the more negative massbalance that we report for the southern glaciers in Bhutanie for the slow flowing presumably downwasting glaciers

Berthier et al (2007) reported a mass balance betweenminus069 andminus085 m we yrminus1 for an ice-covered area of915 km2 around Chhota Shigri Glacier between 1999 (SRTMDEM) and 2004 (SPOT5-HRG DEM) For the same areawe found a mass balance ofminus045plusmn 013 m we yrminus1 over1999ndash2011 The difference in the study period may explainpart of the differences Also the methodology by Berthieret al (2007) did not explicitly take into account the SRTMradar penetration over glaciers and also corrected an appar-ent elevation-dependent bias according to glacier elevationswhich is now known as a resolution issue and is best cor-rected according to the terrain maximum curvature (Gardelleet al 2012b) More details and discussions about these dif-ferences can be found in Vincent et al (2013)

Importantly the results of the eastern Karakoram studysite confirm the balanced or slightly positive mass budgetreported by Gardelle et al (2012a) in the western Karako-ram over the past decade The two Karakoram study sitesare adjacent with a small overlapping area (sim 1100 km2 ofglaciers) where elevation differences agree within 1 m Both

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1280 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 12 Comparison of geodetic mass balances for 10 glaciers inthe Mount Everest area 1999ndash2010 mass balances measured inthe present study are compared with published estimates during2002ndash2007 (Bolch et al 2011) and during 2000ndash2008 (Nuimuraet al 2012) The 1-to-1 line is drawn The mean difference (plusmn1standard deviation) between (i) our values and Bolch et al (2011)are (047plusmn 058 m we yrminus1) and (ii) our values and Nuimura etal (2012) are (012plusmn 021 m we yrminus1) Part of the differences be-tween estimates may be due to the different periods surveyed andthe different glacier outlines used

sites show similar mass balances over slightly different pe-riods +011plusmn 014 m yrminus1 over 1999ndash2010 in the east and+009plusmn 018 m yrminus1 over 1999ndash2008 in the west This con-firms a ldquoKarakoram anomalyrdquo that was first suggested basedon field observations (Hewitt 2005) We report a mass bal-ance of+010plusmn 013 m yrminus1 for Baltoro Glacier (see loca-tion in Fig 8 size= 304 km2) between 1999 and 2010which is consistent with its gradual speed-up reported since2003 (Quincey et al 2009) However given the completelydifferent methods used in our and the latter study we can-not conclude that this demonstrates a real shift from nega-tive to positive mass balance The mass budget of SiachenGlacier (1145 km2 sometimes referred to as the highestbattlefield in the world) is also balanced or slightly pos-itive (+014plusmn 014 m we yrminus1) Thus the alarming state-ment by Rasul (2012) that this glacier retreated by 59 kmand lost 17 of its mass during 1989ndash2009 is very likelyerroneous Again within error bars our regional mass bal-ance in the Karakoram agrees with the mass balance ofminus003plusmn 004 m yrminus1 found by Kaumlaumlb et al (2012) over 2003ndash2008 and with the mass balance ofminus010plusmn 018 m we yrminus1

(2-sigma error level) obtained by Gardner et al (2013) for aregion including both the Karakoram and the Hindu Kush

In the northernmost part of the Pamir in the Alay moun-tain range the mass balance of Abramov Glacier has beenmeasured in the field for 30 yr between 1968 and 1998(WGMS 2012) with a mean value ofminus046 m yrminus1 Herewe report a mass balance ofminus003plusmn 014 m yrminus1 for this

glacier over 1999ndash2011 Our near zero mass budget estimatefor this glacier disagrees with negative mass balances recon-structed for the same period with a model run using biascorrected NCEPNCAR re-analysis data and calibrated withfield data (Barundun et al 2013) An underestimate of ourSRTM C-Band penetration in the Pamir as a whole and inparticular for the small Abramov Glacier may contribute tothese differences

For the whole Pamir study site our mass budget is bal-anced or slightly positive (+014plusmn 013 m we yrminus1) In theeastern Pamir the mass balance of Muztag Ata Glacier was+025 m we yrminus1 between 2005 and 2010 (Yao et al 2012)Although Muztag Ata Glacier is located outside of our Pamirstudy site and his glaciological records only cover the sec-ond half of our study period this measurement is consistentwith our remotely sensed mass balance Thus the ldquoKarako-ram anomalyrdquo should probably be renamed the ldquoPamir-Karakoram anomalyrdquo

This slightly positive or balanced mass budget measuredin the Pamir seems to contradict previous findings by Heidand Kaumlaumlb (2012) They reported rapidly decreasing glacierspeeds in this region between two snapshots of annual ve-locities in 20002001 and 20092010 an observation thatwould fit with negative mass balances (Span and Kuhn 2003Azam et al 2012) We note though that the relationship be-tween glacier mass balance and ice fluxes is not straightfor-ward and might in particular involve a time delay betweena change in mass balance and the related dynamic response(Heid and Kaumlaumlb 2012) Thus the decrease in speed couldwell be due to negative mass balances before or partiallybefore 1999ndash2011 We acknowledge that this discrepancyneeds to be investigated further in particular by examiningcontinuous changes in annual glacier speed through the firstdecade of the 21st Century Indeed Heid and Kaumlaumlb (2012)measured differences between two annual periods which maynot represent mean decadal velocity changes

Jacob et al (2012) using satellite gravimetry (GRACE)measured a mass budget ofminus5plusmn 6 Gt yrminus1 over 2003ndash2010for the ldquoKarakoram-Himalayardquo their region 8c which cor-responds to our study area excluding the Pamir Karakoramand Hindu Kush sub-regions but including some glaciersnorth of the Himalayan ridge already on the Tibetan PlateauFor this subset of our PKH study area we measured amass budget ofminus126plusmn 42 Gt yrminus1 The two estimates over-lap within their error bars especially if our error is dou-bled to match the two standard error level used by Jacob etal (2012) The estimate of Jacob et al (2012) over our com-plete study region (PKH) ieminus6plusmn 8 Gt yrminus1 (sum of theirregions 8b and 8c) also includes mass changes for the KunlunShan sub-region for which we did not measure glacier massbalances Therefore the comparison with our PKH value(minus101plusmn 55 Gt yrminus1) is not straightforward but suggests areasonable agreement especially if we consider the slightmass gain (15plusmn 17 Gt yrminus1) measured with ICESat in theWest Kunlun (Gardner et al 2013)

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1281

Table 6 Annual average of the glacier seasonal and decadal mass loss contributions to Indus Ganges and Brahmaputra discharges Basinarea and seasonal contributions are given by Kaser et al (2010) glacier area in each basin is derived from the RGI v20 (Arendt et al 2012)Numbers in parentheses indicate the percentage of glacierized area within the river basin Seasonal contributions were modeled by Kaser etal (2010) based on the CRU 20 dataset over 1961ndash1990 assuming a balanced glacier mass budget Glacier decadal mass loss contributionsare given as annual average over 1999ndash2011

River Basin area Glacier area Glacier contribution Mean discharge(km2) (km2) (m3 sminus1) (m3 sminus1)

Seasonally delayed Decadal mass lossKaser et al (2010) (this study)

Indus 1 139 814 25 598 (2 ) 105 103plusmn 72 2146 (Chevallier personalcommunication 2012)

Ganges 1 023 609 11 168 (1 ) 47 103plusmn 49 11739 Papa et al (2012Brahmaputra 527 666 15 296 (3 ) 33 147plusmn 64 19710 updated)

54 Reasons for the heterogeneous pattern of massbalance

The PKH glacier mass balance is slightly negative(minus014plusmn 008 m we yrminus1) but changes are not homoge-neous from one sub-region to another and seem to coincidewith the climatic settings of the mountain range For the east-ern sites (from the Hengduan Shan to West Nepal) which areunder the influence of the Indian and south-east Asian sum-mer monsoons the mass balances are moderately negative(sim minus026 m we yrminus1) In the northwest of the PKH at thePamir and Karakoram sites where glacier climate is char-acterized by the westerlies in winter mass balances are bal-anced or slightly positive (sim +010 m we yrminus1) In betweenthese two influences the glaciers of the Spiti Lahaul sitein a monsoon-arid transition zone experienced the strongestmass loss (minus045plusmn 013 m we yrminus1)

This heterogeneous pattern of mass balances seems to berelated to trends in precipitation throughout the PKH Archerand Fowler (2004) reported an increase in winter precipi-tation over the Upper Indus Basin since 1961 a trend con-firmed by Yao et al (2012) over the Pamir and the Karako-ram based on the analysis of the Global Precipitation Cli-matology Project data set (GPCP Adler et al 2003) Thisprecipitation increase is a source of greater accumulation forglaciers and thus a possible explanation for their slightly pos-itive mass budgets In addition in the Upper Indus Basin themean summer temperature has been decreasing since 1961(Fowler and Archer 2006) which is consistent with the find-ings of Shekhar et al (2010) in the Karakoram who reporteda decrease in both maximum and minimum temperature since1988 These increases in winter precipitation and summertemperature likely translate in greater accumulation and re-duced ablation changes which are consistent with the bal-anced or slightly positive glacier mass budget

On the other hand in the east of the PKH the Indiansummer monsoon has been weakening since the 1950s (Bol-lasina et al 2011) which could have reduced the amount

of snowfall over the eastern study sites and thus resultedin negative mass balances In addition maximum and min-imum temperatures increased since 1988 in the western Hi-malaya (Shekhar et al 2010) and maximum temperaturesincreased in the central Himalaya (Nepal) since the mid-1970s (Shrestha et al 1999) Thus both precipitation andtemperature trends in the Himalaya likely contributed to thenegative mass balances measured over the correspondingstudy sites (from Spiti Lahaul to Hengduan Shan Fig 1)

However gridded data sets such as GPCP are not neces-sarily representative of the climate at high altitude IndeedHewitt (2005) reported a 5-to-10-fold increase in precipi-tation between the glacier front and the accumulation areain the Karakoram that is not captured by reanalysis dataas recently confirmed by Immerzeel et al (2012) Thus di-rect measurements of accumulation on High Mountain Asiaglaciers (eg Azam et al 2013) and high-altitude weatherstations are eagerly needed to validate the large scale grid-ded data set (eg reanalysis) test their ability to describe thespecific climate near to PKH glaciers and assess the relativerole played by temperature and precipitation changes

55 Contribution to water resources

Our glacier mass balance can be averaged over large riverbasins to estimate the mean contribution to river runoffdue to a decade of glacier mass loss We assume therebyonly direct river runoff without loss terms For the threerivers for which we cover the entire glacierized area of theircatchments within our sub-regions (Fig 1) these contribu-tions to the river discharge are equal to 103 m3 sminus1 (Indus)103 m3 sminus1 (Ganges) and 147 m3 sminus1 (Brahmaputra) for theperiod 1999ndash2011 (Table 6)

Rivers of PKH being key water resources measures oftheir discharges are highly sensitive and difficult to access forpolitical reasons Papa et al (2012) built recent time series ofdischarges for Ganges and Brahmaputra by using altimetrymeasurements The locations of the corresponding dischargeestimates in Bangladesh are given in Fig 1 We can therefore

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1282 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

evaluate the relative contribution of glacier decadal mass loss(Table 6) to annual river run-off at 09 for the Gangesat Hardinge (basin area ofsim 1 020 000 km2) and 07 forBrahmaputra at Bahadurabad (basin area ofsim 530 000 km2)between 1999 and 2011 Given the discharge measured at theGuddu station by the Surface Water Hydrology Project of theWater and Power Development Authority (Pakistan see lo-cation on Fig 1) between 1999 and 2003 we can estimatethe contribution of glacier mass loss at 48 for the IndusRiver at this station

The seasonal contributions of glaciers to river run-off (iethe part of the annual precipitation that experiences season-ally delayed release from the glaciers) directly results fromseasonal variations of meteorological quantities in contrastto the glacier mass loss contribution which results from thelong-term climatic change Both runoff components directlyaffect the amount of water available in a river at a given lo-cation and at a given time so that their comparison could beof interest Even if the seasonal timing of the glacier massloss contribution is unclear it could in general peak at asimilar time of the year than the seasonal glacier contribu-tion so that both components rather enlarge each other thancancel Seasonal glacier contributions have been modeled byKaser et al (2010) The latter study assumed glaciers to bein equilibrium with the present climate and thus prescribedglacier mass balances equal to 0 Thus their seasonally de-layed glacier contributions do not include the part due todecadal glacier mass loss and is complementary to our es-timates (Table 6) For the eastern basins under the influenceof the Indian summer monsoon (Ganges and Brahmaputra)the contribution of glacier mass loss is higher than the sea-sonal contribution In other words for the first decade of the21st century the mass loss of glaciers is on average moreimportant for water availability in those two rivers than theirseasonal water storage effect The situation is different for theIndus Basin where the glacier melt season is also the dry sea-son which results in a relatively higher seasonally delayedglacier contribution to the river discharge There the season-ally delayed water release from the glaciers and the decadalglacier mass loss contribute on average equally to the riverdischarge Both contributions are strongly (and equally) de-pendent on the location of the discharge measurement andwill be significantly larger (in relative term) in more heav-ily glacierized and smaller mountain catchments located inthe upper part of these major river basins (Bookhagen andBurbank 2010)

6 Conclusion

In this study we assessed the spatial pattern of glaciermass balance over the PKH by comparing the SRTM andSPOT5 DEMs over 9 well-distributed study sites between1999 and 2011 We found slightly positive or balanced massbudgets in the western part (Pamir Karakoram) moder-ate mass loss in the east (Nepal Bhutan Hengduan Shan)

and the most negative mass balance in the monsoon-aridtransition zone (Spiti Lahaul in the western Himalaya)By extrapolating these values to climatically homogeneoussub-regions we estimate the PKH overall mass balanceto have beenminus014plusmn 008 m we yrminus1 between 1999 and2011 which corresponds to a contribution to sea level riseof 0028plusmn 0015 mm yrminus1 This is about 4 of the globalglacier mass loss estimated by Gardner et al (2013) thoughover a slightly shorter time period (2003ndash2009) The largestsource of uncertainties in those geodetic mass balance esti-mates is the poorly constrained penetration of the C-BandSRTM radar signal into snow and ice

Our technique of DEM differencing allows mapping in de-tail the spatial pattern of glacier elevation changes Thosemaps reveal many surge-type behaviors both in the Pamirand the Karakoram It also appears that the glacier-wide massbalance does not seem to be considerably affected by themass transfer from surges over thesim 10 yr time period stud-ied Similar lowering rates over debris-covered and clean iceare observed for four study sites despite the commonly as-sumed insulating effect of debris covers This suggests forthe scale of entire glacier tongues a spatial mixture of re-duced ablation under intact thick debris covers and enhancedablation by supraglacial lakes or ice cliffs or due to thin de-bris layer and very low glacier velocities in ablation areasthat reduce the ice advection downstream

The regionally heterogeneous pattern of ice wastage is inbroad agreement with contrasted trends in large-scale pre-cipitation and temperature However glaciological (eg ac-cumulation) and meteorological records are needed at highaltitude in the vicinity of glaciers to evaluate more closelythe local influence of atmospheric variables on glacier massbalance and fully understand the reasons behind those con-trasted mass budgets in the PKH

Supplementary material related to this article isavailable online at httpwwwthe-cryospherenet712632013tc-7-1263-2013-supplementpdf

Acknowledgements We thank G Cogley A Gardner T NuimuraT Bolch P Wagnon M Barandun M Hoelzle M Huss and ananonymous referee for their constructive comments that led to aconsiderably improved manuscript We thank the Water and PowerDevelopment Authority (WAPDA) for their hydrological data(processed by P Chevallier) and F Papa for sharing his updatedrun-off data ahead of publication We thank T Bolch for sharinghis glacier outlines from the Everest area We thank the USGSfor allowing free access to their Landsat archive CGIAR-SCI forSRTM C-band data DLR for SRTM X-band data and GLIMS forglacier inventories J Gardelle acknowledges a PhD fellowshipfrom the French Space Agency (CNES) and the French NationalResearch Center (CNRS) E Berthier acknowledges support fromCNES through the TOSCA and ISIS proposal no 397 and fromthe Programme National de Teacuteleacutedeacutetection Spatiale E Berthieracknowledges M Masiokas and R Villalba for welcoming him

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1283

at the IANIGLA (Mendoza Argentina) Y Arnaud acknowledgessupport from French National Research Agency through ANR-09-CEP-005-01PAPRIKA A Kaumlaumlb acknowledges support fromESA through Glaciers_cci (400010177810I-AM) and from theEuropean Research Council under the European Unionrsquos SeventhFramework Programme (FP2007-2013)ERC Grant Agreementn 320816

Edited by I M Howat

The publication of this articleis financed by CNRS-INSU

References

Adler R Huffman G Chang A Ferraro R Xie P JanowiakJ Rudolf B Schneider U Curtis S Bolvin D Gruber ASusskind J Arkin P and Nelkin E The Version-2 GlobalPrecipitation Climatology Project (GPCP) Monthly Precipita-tion Analysis (1979ndashPresent) J Hydrometeorol 4 1147ndash11672003

Archer D R and Fowler H J Spatial and temporal variationsin precipitation in the Upper Indus Basin global teleconnectionsand hydrological implications Hydrol Earth Syst Sci 8 47ndash61doi105194hess-8-47-2004 2004

Arendt A Bolch T Cogley J G Gardner A Hagen J-OHock R Kaser G Pfeffer W T Moholdt G Paul F RadicV Andreassen L Bajracharya S Beedle M Berthier EBhambri R Bliss A Brown I Burgess E Burgess DCawkwell F Chinn T Copland L Davies B de AngelisH Dolgova E Filbert K Forester R Fountain A Frey HGiffen B Glasser N Gurney S Hagg W Hall D Hari-tashya U K Hartmann G Helm C Herreid S Howat IKapustin G Khromova T Kienholz C Koenig M KohlerJ Kriegel D Kutuzov S Lavrentiev I LeBris R Lund JManley W Mayer C Miles E Li X Menounos B Mer-cer A Moelg N Mool P Nosenko G Negrete A NuthC Pettersson R Racoviteanu A Ranzi R Rastner P RauF Rich J Rott H Schneider C Seliverstov Y Sharp MSigurethsson O Stokes C Wheate R Winsvold S WolkenG Wyatt F and Zheltyhina N Randolph Glacier Inventory[v20] A Dataset of Global Glacier Outlines Global Land IceMeasurements from Space Boulder Colorado USA Digital Me-dia httpwwwglimsorgRGIrandolphhtml 2012

Azam M Wagnon P Ramanathan A Vincent C Sharma PArnaud Y Linda A Pottakkal J Chevallier P Singh V andBerthier E From balance to imbalance a shift in the dynamicbehaviour of Chhota Shigri glacier western Himalaya India JGlaciol 58 315ndash324 doi1031892012JoG11J123 2012

Azam M F Wagnon P Vincent C Ramanathan A Linda Aand Singh V B Reconstruction of the annual mass balanceof Chhota Shigri Glacier (Western Himalaya India) since 1969Ann Glaciol in revision 2013

Barrand N and Murray T Multivariate Controls on the In-cidence of Glacier Surging in the Karakoram Himalaya

Arct Antarct Alp Res 38 489ndash498 doi1016571523-0430(2006)38[489MCOTIO]20CO2 2006

Barundun M Huss M Azisov E Gafurov A HoelzleM Merkushkin A Salzmann N and Usubaliev R Re-establishing seasonal mass balance observation at AbramovGlacier Kyrgyzstan from 1968ndash2012 Abstract EGU2013-5668European Geophysical Union Vienna 2013

Benn D Bolch T Hands K Gulley J Luckman ANicholson L Quincey D Thompson S Toumi R andWiseman S Response of debris-covered glaciers in theMount Everest region to recent warming and implicationsfor outburst flood hazards Earth-Sci Rev 114 156ndash174doi101016jearscirev201203008 2012

Berthier E and Vincent C Relative contribution of surface mass-balance and ice-flux changes to the accelerated thinning of Merde Glace French Alps over 1979ndash2008 J Glaciol 58 501ndash512doi1031892012JoG11J083 2012

Berthier E Arnaud Y Baratoux D Vincent C and Reacutemy FRecent rapid thinning of the ldquo Mer de Glace rdquo glacier derivedfrom satellite optical images Geophys Res Lett 31 L17401doi1010292004GL020706 2004

Berthier E Arnaud Y Kumar R Ahmad S Wagnon P andChevallier P Remote sensing estimates of glacier mass balancesin the Himachal Pradesh (Western Himalaya India) RemoteSens Environ 108 327ndash338 doi101016jrse2006110172007

Bhambri R Bolch T Kawishwar P Dobhal D P SrivastavaD and Pratap B Heterogeneity in Glacier response from 1973to 2011 in the Shyok valley Karakoram India The CryosphereDiscuss 6 3049ndash3078 doi105194tcd-6-3049-2012 2012

Bhutiyani M Mass-balance studies on Siachen Glacier in theNubra valley Karakoram Himalaya India J Glaciol 45 112ndash118 1999

Bishop M P Schroder J F and Ward J L SPOT multispec-tral analysis for producing supraglacial debris-load estimates forBatura glacier Pakistan Geocarto Int 10 81ndash90 1995

Bolch T Pieczonka T and Benn D I Multi-decadal mass lossof glaciers in the Everest area (Nepal Himalaya) derived fromstereo imagery The Cryosphere 5 349ndash358 doi105194tc-5-349-2011 2011

Bolch T Kulkarni A Kaumlaumlb A Huggel C Paul F Cogley JFrey H Kargel J Fujita K Scheel M Bajracharya S andStoffel M The state and fate of Himalayan Glaciers Science336 310ndash314 doi101126science1215828 2012

Bollasina M Ming Y and Ramaswamy V Anthropogenicaerosols and the weakening of the South Asian summer mon-soon Science 334 502ndash505 doi101126science12049942011

Bookhagen B and Burbank D Toward a complete Himalayan hy-drological budget spatiotemporal distribution of snowmelt andrainfall and their impact on river discharge J Geophys Res115 F03019 doi1010292009JF001426 2010

Bretherton C Widmann M Dymnikov V Wallace J and BladeacuteI The effective number of spatial degrees of freedom of a time-varying field J Climate 12 1990ndash2009 1999

Copland L Pope S Bishop M Shroder J Clendon PBush A Kamp U Seong Y and Owen L Glacier veloc-ities across the central Karakoram Ann Glaciol 50 41ndash49doi103189172756409789624229 2009

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1271

Fig 2 Glacier elevation changes over the Hengduan Shan study site Inset distribution of the elevation differences off glacier The medianand the standard deviation of the elevation difference and the number of pixels off glaciers are given (by construction the mean elevationdifference is null)

Fig 3Same as Fig 2 but for the Bhutan study site

from minus033plusmn 014 m we yrminus1 to minus022plusmn 012 m we yrminus1The most negative mass balance is measured in Spiti Lahaul(western Himalaya) atminus045plusmn 013 m we yrminus1 (Table 4)

However within a study site mass balances can be highlyvariable from one glacier to another In the Everest area twolarge glaciers (71 km2 each) experienced distinctly differentmass balance withminus016plusmn 016 m we yrminus1 for the south-flowing Ngozumpa Glacier andminus059plusmn 016 m we yrminus1

for the north-flowing Rongbuk Glacier (Fig 4) In Bhutan(Fig 3) mass loss is higher south (minus025plusmn 013 m we yrminus1)

than north (minus014plusmn 012 m we yrminus1) of the main Hi-malayan ridge though not at a statistically significant level

The region-wide mass budget of PKH glaciers isnegative minus101plusmn 55 Gt yrminus1 (minus014plusmn 008 m we yrminus1)which corresponds to a sea level rise contribution of0028plusmn 0015 mm yrminus1 over 1999ndash2011

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1272 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 4Same as Fig 2 but for the Everest study site

Fig 5Same as Fig 2 but for the West Nepal study site

44 Glacier surges

In the Pamir and the Karakoram the spatial pattern of ele-vation changes of many glaciers is typical of surge events orflow instabilities They can be identified because of their veryhigh thinning and thickening rates that can both reach up to16 m yrminus1 (Figs 8ndash10)

Most of them have already been reported to be surge-type glaciers based on their velocity (Kotlyakov et al 2008Quincey et al 2011 Copland et al 2009 Heid and Kaumlaumlb2012) morphology (Copland et al 2011 Barrand and Mur-ray 2006 Hewitt 2007) or elevation changes (Gardelle etal 2012a) Here we only identify surge-type glaciers for thesake of mass balance calculation and we do not attempt toprovide an exhaustive up-to-date surge inventory This is why

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1273

Fig 6 Same as Fig 2 but for the Spiti Lahaul study site The map of elevation changes contains many voids due to the presence ofthin clouds during the acquisition of the SPOT5 DEM The mass balance of the Chhota Shigri and Bara Shigri glaciers are respectivelyminus039plusmn 018 m we yrminus1 andminus048plusmn 018 m we yrminus1

the published surge inventories in the Pamir and the Karako-ram do not necessarily match the surge-type glaciers labeledon the elevation change maps as these refer only to our ob-servation period (1999ndash2011)

The mass balance of surge-type glaciers (respectivelynon-surge-type glaciers) is +019plusmn 022 m we yrminus1

(+012plusmn 011 m we yrminus1) in the Pamir+009plusmn 030 m we yrminus1 (+009plusmn 015 m we yrminus1) inthe western Karakoram and+010plusmn 025 m we yrminus1

(+012plusmn 012 m we yrminus1) in the eastern Karakoram Theoverall similarity between the mass budgets for surge-typeand non-surge-type glaciers shows that those flow instabili-ties seem not to affect the glacier-wide mass balance at leastover short time periods here 10 yr

5 Discussion

51 SRTM penetration

Alternative estimates of SRTMC-bandpenetration for the PKHregion were obtained by Kaumlaumlb et al (2012) by comparing el-evations of the ICESat altimeter (Zwally et al 2002) with theSRTMC-bandDEM in PKH and extrapolating the 2003ndash2008trend of elevation changes back to the acquisition date ofSRTM The difference between east and west is more dis-tinct in our case with a mean penetration in the Karakoram of34 m (24plusmn 03 m in Kaumlaumlb et al 2012) 24 m in Bhutan and14 m around the Everest area (25plusmn 05 m for a wider areaincluding east Nepal and Bhutan in Kaumlaumlb et al 2012) Thiseastwest gradient is expected given that in February 2000when the SRTM mission was flown the snowfirn thicknesswas probably larger in the western PKH where glaciers areof winter-accumulation type than in the eastern PKH where

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1274 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 7Same as Fig 2 but for the Hindu Kush study site

glaciers are of summer-accumulation type Further studieswould be necessary to investigate the possible reasons be-hind the relatively low SRTMC-bandpenetration for the Pamir(18 m) that do not follow the eastwest gradient mentionedabove

Both our and Kaumlaumlb et alrsquos (2012) estimates agree withintheir error bars and seem thus to provide robust first-orderestimates for the SRTMC-bandradar penetration in the regionHowever it remains to be verified if a penetration depth com-puted at the regional scale is appropriate for a single smallglacier whose altitude distribution is different from the oneof whole study site (eg Abramov Glacier in the far north ofour Pamir study site)

52 Thinning over debris-covered ice

PKH glaciers are heavily debris-covered in their ablation ar-eas because of steep rock walls that surround them and in-tense avalanche activity Twelve percent (12 ) of the glacierarea in our study sites are covered with debris (up to 22 inthe Everest area Table 1) which is in agreement with pre-vious estimates for the Himalaya and the Karakoram 13 in Kaumlaumlb et al (2012) and 10 in Bolch et al (2012) Thedebris layer is expected to modify the ablation of the under-

lying ice It will increase ablation if its thickness is just a fewcentimeters (dominance of the albedo decrease) or decreaseablation it if the layer is thick enough to protect the ice fromsolar radiation (Mattson et al 1993 Mihalcea et al 2006Benn et al 2012 Lejeune et al 2013)

However recent studies have reported similar overall thin-ning rates over debris-covered and debris-free ice in thePKH (Kaumlaumlb et al 2012 Gardelle et al 2012a) or evenhigher lowering rates over debris-covered parts in the Ever-est area (Nuimura et al 2012) Our findings are consistentwith Nuimura et al (2012) for the Everest study site with astronger thinning (rate of elevation change ofminus097 m yrminus1)

in debris-covered areas than over clean ice (rate of elevationchange ofminus057 m yrminus1) In the Pamir Karakoram and SpitiLahaul study sites these rates are similar while in the HinduKush West Nepal Bhutan and Hengduan Shan study sitesthinning is greater over clean ice in agreement with the com-monly assumed protective effect of debris (Fig 4) Thosedifferences between our 9 study sites suggest a complex re-lationship between debris cover and glacier wastage

As local glacier thickness changes are a result of bothmass balance processes and a dynamic component theseunexpected high thinning rates over debris-covered parts

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1275

Fig 8Same as Fig 2 but for the Karakoram east study site Surge-type glaciers in an active surge phase (resp quiescent phase) are identifiedwith a solid triangle (resp solid circle)

are potentially the result of glacier-wide processes Over aglacier tongue the ablation can be enhanced by the pres-ence of steep exposed ice cliffs or supraglacial lakes andponds (Sakai et al 2000 2002) In addition it is also pos-sible that the glacier tongues are mostly covered by thindebris layers such that the albedo effect dominates the in-sulating effect resulting in an increased tongue-wide abla-tion The prevalence of thin debris was recently suggestedfor debris-covered glaciers around the Mount Gongga in thesouth-eastern Tibetan Plateau (Zhang et al 2013b) Finallyice-flow rates could be very low at debris-covered parts asshown by Quincey et al (2007) in the Everest area Kaumlaumlb(2005) for Bhutan and Scherler et al (2011b) in the HinduKush-Karakoram-Himalaya Thus all three factors are likelyto increase the overall thinning under debris despite the un-doubted protective effect of a continuous thick debris coverat a local scale Future investigations combining surface ve-locity fields and maps of elevation changes could allow eval-uating more closely the relative role of ablation and ice fluxesin thinning rates over PKH glacier tongues (eg Berthierand Vincent 2012 Nuth et al 2012) Measurements of the

spatial distribution of debris thickness over the PKH glaciertongues are also needed

Note that in the case of debris-covered tongues the ele-vation change measurement represents the glacier thicknesschange but also the possible debris thickness evolution Thelatter remains poorly known in the PKH or in any othermountain range But based on the debris discharges givenby Bishop et al (1995) for Batura Glacier (Pakistan) 48to 90times 103 m3 yrminus1 the thickening of the debris can be es-timated between 27 and 55 mm yrminus1 which is negligiblegiven the magnitude of the glacier elevation changes

53 Comparison to previous mass balance estimates

Throughout the PKH there is only a single peer-reviewedglaciological record (on Chhota Shigri Glacier) coveringmost of our study period (Wagnon et al 2007 Vincent etal 2013) Our geodetic mass balance estimate for the SpitiLahaul study site has been recently compared to those glacio-logical measurements on Chhota Shigri Glacier (Vincent etal 2013) Mass balance records reported in Yao et al (2012)for the Tibetan Plateau and the surroundings mountain ranges

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1276 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 9Same as Fig 8 but for the Karakoram west study site

only start during glaciological year 20052006 This is whyhere we do not compare our mass balance estimates toglaciological records and limit our comparison to others re-gional estimates obtained using the geodetic or gravimetricmethods

We stress that the comparison of our mass balance mea-surements to published estimates is complicated by thefact that generally they do not cover the same time pe-riod Little is known about the inter-annual variability inthe PKH during the 21st century due to the limited num-ber of long glaciological time series The 9 yr mass balance

record on Chhota Shigri Glacier reveals a relatively highinter-annual variability of the annual mass balance (stan-dard deviationplusmn 074 m we yrminus1 during 2003ndash2011 Vin-cent et al 2013) which is similar to the one obtained be-tween 1968 and 1998 on Abramov Glacier (standard devi-ation 069 m we yrminus1 WGMS 2012) Thus inter-annualvariability can explain part of the difference between two cu-mulative mass balance estimates over 5ndash10 yr even if they areoffset by only one or two years

Based on ICESat repeat track measurements and theSRTM DEM Kaumlaumlb et al (2012) measured a region-wide

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1277

Fig 10Same as Fig 8 but for the Pamir study site

mass balance ofminus021plusmn 005 m we yrminus1 between 2003and 2008 for Hindu Kush-Karakoram-Himalaya glaciersFor that same area (ie excluding the Pamir and theHengduan Shan study sites) we found a mass balance ofminus015plusmn 007 m we yrminus1 Thus our result agrees with Kaumlaumlbet al (2012) within the error bars although our study periodis longer (1999ndash2011) and our measurement principle andextrapolation approach are different In addition we havecalculated updated mass balances from Kaumlaumlb et al (2012)ie averages over 3times 3 degree cells centered over our studysites and find that they are consistent within error barswith our mass balance estimates (Table 5) Similar resultsare found when our mass balances are compared to a spa-tially extended ICESat analysis for 2003ndash2009 (Gardner etal 2013) Mass losses measured with ICESat (Kaumlaumlb et al2012 Gardner et al 2013) tend to be slightly larger thanours This could be due to the difference in time periodsAlso our smaller mass losses could be partly explained if wesystematically underestimated the poorly constrained pene-

tration of the C-band andor X-band radar signal into ice andsnow

At a more local scale Bolch et al (2011) reported a massbalance ofminus079plusmn 052 m yrminus1 between 2002 and 2007 byDEM differencing over a glacier area of 62 km2 includingten glaciers south and west of Mt Everest For these sameglaciers Nuimura et al (2012) found a mean mass balanceof minus045plusmn 060 m yrminus1 during 2000ndash2008 while they com-puted a mass balance ofminus040plusmn 025 m yrminus1 between 1992and 2008 over a glacier area of 183 km2 around Mt EverestIn the present study we found an average mass balance ofminus041plusmn 021 m yrminus1 over the ten glaciers mentioned above(Table 5 Fig S1) and a value ofminus026plusmn 013 m yrminus1 overthe whole Everest study site between 2000 and 2011 Ourvalues are comparable to those of Nuimura et al (2012)Our mass balance is less negative than the 2002ndash2007 massbalance of Bolch et al (2011) but not statistically differentwhen error bars are considered (Fig 12) Indeed Bolch etal (2011) acknowledged some high uncertainties for their

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1278 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Table 4Glacier mass budget of the nine sub-regions to which the results of the study sites (mass balances) have been extrapolated For eachof them the total glacierized area as well as the percentage of the glacier area over which the mass balance was computed (ie study sites)are given The total glacier area is derived from the RGI v20 (Arendt et al 2012) The error for the mass balance in each sub-region is theroot of sum of squares (RSS) of the mass balance error of each study site and the regional extrapolation error The error for the mass budgetin each sub-region includes additionally the error from the total glacier area in the RGI v20

Sub-region Total glacier Measured glacier Mass balance Mass budgetarea (km2) area () (m we yrminus1) (Gt yrminus1)

Hengduan Shan 11584 12 minus033plusmn 014 minus38plusmn 38Bhutan 4021 34 minus022plusmn 013 minus09plusmn 05Everest 6226 23 minus026plusmn 014 minus16plusmn 08West Nepal 6849 13 minus032plusmn 014 minus22plusmn 10Spiti Lahaul 9043 23 minus045plusmn 014 minus41plusmn 13Hindu Kush 6135 13 minus012plusmn 016 minus07plusmn 10Karakoram East

1902428 +011 plusmn 014 +19 plusmn 31

Karakoram West 30 +009 plusmn 018Pamir 9369 34 +014 plusmn 014 +13 plusmn 13

TotalArea- 72251 30 minus014plusmn 008 minus101plusmn 55weighted average

Table 5Comparison of geodetic mass balances between this study and previous published results with overlapping study periods and similargeographic locations Updated figures from Kaumlaumlb et al (2012) are averaged over 3 times 3 cells centered over the study sites of the presentstudy

GlacierSite name This study Bolch et al (2011) Nuimura et al (2012) Kaumlaumlb et al (2012 updated)2002ndash2007 2000ndash2008 2003ndash2008

Study Mass balance Area (km2)a Mass balance Area Mass balance Mass balanceperiod (m we yrminus1) (m we yrminus1) (km2) (m we yrminus1) (m we yrminus1)

Hengduan San 1999ndash2010 minus022 plusmn 014 1303 (55 )Bhutan 1999ndash2010 minus022 plusmn 012 1384 (64 ) minus052 plusmn 016b

Everest

1999ndash2011

minus026 plusmn 013 1461 (58 ) minus039 plusmn 011AX010 minus090plusmn 034 04Changri SharNup minus042plusmn 017 161 (79 ) minus029plusmn 052 130 minus055plusmn 038Khumbu minus051plusmn 019 203 (47 ) minus045plusmn 052 170 minus076plusmn 052Nuptse minus037plusmn 020 50 (74 ) minus040plusmn 053 40 minus034plusmn 027Lhotse Nup minus021plusmn 027 24 (70 ) minus103plusmn 051 19 minus022plusmn 047Lhotse minus043plusmn 018 85 (83 ) minus110plusmn 052 65 minus067plusmn 051Lhotse SharImja minus070plusmn 052 98 (44 ) minus145plusmn 052 107 minus093plusmn 060Amphu Laptse minus046plusmn 034 25 (38 ) minus077plusmn 052 15 minus018plusmn 094Chukhung +044 plusmn 024 42 (47 ) +004 plusmn 054 38 +043 plusmn 081Ama Dablam minus049plusmn 017 36 (54 ) minus056plusmn 052 22 minus056plusmn 073Duwo minus016plusmn 026 19 (18 ) minus196plusmn 053 10 minus068plusmn 074Total Khumbu minus041plusmn 021 744 minus079plusmn 052 617 minus045plusmn 060(10 Glaciers above)

West Nepal 1999ndash2011 minus032 plusmn 013 908 (40 ) minus032 plusmn 012Spiti Lahaul 1999ndash2011 minus045 plusmn 013 2110 (46 ) minus038 plusmn 006Karakoram East 1999ndash2010 +011 plusmn 014 5328 (42 ) minus004 plusmn 004Karakoram West 1999ndash2008 +009 plusmn 018 5434 (45 )Hindu Kush 1999ndash2008 minus012 plusmn 016 793 (80 ) minus020 plusmn 006Pamir 1999ndash2011 +014 plusmn 013 3178 (50 )

a The total glacier area is given in km2 and in parenthesis the of the glacier area actually covered with measurementsb For this cell the ICESat coverage is insufficient and does not sample all glacier elevations

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1279

Fig 11 Elevation changes (SPOT5-SRTM) in the ablation area for clean ice (blue circles) and debris-covered ice (black circles) for eachstudy site For the Karakoram (east and west) and the Pamir study sites only non-surge-type glaciers are considered Standard error of theelevation differences are typically less thanplusmn2 m with each 100 m elevation band (not shown for the sake of clarity) Note elevation changesover separate sections of a glacier cannot be treated as mass changes due to the disregard of glacier dynamics

estimate over this short time period The differences in sur-vey periods and in the glacier outlines may also explainpart of these discrepancies In particular the delimitation ofthe accumulation area is notoriously difficult and Bolch etal (2011) did not include some of the steep high altitudeslopes which we included due to their potential avalanchecontribution For the 10 glaciers mentioned above our massbalances would become 012 m we yrminus1 more negative ifwe used the exact same outlines as Bolch et al (2011)Our positive mass balance for the small Chukhung Glacier(+044plusmn 024 m we yrminus1) is in agreement with Nuimura etal (2012) but enigmatic in a context of regional glacier lossand deserves further investigation If the 2002ndash2007 massbalance estimate by Bolch et al (2011) is excluded all otherrecent mass balance estimates suggest no acceleration of themass loss in the Everest area when compared to the long-term(1970ndash2007) mass balance measured by Bolch et al (2011)in this regionminus032plusmn 008 m weyrminus1

In Bhutan our results indicate a stronger mass loss forsouthern glaciers than for northern ones Interestingly in thisarea Kaumlaumlb (2005) measured high speeds (up to 200 m yrminus1)

for glaciers north of the Himalayan main ridge and ratherlow velocities over southern glacier tongues by using repeat

ASTER data This is consistent with the more negative massbalance that we report for the southern glaciers in Bhutanie for the slow flowing presumably downwasting glaciers

Berthier et al (2007) reported a mass balance betweenminus069 andminus085 m we yrminus1 for an ice-covered area of915 km2 around Chhota Shigri Glacier between 1999 (SRTMDEM) and 2004 (SPOT5-HRG DEM) For the same areawe found a mass balance ofminus045plusmn 013 m we yrminus1 over1999ndash2011 The difference in the study period may explainpart of the differences Also the methodology by Berthieret al (2007) did not explicitly take into account the SRTMradar penetration over glaciers and also corrected an appar-ent elevation-dependent bias according to glacier elevationswhich is now known as a resolution issue and is best cor-rected according to the terrain maximum curvature (Gardelleet al 2012b) More details and discussions about these dif-ferences can be found in Vincent et al (2013)

Importantly the results of the eastern Karakoram studysite confirm the balanced or slightly positive mass budgetreported by Gardelle et al (2012a) in the western Karako-ram over the past decade The two Karakoram study sitesare adjacent with a small overlapping area (sim 1100 km2 ofglaciers) where elevation differences agree within 1 m Both

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1280 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 12 Comparison of geodetic mass balances for 10 glaciers inthe Mount Everest area 1999ndash2010 mass balances measured inthe present study are compared with published estimates during2002ndash2007 (Bolch et al 2011) and during 2000ndash2008 (Nuimuraet al 2012) The 1-to-1 line is drawn The mean difference (plusmn1standard deviation) between (i) our values and Bolch et al (2011)are (047plusmn 058 m we yrminus1) and (ii) our values and Nuimura etal (2012) are (012plusmn 021 m we yrminus1) Part of the differences be-tween estimates may be due to the different periods surveyed andthe different glacier outlines used

sites show similar mass balances over slightly different pe-riods +011plusmn 014 m yrminus1 over 1999ndash2010 in the east and+009plusmn 018 m yrminus1 over 1999ndash2008 in the west This con-firms a ldquoKarakoram anomalyrdquo that was first suggested basedon field observations (Hewitt 2005) We report a mass bal-ance of+010plusmn 013 m yrminus1 for Baltoro Glacier (see loca-tion in Fig 8 size= 304 km2) between 1999 and 2010which is consistent with its gradual speed-up reported since2003 (Quincey et al 2009) However given the completelydifferent methods used in our and the latter study we can-not conclude that this demonstrates a real shift from nega-tive to positive mass balance The mass budget of SiachenGlacier (1145 km2 sometimes referred to as the highestbattlefield in the world) is also balanced or slightly pos-itive (+014plusmn 014 m we yrminus1) Thus the alarming state-ment by Rasul (2012) that this glacier retreated by 59 kmand lost 17 of its mass during 1989ndash2009 is very likelyerroneous Again within error bars our regional mass bal-ance in the Karakoram agrees with the mass balance ofminus003plusmn 004 m yrminus1 found by Kaumlaumlb et al (2012) over 2003ndash2008 and with the mass balance ofminus010plusmn 018 m we yrminus1

(2-sigma error level) obtained by Gardner et al (2013) for aregion including both the Karakoram and the Hindu Kush

In the northernmost part of the Pamir in the Alay moun-tain range the mass balance of Abramov Glacier has beenmeasured in the field for 30 yr between 1968 and 1998(WGMS 2012) with a mean value ofminus046 m yrminus1 Herewe report a mass balance ofminus003plusmn 014 m yrminus1 for this

glacier over 1999ndash2011 Our near zero mass budget estimatefor this glacier disagrees with negative mass balances recon-structed for the same period with a model run using biascorrected NCEPNCAR re-analysis data and calibrated withfield data (Barundun et al 2013) An underestimate of ourSRTM C-Band penetration in the Pamir as a whole and inparticular for the small Abramov Glacier may contribute tothese differences

For the whole Pamir study site our mass budget is bal-anced or slightly positive (+014plusmn 013 m we yrminus1) In theeastern Pamir the mass balance of Muztag Ata Glacier was+025 m we yrminus1 between 2005 and 2010 (Yao et al 2012)Although Muztag Ata Glacier is located outside of our Pamirstudy site and his glaciological records only cover the sec-ond half of our study period this measurement is consistentwith our remotely sensed mass balance Thus the ldquoKarako-ram anomalyrdquo should probably be renamed the ldquoPamir-Karakoram anomalyrdquo

This slightly positive or balanced mass budget measuredin the Pamir seems to contradict previous findings by Heidand Kaumlaumlb (2012) They reported rapidly decreasing glacierspeeds in this region between two snapshots of annual ve-locities in 20002001 and 20092010 an observation thatwould fit with negative mass balances (Span and Kuhn 2003Azam et al 2012) We note though that the relationship be-tween glacier mass balance and ice fluxes is not straightfor-ward and might in particular involve a time delay betweena change in mass balance and the related dynamic response(Heid and Kaumlaumlb 2012) Thus the decrease in speed couldwell be due to negative mass balances before or partiallybefore 1999ndash2011 We acknowledge that this discrepancyneeds to be investigated further in particular by examiningcontinuous changes in annual glacier speed through the firstdecade of the 21st Century Indeed Heid and Kaumlaumlb (2012)measured differences between two annual periods which maynot represent mean decadal velocity changes

Jacob et al (2012) using satellite gravimetry (GRACE)measured a mass budget ofminus5plusmn 6 Gt yrminus1 over 2003ndash2010for the ldquoKarakoram-Himalayardquo their region 8c which cor-responds to our study area excluding the Pamir Karakoramand Hindu Kush sub-regions but including some glaciersnorth of the Himalayan ridge already on the Tibetan PlateauFor this subset of our PKH study area we measured amass budget ofminus126plusmn 42 Gt yrminus1 The two estimates over-lap within their error bars especially if our error is dou-bled to match the two standard error level used by Jacob etal (2012) The estimate of Jacob et al (2012) over our com-plete study region (PKH) ieminus6plusmn 8 Gt yrminus1 (sum of theirregions 8b and 8c) also includes mass changes for the KunlunShan sub-region for which we did not measure glacier massbalances Therefore the comparison with our PKH value(minus101plusmn 55 Gt yrminus1) is not straightforward but suggests areasonable agreement especially if we consider the slightmass gain (15plusmn 17 Gt yrminus1) measured with ICESat in theWest Kunlun (Gardner et al 2013)

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1281

Table 6 Annual average of the glacier seasonal and decadal mass loss contributions to Indus Ganges and Brahmaputra discharges Basinarea and seasonal contributions are given by Kaser et al (2010) glacier area in each basin is derived from the RGI v20 (Arendt et al 2012)Numbers in parentheses indicate the percentage of glacierized area within the river basin Seasonal contributions were modeled by Kaser etal (2010) based on the CRU 20 dataset over 1961ndash1990 assuming a balanced glacier mass budget Glacier decadal mass loss contributionsare given as annual average over 1999ndash2011

River Basin area Glacier area Glacier contribution Mean discharge(km2) (km2) (m3 sminus1) (m3 sminus1)

Seasonally delayed Decadal mass lossKaser et al (2010) (this study)

Indus 1 139 814 25 598 (2 ) 105 103plusmn 72 2146 (Chevallier personalcommunication 2012)

Ganges 1 023 609 11 168 (1 ) 47 103plusmn 49 11739 Papa et al (2012Brahmaputra 527 666 15 296 (3 ) 33 147plusmn 64 19710 updated)

54 Reasons for the heterogeneous pattern of massbalance

The PKH glacier mass balance is slightly negative(minus014plusmn 008 m we yrminus1) but changes are not homoge-neous from one sub-region to another and seem to coincidewith the climatic settings of the mountain range For the east-ern sites (from the Hengduan Shan to West Nepal) which areunder the influence of the Indian and south-east Asian sum-mer monsoons the mass balances are moderately negative(sim minus026 m we yrminus1) In the northwest of the PKH at thePamir and Karakoram sites where glacier climate is char-acterized by the westerlies in winter mass balances are bal-anced or slightly positive (sim +010 m we yrminus1) In betweenthese two influences the glaciers of the Spiti Lahaul sitein a monsoon-arid transition zone experienced the strongestmass loss (minus045plusmn 013 m we yrminus1)

This heterogeneous pattern of mass balances seems to berelated to trends in precipitation throughout the PKH Archerand Fowler (2004) reported an increase in winter precipi-tation over the Upper Indus Basin since 1961 a trend con-firmed by Yao et al (2012) over the Pamir and the Karako-ram based on the analysis of the Global Precipitation Cli-matology Project data set (GPCP Adler et al 2003) Thisprecipitation increase is a source of greater accumulation forglaciers and thus a possible explanation for their slightly pos-itive mass budgets In addition in the Upper Indus Basin themean summer temperature has been decreasing since 1961(Fowler and Archer 2006) which is consistent with the find-ings of Shekhar et al (2010) in the Karakoram who reporteda decrease in both maximum and minimum temperature since1988 These increases in winter precipitation and summertemperature likely translate in greater accumulation and re-duced ablation changes which are consistent with the bal-anced or slightly positive glacier mass budget

On the other hand in the east of the PKH the Indiansummer monsoon has been weakening since the 1950s (Bol-lasina et al 2011) which could have reduced the amount

of snowfall over the eastern study sites and thus resultedin negative mass balances In addition maximum and min-imum temperatures increased since 1988 in the western Hi-malaya (Shekhar et al 2010) and maximum temperaturesincreased in the central Himalaya (Nepal) since the mid-1970s (Shrestha et al 1999) Thus both precipitation andtemperature trends in the Himalaya likely contributed to thenegative mass balances measured over the correspondingstudy sites (from Spiti Lahaul to Hengduan Shan Fig 1)

However gridded data sets such as GPCP are not neces-sarily representative of the climate at high altitude IndeedHewitt (2005) reported a 5-to-10-fold increase in precipi-tation between the glacier front and the accumulation areain the Karakoram that is not captured by reanalysis dataas recently confirmed by Immerzeel et al (2012) Thus di-rect measurements of accumulation on High Mountain Asiaglaciers (eg Azam et al 2013) and high-altitude weatherstations are eagerly needed to validate the large scale grid-ded data set (eg reanalysis) test their ability to describe thespecific climate near to PKH glaciers and assess the relativerole played by temperature and precipitation changes

55 Contribution to water resources

Our glacier mass balance can be averaged over large riverbasins to estimate the mean contribution to river runoffdue to a decade of glacier mass loss We assume therebyonly direct river runoff without loss terms For the threerivers for which we cover the entire glacierized area of theircatchments within our sub-regions (Fig 1) these contribu-tions to the river discharge are equal to 103 m3 sminus1 (Indus)103 m3 sminus1 (Ganges) and 147 m3 sminus1 (Brahmaputra) for theperiod 1999ndash2011 (Table 6)

Rivers of PKH being key water resources measures oftheir discharges are highly sensitive and difficult to access forpolitical reasons Papa et al (2012) built recent time series ofdischarges for Ganges and Brahmaputra by using altimetrymeasurements The locations of the corresponding dischargeestimates in Bangladesh are given in Fig 1 We can therefore

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1282 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

evaluate the relative contribution of glacier decadal mass loss(Table 6) to annual river run-off at 09 for the Gangesat Hardinge (basin area ofsim 1 020 000 km2) and 07 forBrahmaputra at Bahadurabad (basin area ofsim 530 000 km2)between 1999 and 2011 Given the discharge measured at theGuddu station by the Surface Water Hydrology Project of theWater and Power Development Authority (Pakistan see lo-cation on Fig 1) between 1999 and 2003 we can estimatethe contribution of glacier mass loss at 48 for the IndusRiver at this station

The seasonal contributions of glaciers to river run-off (iethe part of the annual precipitation that experiences season-ally delayed release from the glaciers) directly results fromseasonal variations of meteorological quantities in contrastto the glacier mass loss contribution which results from thelong-term climatic change Both runoff components directlyaffect the amount of water available in a river at a given lo-cation and at a given time so that their comparison could beof interest Even if the seasonal timing of the glacier massloss contribution is unclear it could in general peak at asimilar time of the year than the seasonal glacier contribu-tion so that both components rather enlarge each other thancancel Seasonal glacier contributions have been modeled byKaser et al (2010) The latter study assumed glaciers to bein equilibrium with the present climate and thus prescribedglacier mass balances equal to 0 Thus their seasonally de-layed glacier contributions do not include the part due todecadal glacier mass loss and is complementary to our es-timates (Table 6) For the eastern basins under the influenceof the Indian summer monsoon (Ganges and Brahmaputra)the contribution of glacier mass loss is higher than the sea-sonal contribution In other words for the first decade of the21st century the mass loss of glaciers is on average moreimportant for water availability in those two rivers than theirseasonal water storage effect The situation is different for theIndus Basin where the glacier melt season is also the dry sea-son which results in a relatively higher seasonally delayedglacier contribution to the river discharge There the season-ally delayed water release from the glaciers and the decadalglacier mass loss contribute on average equally to the riverdischarge Both contributions are strongly (and equally) de-pendent on the location of the discharge measurement andwill be significantly larger (in relative term) in more heav-ily glacierized and smaller mountain catchments located inthe upper part of these major river basins (Bookhagen andBurbank 2010)

6 Conclusion

In this study we assessed the spatial pattern of glaciermass balance over the PKH by comparing the SRTM andSPOT5 DEMs over 9 well-distributed study sites between1999 and 2011 We found slightly positive or balanced massbudgets in the western part (Pamir Karakoram) moder-ate mass loss in the east (Nepal Bhutan Hengduan Shan)

and the most negative mass balance in the monsoon-aridtransition zone (Spiti Lahaul in the western Himalaya)By extrapolating these values to climatically homogeneoussub-regions we estimate the PKH overall mass balanceto have beenminus014plusmn 008 m we yrminus1 between 1999 and2011 which corresponds to a contribution to sea level riseof 0028plusmn 0015 mm yrminus1 This is about 4 of the globalglacier mass loss estimated by Gardner et al (2013) thoughover a slightly shorter time period (2003ndash2009) The largestsource of uncertainties in those geodetic mass balance esti-mates is the poorly constrained penetration of the C-BandSRTM radar signal into snow and ice

Our technique of DEM differencing allows mapping in de-tail the spatial pattern of glacier elevation changes Thosemaps reveal many surge-type behaviors both in the Pamirand the Karakoram It also appears that the glacier-wide massbalance does not seem to be considerably affected by themass transfer from surges over thesim 10 yr time period stud-ied Similar lowering rates over debris-covered and clean iceare observed for four study sites despite the commonly as-sumed insulating effect of debris covers This suggests forthe scale of entire glacier tongues a spatial mixture of re-duced ablation under intact thick debris covers and enhancedablation by supraglacial lakes or ice cliffs or due to thin de-bris layer and very low glacier velocities in ablation areasthat reduce the ice advection downstream

The regionally heterogeneous pattern of ice wastage is inbroad agreement with contrasted trends in large-scale pre-cipitation and temperature However glaciological (eg ac-cumulation) and meteorological records are needed at highaltitude in the vicinity of glaciers to evaluate more closelythe local influence of atmospheric variables on glacier massbalance and fully understand the reasons behind those con-trasted mass budgets in the PKH

Supplementary material related to this article isavailable online at httpwwwthe-cryospherenet712632013tc-7-1263-2013-supplementpdf

Acknowledgements We thank G Cogley A Gardner T NuimuraT Bolch P Wagnon M Barandun M Hoelzle M Huss and ananonymous referee for their constructive comments that led to aconsiderably improved manuscript We thank the Water and PowerDevelopment Authority (WAPDA) for their hydrological data(processed by P Chevallier) and F Papa for sharing his updatedrun-off data ahead of publication We thank T Bolch for sharinghis glacier outlines from the Everest area We thank the USGSfor allowing free access to their Landsat archive CGIAR-SCI forSRTM C-band data DLR for SRTM X-band data and GLIMS forglacier inventories J Gardelle acknowledges a PhD fellowshipfrom the French Space Agency (CNES) and the French NationalResearch Center (CNRS) E Berthier acknowledges support fromCNES through the TOSCA and ISIS proposal no 397 and fromthe Programme National de Teacuteleacutedeacutetection Spatiale E Berthieracknowledges M Masiokas and R Villalba for welcoming him

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1283

at the IANIGLA (Mendoza Argentina) Y Arnaud acknowledgessupport from French National Research Agency through ANR-09-CEP-005-01PAPRIKA A Kaumlaumlb acknowledges support fromESA through Glaciers_cci (400010177810I-AM) and from theEuropean Research Council under the European Unionrsquos SeventhFramework Programme (FP2007-2013)ERC Grant Agreementn 320816

Edited by I M Howat

The publication of this articleis financed by CNRS-INSU

References

Adler R Huffman G Chang A Ferraro R Xie P JanowiakJ Rudolf B Schneider U Curtis S Bolvin D Gruber ASusskind J Arkin P and Nelkin E The Version-2 GlobalPrecipitation Climatology Project (GPCP) Monthly Precipita-tion Analysis (1979ndashPresent) J Hydrometeorol 4 1147ndash11672003

Archer D R and Fowler H J Spatial and temporal variationsin precipitation in the Upper Indus Basin global teleconnectionsand hydrological implications Hydrol Earth Syst Sci 8 47ndash61doi105194hess-8-47-2004 2004

Arendt A Bolch T Cogley J G Gardner A Hagen J-OHock R Kaser G Pfeffer W T Moholdt G Paul F RadicV Andreassen L Bajracharya S Beedle M Berthier EBhambri R Bliss A Brown I Burgess E Burgess DCawkwell F Chinn T Copland L Davies B de AngelisH Dolgova E Filbert K Forester R Fountain A Frey HGiffen B Glasser N Gurney S Hagg W Hall D Hari-tashya U K Hartmann G Helm C Herreid S Howat IKapustin G Khromova T Kienholz C Koenig M KohlerJ Kriegel D Kutuzov S Lavrentiev I LeBris R Lund JManley W Mayer C Miles E Li X Menounos B Mer-cer A Moelg N Mool P Nosenko G Negrete A NuthC Pettersson R Racoviteanu A Ranzi R Rastner P RauF Rich J Rott H Schneider C Seliverstov Y Sharp MSigurethsson O Stokes C Wheate R Winsvold S WolkenG Wyatt F and Zheltyhina N Randolph Glacier Inventory[v20] A Dataset of Global Glacier Outlines Global Land IceMeasurements from Space Boulder Colorado USA Digital Me-dia httpwwwglimsorgRGIrandolphhtml 2012

Azam M Wagnon P Ramanathan A Vincent C Sharma PArnaud Y Linda A Pottakkal J Chevallier P Singh V andBerthier E From balance to imbalance a shift in the dynamicbehaviour of Chhota Shigri glacier western Himalaya India JGlaciol 58 315ndash324 doi1031892012JoG11J123 2012

Azam M F Wagnon P Vincent C Ramanathan A Linda Aand Singh V B Reconstruction of the annual mass balanceof Chhota Shigri Glacier (Western Himalaya India) since 1969Ann Glaciol in revision 2013

Barrand N and Murray T Multivariate Controls on the In-cidence of Glacier Surging in the Karakoram Himalaya

Arct Antarct Alp Res 38 489ndash498 doi1016571523-0430(2006)38[489MCOTIO]20CO2 2006

Barundun M Huss M Azisov E Gafurov A HoelzleM Merkushkin A Salzmann N and Usubaliev R Re-establishing seasonal mass balance observation at AbramovGlacier Kyrgyzstan from 1968ndash2012 Abstract EGU2013-5668European Geophysical Union Vienna 2013

Benn D Bolch T Hands K Gulley J Luckman ANicholson L Quincey D Thompson S Toumi R andWiseman S Response of debris-covered glaciers in theMount Everest region to recent warming and implicationsfor outburst flood hazards Earth-Sci Rev 114 156ndash174doi101016jearscirev201203008 2012

Berthier E and Vincent C Relative contribution of surface mass-balance and ice-flux changes to the accelerated thinning of Merde Glace French Alps over 1979ndash2008 J Glaciol 58 501ndash512doi1031892012JoG11J083 2012

Berthier E Arnaud Y Baratoux D Vincent C and Reacutemy FRecent rapid thinning of the ldquo Mer de Glace rdquo glacier derivedfrom satellite optical images Geophys Res Lett 31 L17401doi1010292004GL020706 2004

Berthier E Arnaud Y Kumar R Ahmad S Wagnon P andChevallier P Remote sensing estimates of glacier mass balancesin the Himachal Pradesh (Western Himalaya India) RemoteSens Environ 108 327ndash338 doi101016jrse2006110172007

Bhambri R Bolch T Kawishwar P Dobhal D P SrivastavaD and Pratap B Heterogeneity in Glacier response from 1973to 2011 in the Shyok valley Karakoram India The CryosphereDiscuss 6 3049ndash3078 doi105194tcd-6-3049-2012 2012

Bhutiyani M Mass-balance studies on Siachen Glacier in theNubra valley Karakoram Himalaya India J Glaciol 45 112ndash118 1999

Bishop M P Schroder J F and Ward J L SPOT multispec-tral analysis for producing supraglacial debris-load estimates forBatura glacier Pakistan Geocarto Int 10 81ndash90 1995

Bolch T Pieczonka T and Benn D I Multi-decadal mass lossof glaciers in the Everest area (Nepal Himalaya) derived fromstereo imagery The Cryosphere 5 349ndash358 doi105194tc-5-349-2011 2011

Bolch T Kulkarni A Kaumlaumlb A Huggel C Paul F Cogley JFrey H Kargel J Fujita K Scheel M Bajracharya S andStoffel M The state and fate of Himalayan Glaciers Science336 310ndash314 doi101126science1215828 2012

Bollasina M Ming Y and Ramaswamy V Anthropogenicaerosols and the weakening of the South Asian summer mon-soon Science 334 502ndash505 doi101126science12049942011

Bookhagen B and Burbank D Toward a complete Himalayan hy-drological budget spatiotemporal distribution of snowmelt andrainfall and their impact on river discharge J Geophys Res115 F03019 doi1010292009JF001426 2010

Bretherton C Widmann M Dymnikov V Wallace J and BladeacuteI The effective number of spatial degrees of freedom of a time-varying field J Climate 12 1990ndash2009 1999

Copland L Pope S Bishop M Shroder J Clendon PBush A Kamp U Seong Y and Owen L Glacier veloc-ities across the central Karakoram Ann Glaciol 50 41ndash49doi103189172756409789624229 2009

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

1272 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 4Same as Fig 2 but for the Everest study site

Fig 5Same as Fig 2 but for the West Nepal study site

44 Glacier surges

In the Pamir and the Karakoram the spatial pattern of ele-vation changes of many glaciers is typical of surge events orflow instabilities They can be identified because of their veryhigh thinning and thickening rates that can both reach up to16 m yrminus1 (Figs 8ndash10)

Most of them have already been reported to be surge-type glaciers based on their velocity (Kotlyakov et al 2008Quincey et al 2011 Copland et al 2009 Heid and Kaumlaumlb2012) morphology (Copland et al 2011 Barrand and Mur-ray 2006 Hewitt 2007) or elevation changes (Gardelle etal 2012a) Here we only identify surge-type glaciers for thesake of mass balance calculation and we do not attempt toprovide an exhaustive up-to-date surge inventory This is why

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1273

Fig 6 Same as Fig 2 but for the Spiti Lahaul study site The map of elevation changes contains many voids due to the presence ofthin clouds during the acquisition of the SPOT5 DEM The mass balance of the Chhota Shigri and Bara Shigri glaciers are respectivelyminus039plusmn 018 m we yrminus1 andminus048plusmn 018 m we yrminus1

the published surge inventories in the Pamir and the Karako-ram do not necessarily match the surge-type glaciers labeledon the elevation change maps as these refer only to our ob-servation period (1999ndash2011)

The mass balance of surge-type glaciers (respectivelynon-surge-type glaciers) is +019plusmn 022 m we yrminus1

(+012plusmn 011 m we yrminus1) in the Pamir+009plusmn 030 m we yrminus1 (+009plusmn 015 m we yrminus1) inthe western Karakoram and+010plusmn 025 m we yrminus1

(+012plusmn 012 m we yrminus1) in the eastern Karakoram Theoverall similarity between the mass budgets for surge-typeand non-surge-type glaciers shows that those flow instabili-ties seem not to affect the glacier-wide mass balance at leastover short time periods here 10 yr

5 Discussion

51 SRTM penetration

Alternative estimates of SRTMC-bandpenetration for the PKHregion were obtained by Kaumlaumlb et al (2012) by comparing el-evations of the ICESat altimeter (Zwally et al 2002) with theSRTMC-bandDEM in PKH and extrapolating the 2003ndash2008trend of elevation changes back to the acquisition date ofSRTM The difference between east and west is more dis-tinct in our case with a mean penetration in the Karakoram of34 m (24plusmn 03 m in Kaumlaumlb et al 2012) 24 m in Bhutan and14 m around the Everest area (25plusmn 05 m for a wider areaincluding east Nepal and Bhutan in Kaumlaumlb et al 2012) Thiseastwest gradient is expected given that in February 2000when the SRTM mission was flown the snowfirn thicknesswas probably larger in the western PKH where glaciers areof winter-accumulation type than in the eastern PKH where

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1274 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 7Same as Fig 2 but for the Hindu Kush study site

glaciers are of summer-accumulation type Further studieswould be necessary to investigate the possible reasons be-hind the relatively low SRTMC-bandpenetration for the Pamir(18 m) that do not follow the eastwest gradient mentionedabove

Both our and Kaumlaumlb et alrsquos (2012) estimates agree withintheir error bars and seem thus to provide robust first-orderestimates for the SRTMC-bandradar penetration in the regionHowever it remains to be verified if a penetration depth com-puted at the regional scale is appropriate for a single smallglacier whose altitude distribution is different from the oneof whole study site (eg Abramov Glacier in the far north ofour Pamir study site)

52 Thinning over debris-covered ice

PKH glaciers are heavily debris-covered in their ablation ar-eas because of steep rock walls that surround them and in-tense avalanche activity Twelve percent (12 ) of the glacierarea in our study sites are covered with debris (up to 22 inthe Everest area Table 1) which is in agreement with pre-vious estimates for the Himalaya and the Karakoram 13 in Kaumlaumlb et al (2012) and 10 in Bolch et al (2012) Thedebris layer is expected to modify the ablation of the under-

lying ice It will increase ablation if its thickness is just a fewcentimeters (dominance of the albedo decrease) or decreaseablation it if the layer is thick enough to protect the ice fromsolar radiation (Mattson et al 1993 Mihalcea et al 2006Benn et al 2012 Lejeune et al 2013)

However recent studies have reported similar overall thin-ning rates over debris-covered and debris-free ice in thePKH (Kaumlaumlb et al 2012 Gardelle et al 2012a) or evenhigher lowering rates over debris-covered parts in the Ever-est area (Nuimura et al 2012) Our findings are consistentwith Nuimura et al (2012) for the Everest study site with astronger thinning (rate of elevation change ofminus097 m yrminus1)

in debris-covered areas than over clean ice (rate of elevationchange ofminus057 m yrminus1) In the Pamir Karakoram and SpitiLahaul study sites these rates are similar while in the HinduKush West Nepal Bhutan and Hengduan Shan study sitesthinning is greater over clean ice in agreement with the com-monly assumed protective effect of debris (Fig 4) Thosedifferences between our 9 study sites suggest a complex re-lationship between debris cover and glacier wastage

As local glacier thickness changes are a result of bothmass balance processes and a dynamic component theseunexpected high thinning rates over debris-covered parts

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1275

Fig 8Same as Fig 2 but for the Karakoram east study site Surge-type glaciers in an active surge phase (resp quiescent phase) are identifiedwith a solid triangle (resp solid circle)

are potentially the result of glacier-wide processes Over aglacier tongue the ablation can be enhanced by the pres-ence of steep exposed ice cliffs or supraglacial lakes andponds (Sakai et al 2000 2002) In addition it is also pos-sible that the glacier tongues are mostly covered by thindebris layers such that the albedo effect dominates the in-sulating effect resulting in an increased tongue-wide abla-tion The prevalence of thin debris was recently suggestedfor debris-covered glaciers around the Mount Gongga in thesouth-eastern Tibetan Plateau (Zhang et al 2013b) Finallyice-flow rates could be very low at debris-covered parts asshown by Quincey et al (2007) in the Everest area Kaumlaumlb(2005) for Bhutan and Scherler et al (2011b) in the HinduKush-Karakoram-Himalaya Thus all three factors are likelyto increase the overall thinning under debris despite the un-doubted protective effect of a continuous thick debris coverat a local scale Future investigations combining surface ve-locity fields and maps of elevation changes could allow eval-uating more closely the relative role of ablation and ice fluxesin thinning rates over PKH glacier tongues (eg Berthierand Vincent 2012 Nuth et al 2012) Measurements of the

spatial distribution of debris thickness over the PKH glaciertongues are also needed

Note that in the case of debris-covered tongues the ele-vation change measurement represents the glacier thicknesschange but also the possible debris thickness evolution Thelatter remains poorly known in the PKH or in any othermountain range But based on the debris discharges givenby Bishop et al (1995) for Batura Glacier (Pakistan) 48to 90times 103 m3 yrminus1 the thickening of the debris can be es-timated between 27 and 55 mm yrminus1 which is negligiblegiven the magnitude of the glacier elevation changes

53 Comparison to previous mass balance estimates

Throughout the PKH there is only a single peer-reviewedglaciological record (on Chhota Shigri Glacier) coveringmost of our study period (Wagnon et al 2007 Vincent etal 2013) Our geodetic mass balance estimate for the SpitiLahaul study site has been recently compared to those glacio-logical measurements on Chhota Shigri Glacier (Vincent etal 2013) Mass balance records reported in Yao et al (2012)for the Tibetan Plateau and the surroundings mountain ranges

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1276 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 9Same as Fig 8 but for the Karakoram west study site

only start during glaciological year 20052006 This is whyhere we do not compare our mass balance estimates toglaciological records and limit our comparison to others re-gional estimates obtained using the geodetic or gravimetricmethods

We stress that the comparison of our mass balance mea-surements to published estimates is complicated by thefact that generally they do not cover the same time pe-riod Little is known about the inter-annual variability inthe PKH during the 21st century due to the limited num-ber of long glaciological time series The 9 yr mass balance

record on Chhota Shigri Glacier reveals a relatively highinter-annual variability of the annual mass balance (stan-dard deviationplusmn 074 m we yrminus1 during 2003ndash2011 Vin-cent et al 2013) which is similar to the one obtained be-tween 1968 and 1998 on Abramov Glacier (standard devi-ation 069 m we yrminus1 WGMS 2012) Thus inter-annualvariability can explain part of the difference between two cu-mulative mass balance estimates over 5ndash10 yr even if they areoffset by only one or two years

Based on ICESat repeat track measurements and theSRTM DEM Kaumlaumlb et al (2012) measured a region-wide

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1277

Fig 10Same as Fig 8 but for the Pamir study site

mass balance ofminus021plusmn 005 m we yrminus1 between 2003and 2008 for Hindu Kush-Karakoram-Himalaya glaciersFor that same area (ie excluding the Pamir and theHengduan Shan study sites) we found a mass balance ofminus015plusmn 007 m we yrminus1 Thus our result agrees with Kaumlaumlbet al (2012) within the error bars although our study periodis longer (1999ndash2011) and our measurement principle andextrapolation approach are different In addition we havecalculated updated mass balances from Kaumlaumlb et al (2012)ie averages over 3times 3 degree cells centered over our studysites and find that they are consistent within error barswith our mass balance estimates (Table 5) Similar resultsare found when our mass balances are compared to a spa-tially extended ICESat analysis for 2003ndash2009 (Gardner etal 2013) Mass losses measured with ICESat (Kaumlaumlb et al2012 Gardner et al 2013) tend to be slightly larger thanours This could be due to the difference in time periodsAlso our smaller mass losses could be partly explained if wesystematically underestimated the poorly constrained pene-

tration of the C-band andor X-band radar signal into ice andsnow

At a more local scale Bolch et al (2011) reported a massbalance ofminus079plusmn 052 m yrminus1 between 2002 and 2007 byDEM differencing over a glacier area of 62 km2 includingten glaciers south and west of Mt Everest For these sameglaciers Nuimura et al (2012) found a mean mass balanceof minus045plusmn 060 m yrminus1 during 2000ndash2008 while they com-puted a mass balance ofminus040plusmn 025 m yrminus1 between 1992and 2008 over a glacier area of 183 km2 around Mt EverestIn the present study we found an average mass balance ofminus041plusmn 021 m yrminus1 over the ten glaciers mentioned above(Table 5 Fig S1) and a value ofminus026plusmn 013 m yrminus1 overthe whole Everest study site between 2000 and 2011 Ourvalues are comparable to those of Nuimura et al (2012)Our mass balance is less negative than the 2002ndash2007 massbalance of Bolch et al (2011) but not statistically differentwhen error bars are considered (Fig 12) Indeed Bolch etal (2011) acknowledged some high uncertainties for their

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1278 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Table 4Glacier mass budget of the nine sub-regions to which the results of the study sites (mass balances) have been extrapolated For eachof them the total glacierized area as well as the percentage of the glacier area over which the mass balance was computed (ie study sites)are given The total glacier area is derived from the RGI v20 (Arendt et al 2012) The error for the mass balance in each sub-region is theroot of sum of squares (RSS) of the mass balance error of each study site and the regional extrapolation error The error for the mass budgetin each sub-region includes additionally the error from the total glacier area in the RGI v20

Sub-region Total glacier Measured glacier Mass balance Mass budgetarea (km2) area () (m we yrminus1) (Gt yrminus1)

Hengduan Shan 11584 12 minus033plusmn 014 minus38plusmn 38Bhutan 4021 34 minus022plusmn 013 minus09plusmn 05Everest 6226 23 minus026plusmn 014 minus16plusmn 08West Nepal 6849 13 minus032plusmn 014 minus22plusmn 10Spiti Lahaul 9043 23 minus045plusmn 014 minus41plusmn 13Hindu Kush 6135 13 minus012plusmn 016 minus07plusmn 10Karakoram East

1902428 +011 plusmn 014 +19 plusmn 31

Karakoram West 30 +009 plusmn 018Pamir 9369 34 +014 plusmn 014 +13 plusmn 13

TotalArea- 72251 30 minus014plusmn 008 minus101plusmn 55weighted average

Table 5Comparison of geodetic mass balances between this study and previous published results with overlapping study periods and similargeographic locations Updated figures from Kaumlaumlb et al (2012) are averaged over 3 times 3 cells centered over the study sites of the presentstudy

GlacierSite name This study Bolch et al (2011) Nuimura et al (2012) Kaumlaumlb et al (2012 updated)2002ndash2007 2000ndash2008 2003ndash2008

Study Mass balance Area (km2)a Mass balance Area Mass balance Mass balanceperiod (m we yrminus1) (m we yrminus1) (km2) (m we yrminus1) (m we yrminus1)

Hengduan San 1999ndash2010 minus022 plusmn 014 1303 (55 )Bhutan 1999ndash2010 minus022 plusmn 012 1384 (64 ) minus052 plusmn 016b

Everest

1999ndash2011

minus026 plusmn 013 1461 (58 ) minus039 plusmn 011AX010 minus090plusmn 034 04Changri SharNup minus042plusmn 017 161 (79 ) minus029plusmn 052 130 minus055plusmn 038Khumbu minus051plusmn 019 203 (47 ) minus045plusmn 052 170 minus076plusmn 052Nuptse minus037plusmn 020 50 (74 ) minus040plusmn 053 40 minus034plusmn 027Lhotse Nup minus021plusmn 027 24 (70 ) minus103plusmn 051 19 minus022plusmn 047Lhotse minus043plusmn 018 85 (83 ) minus110plusmn 052 65 minus067plusmn 051Lhotse SharImja minus070plusmn 052 98 (44 ) minus145plusmn 052 107 minus093plusmn 060Amphu Laptse minus046plusmn 034 25 (38 ) minus077plusmn 052 15 minus018plusmn 094Chukhung +044 plusmn 024 42 (47 ) +004 plusmn 054 38 +043 plusmn 081Ama Dablam minus049plusmn 017 36 (54 ) minus056plusmn 052 22 minus056plusmn 073Duwo minus016plusmn 026 19 (18 ) minus196plusmn 053 10 minus068plusmn 074Total Khumbu minus041plusmn 021 744 minus079plusmn 052 617 minus045plusmn 060(10 Glaciers above)

West Nepal 1999ndash2011 minus032 plusmn 013 908 (40 ) minus032 plusmn 012Spiti Lahaul 1999ndash2011 minus045 plusmn 013 2110 (46 ) minus038 plusmn 006Karakoram East 1999ndash2010 +011 plusmn 014 5328 (42 ) minus004 plusmn 004Karakoram West 1999ndash2008 +009 plusmn 018 5434 (45 )Hindu Kush 1999ndash2008 minus012 plusmn 016 793 (80 ) minus020 plusmn 006Pamir 1999ndash2011 +014 plusmn 013 3178 (50 )

a The total glacier area is given in km2 and in parenthesis the of the glacier area actually covered with measurementsb For this cell the ICESat coverage is insufficient and does not sample all glacier elevations

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1279

Fig 11 Elevation changes (SPOT5-SRTM) in the ablation area for clean ice (blue circles) and debris-covered ice (black circles) for eachstudy site For the Karakoram (east and west) and the Pamir study sites only non-surge-type glaciers are considered Standard error of theelevation differences are typically less thanplusmn2 m with each 100 m elevation band (not shown for the sake of clarity) Note elevation changesover separate sections of a glacier cannot be treated as mass changes due to the disregard of glacier dynamics

estimate over this short time period The differences in sur-vey periods and in the glacier outlines may also explainpart of these discrepancies In particular the delimitation ofthe accumulation area is notoriously difficult and Bolch etal (2011) did not include some of the steep high altitudeslopes which we included due to their potential avalanchecontribution For the 10 glaciers mentioned above our massbalances would become 012 m we yrminus1 more negative ifwe used the exact same outlines as Bolch et al (2011)Our positive mass balance for the small Chukhung Glacier(+044plusmn 024 m we yrminus1) is in agreement with Nuimura etal (2012) but enigmatic in a context of regional glacier lossand deserves further investigation If the 2002ndash2007 massbalance estimate by Bolch et al (2011) is excluded all otherrecent mass balance estimates suggest no acceleration of themass loss in the Everest area when compared to the long-term(1970ndash2007) mass balance measured by Bolch et al (2011)in this regionminus032plusmn 008 m weyrminus1

In Bhutan our results indicate a stronger mass loss forsouthern glaciers than for northern ones Interestingly in thisarea Kaumlaumlb (2005) measured high speeds (up to 200 m yrminus1)

for glaciers north of the Himalayan main ridge and ratherlow velocities over southern glacier tongues by using repeat

ASTER data This is consistent with the more negative massbalance that we report for the southern glaciers in Bhutanie for the slow flowing presumably downwasting glaciers

Berthier et al (2007) reported a mass balance betweenminus069 andminus085 m we yrminus1 for an ice-covered area of915 km2 around Chhota Shigri Glacier between 1999 (SRTMDEM) and 2004 (SPOT5-HRG DEM) For the same areawe found a mass balance ofminus045plusmn 013 m we yrminus1 over1999ndash2011 The difference in the study period may explainpart of the differences Also the methodology by Berthieret al (2007) did not explicitly take into account the SRTMradar penetration over glaciers and also corrected an appar-ent elevation-dependent bias according to glacier elevationswhich is now known as a resolution issue and is best cor-rected according to the terrain maximum curvature (Gardelleet al 2012b) More details and discussions about these dif-ferences can be found in Vincent et al (2013)

Importantly the results of the eastern Karakoram studysite confirm the balanced or slightly positive mass budgetreported by Gardelle et al (2012a) in the western Karako-ram over the past decade The two Karakoram study sitesare adjacent with a small overlapping area (sim 1100 km2 ofglaciers) where elevation differences agree within 1 m Both

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1280 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 12 Comparison of geodetic mass balances for 10 glaciers inthe Mount Everest area 1999ndash2010 mass balances measured inthe present study are compared with published estimates during2002ndash2007 (Bolch et al 2011) and during 2000ndash2008 (Nuimuraet al 2012) The 1-to-1 line is drawn The mean difference (plusmn1standard deviation) between (i) our values and Bolch et al (2011)are (047plusmn 058 m we yrminus1) and (ii) our values and Nuimura etal (2012) are (012plusmn 021 m we yrminus1) Part of the differences be-tween estimates may be due to the different periods surveyed andthe different glacier outlines used

sites show similar mass balances over slightly different pe-riods +011plusmn 014 m yrminus1 over 1999ndash2010 in the east and+009plusmn 018 m yrminus1 over 1999ndash2008 in the west This con-firms a ldquoKarakoram anomalyrdquo that was first suggested basedon field observations (Hewitt 2005) We report a mass bal-ance of+010plusmn 013 m yrminus1 for Baltoro Glacier (see loca-tion in Fig 8 size= 304 km2) between 1999 and 2010which is consistent with its gradual speed-up reported since2003 (Quincey et al 2009) However given the completelydifferent methods used in our and the latter study we can-not conclude that this demonstrates a real shift from nega-tive to positive mass balance The mass budget of SiachenGlacier (1145 km2 sometimes referred to as the highestbattlefield in the world) is also balanced or slightly pos-itive (+014plusmn 014 m we yrminus1) Thus the alarming state-ment by Rasul (2012) that this glacier retreated by 59 kmand lost 17 of its mass during 1989ndash2009 is very likelyerroneous Again within error bars our regional mass bal-ance in the Karakoram agrees with the mass balance ofminus003plusmn 004 m yrminus1 found by Kaumlaumlb et al (2012) over 2003ndash2008 and with the mass balance ofminus010plusmn 018 m we yrminus1

(2-sigma error level) obtained by Gardner et al (2013) for aregion including both the Karakoram and the Hindu Kush

In the northernmost part of the Pamir in the Alay moun-tain range the mass balance of Abramov Glacier has beenmeasured in the field for 30 yr between 1968 and 1998(WGMS 2012) with a mean value ofminus046 m yrminus1 Herewe report a mass balance ofminus003plusmn 014 m yrminus1 for this

glacier over 1999ndash2011 Our near zero mass budget estimatefor this glacier disagrees with negative mass balances recon-structed for the same period with a model run using biascorrected NCEPNCAR re-analysis data and calibrated withfield data (Barundun et al 2013) An underestimate of ourSRTM C-Band penetration in the Pamir as a whole and inparticular for the small Abramov Glacier may contribute tothese differences

For the whole Pamir study site our mass budget is bal-anced or slightly positive (+014plusmn 013 m we yrminus1) In theeastern Pamir the mass balance of Muztag Ata Glacier was+025 m we yrminus1 between 2005 and 2010 (Yao et al 2012)Although Muztag Ata Glacier is located outside of our Pamirstudy site and his glaciological records only cover the sec-ond half of our study period this measurement is consistentwith our remotely sensed mass balance Thus the ldquoKarako-ram anomalyrdquo should probably be renamed the ldquoPamir-Karakoram anomalyrdquo

This slightly positive or balanced mass budget measuredin the Pamir seems to contradict previous findings by Heidand Kaumlaumlb (2012) They reported rapidly decreasing glacierspeeds in this region between two snapshots of annual ve-locities in 20002001 and 20092010 an observation thatwould fit with negative mass balances (Span and Kuhn 2003Azam et al 2012) We note though that the relationship be-tween glacier mass balance and ice fluxes is not straightfor-ward and might in particular involve a time delay betweena change in mass balance and the related dynamic response(Heid and Kaumlaumlb 2012) Thus the decrease in speed couldwell be due to negative mass balances before or partiallybefore 1999ndash2011 We acknowledge that this discrepancyneeds to be investigated further in particular by examiningcontinuous changes in annual glacier speed through the firstdecade of the 21st Century Indeed Heid and Kaumlaumlb (2012)measured differences between two annual periods which maynot represent mean decadal velocity changes

Jacob et al (2012) using satellite gravimetry (GRACE)measured a mass budget ofminus5plusmn 6 Gt yrminus1 over 2003ndash2010for the ldquoKarakoram-Himalayardquo their region 8c which cor-responds to our study area excluding the Pamir Karakoramand Hindu Kush sub-regions but including some glaciersnorth of the Himalayan ridge already on the Tibetan PlateauFor this subset of our PKH study area we measured amass budget ofminus126plusmn 42 Gt yrminus1 The two estimates over-lap within their error bars especially if our error is dou-bled to match the two standard error level used by Jacob etal (2012) The estimate of Jacob et al (2012) over our com-plete study region (PKH) ieminus6plusmn 8 Gt yrminus1 (sum of theirregions 8b and 8c) also includes mass changes for the KunlunShan sub-region for which we did not measure glacier massbalances Therefore the comparison with our PKH value(minus101plusmn 55 Gt yrminus1) is not straightforward but suggests areasonable agreement especially if we consider the slightmass gain (15plusmn 17 Gt yrminus1) measured with ICESat in theWest Kunlun (Gardner et al 2013)

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1281

Table 6 Annual average of the glacier seasonal and decadal mass loss contributions to Indus Ganges and Brahmaputra discharges Basinarea and seasonal contributions are given by Kaser et al (2010) glacier area in each basin is derived from the RGI v20 (Arendt et al 2012)Numbers in parentheses indicate the percentage of glacierized area within the river basin Seasonal contributions were modeled by Kaser etal (2010) based on the CRU 20 dataset over 1961ndash1990 assuming a balanced glacier mass budget Glacier decadal mass loss contributionsare given as annual average over 1999ndash2011

River Basin area Glacier area Glacier contribution Mean discharge(km2) (km2) (m3 sminus1) (m3 sminus1)

Seasonally delayed Decadal mass lossKaser et al (2010) (this study)

Indus 1 139 814 25 598 (2 ) 105 103plusmn 72 2146 (Chevallier personalcommunication 2012)

Ganges 1 023 609 11 168 (1 ) 47 103plusmn 49 11739 Papa et al (2012Brahmaputra 527 666 15 296 (3 ) 33 147plusmn 64 19710 updated)

54 Reasons for the heterogeneous pattern of massbalance

The PKH glacier mass balance is slightly negative(minus014plusmn 008 m we yrminus1) but changes are not homoge-neous from one sub-region to another and seem to coincidewith the climatic settings of the mountain range For the east-ern sites (from the Hengduan Shan to West Nepal) which areunder the influence of the Indian and south-east Asian sum-mer monsoons the mass balances are moderately negative(sim minus026 m we yrminus1) In the northwest of the PKH at thePamir and Karakoram sites where glacier climate is char-acterized by the westerlies in winter mass balances are bal-anced or slightly positive (sim +010 m we yrminus1) In betweenthese two influences the glaciers of the Spiti Lahaul sitein a monsoon-arid transition zone experienced the strongestmass loss (minus045plusmn 013 m we yrminus1)

This heterogeneous pattern of mass balances seems to berelated to trends in precipitation throughout the PKH Archerand Fowler (2004) reported an increase in winter precipi-tation over the Upper Indus Basin since 1961 a trend con-firmed by Yao et al (2012) over the Pamir and the Karako-ram based on the analysis of the Global Precipitation Cli-matology Project data set (GPCP Adler et al 2003) Thisprecipitation increase is a source of greater accumulation forglaciers and thus a possible explanation for their slightly pos-itive mass budgets In addition in the Upper Indus Basin themean summer temperature has been decreasing since 1961(Fowler and Archer 2006) which is consistent with the find-ings of Shekhar et al (2010) in the Karakoram who reporteda decrease in both maximum and minimum temperature since1988 These increases in winter precipitation and summertemperature likely translate in greater accumulation and re-duced ablation changes which are consistent with the bal-anced or slightly positive glacier mass budget

On the other hand in the east of the PKH the Indiansummer monsoon has been weakening since the 1950s (Bol-lasina et al 2011) which could have reduced the amount

of snowfall over the eastern study sites and thus resultedin negative mass balances In addition maximum and min-imum temperatures increased since 1988 in the western Hi-malaya (Shekhar et al 2010) and maximum temperaturesincreased in the central Himalaya (Nepal) since the mid-1970s (Shrestha et al 1999) Thus both precipitation andtemperature trends in the Himalaya likely contributed to thenegative mass balances measured over the correspondingstudy sites (from Spiti Lahaul to Hengduan Shan Fig 1)

However gridded data sets such as GPCP are not neces-sarily representative of the climate at high altitude IndeedHewitt (2005) reported a 5-to-10-fold increase in precipi-tation between the glacier front and the accumulation areain the Karakoram that is not captured by reanalysis dataas recently confirmed by Immerzeel et al (2012) Thus di-rect measurements of accumulation on High Mountain Asiaglaciers (eg Azam et al 2013) and high-altitude weatherstations are eagerly needed to validate the large scale grid-ded data set (eg reanalysis) test their ability to describe thespecific climate near to PKH glaciers and assess the relativerole played by temperature and precipitation changes

55 Contribution to water resources

Our glacier mass balance can be averaged over large riverbasins to estimate the mean contribution to river runoffdue to a decade of glacier mass loss We assume therebyonly direct river runoff without loss terms For the threerivers for which we cover the entire glacierized area of theircatchments within our sub-regions (Fig 1) these contribu-tions to the river discharge are equal to 103 m3 sminus1 (Indus)103 m3 sminus1 (Ganges) and 147 m3 sminus1 (Brahmaputra) for theperiod 1999ndash2011 (Table 6)

Rivers of PKH being key water resources measures oftheir discharges are highly sensitive and difficult to access forpolitical reasons Papa et al (2012) built recent time series ofdischarges for Ganges and Brahmaputra by using altimetrymeasurements The locations of the corresponding dischargeestimates in Bangladesh are given in Fig 1 We can therefore

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1282 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

evaluate the relative contribution of glacier decadal mass loss(Table 6) to annual river run-off at 09 for the Gangesat Hardinge (basin area ofsim 1 020 000 km2) and 07 forBrahmaputra at Bahadurabad (basin area ofsim 530 000 km2)between 1999 and 2011 Given the discharge measured at theGuddu station by the Surface Water Hydrology Project of theWater and Power Development Authority (Pakistan see lo-cation on Fig 1) between 1999 and 2003 we can estimatethe contribution of glacier mass loss at 48 for the IndusRiver at this station

The seasonal contributions of glaciers to river run-off (iethe part of the annual precipitation that experiences season-ally delayed release from the glaciers) directly results fromseasonal variations of meteorological quantities in contrastto the glacier mass loss contribution which results from thelong-term climatic change Both runoff components directlyaffect the amount of water available in a river at a given lo-cation and at a given time so that their comparison could beof interest Even if the seasonal timing of the glacier massloss contribution is unclear it could in general peak at asimilar time of the year than the seasonal glacier contribu-tion so that both components rather enlarge each other thancancel Seasonal glacier contributions have been modeled byKaser et al (2010) The latter study assumed glaciers to bein equilibrium with the present climate and thus prescribedglacier mass balances equal to 0 Thus their seasonally de-layed glacier contributions do not include the part due todecadal glacier mass loss and is complementary to our es-timates (Table 6) For the eastern basins under the influenceof the Indian summer monsoon (Ganges and Brahmaputra)the contribution of glacier mass loss is higher than the sea-sonal contribution In other words for the first decade of the21st century the mass loss of glaciers is on average moreimportant for water availability in those two rivers than theirseasonal water storage effect The situation is different for theIndus Basin where the glacier melt season is also the dry sea-son which results in a relatively higher seasonally delayedglacier contribution to the river discharge There the season-ally delayed water release from the glaciers and the decadalglacier mass loss contribute on average equally to the riverdischarge Both contributions are strongly (and equally) de-pendent on the location of the discharge measurement andwill be significantly larger (in relative term) in more heav-ily glacierized and smaller mountain catchments located inthe upper part of these major river basins (Bookhagen andBurbank 2010)

6 Conclusion

In this study we assessed the spatial pattern of glaciermass balance over the PKH by comparing the SRTM andSPOT5 DEMs over 9 well-distributed study sites between1999 and 2011 We found slightly positive or balanced massbudgets in the western part (Pamir Karakoram) moder-ate mass loss in the east (Nepal Bhutan Hengduan Shan)

and the most negative mass balance in the monsoon-aridtransition zone (Spiti Lahaul in the western Himalaya)By extrapolating these values to climatically homogeneoussub-regions we estimate the PKH overall mass balanceto have beenminus014plusmn 008 m we yrminus1 between 1999 and2011 which corresponds to a contribution to sea level riseof 0028plusmn 0015 mm yrminus1 This is about 4 of the globalglacier mass loss estimated by Gardner et al (2013) thoughover a slightly shorter time period (2003ndash2009) The largestsource of uncertainties in those geodetic mass balance esti-mates is the poorly constrained penetration of the C-BandSRTM radar signal into snow and ice

Our technique of DEM differencing allows mapping in de-tail the spatial pattern of glacier elevation changes Thosemaps reveal many surge-type behaviors both in the Pamirand the Karakoram It also appears that the glacier-wide massbalance does not seem to be considerably affected by themass transfer from surges over thesim 10 yr time period stud-ied Similar lowering rates over debris-covered and clean iceare observed for four study sites despite the commonly as-sumed insulating effect of debris covers This suggests forthe scale of entire glacier tongues a spatial mixture of re-duced ablation under intact thick debris covers and enhancedablation by supraglacial lakes or ice cliffs or due to thin de-bris layer and very low glacier velocities in ablation areasthat reduce the ice advection downstream

The regionally heterogeneous pattern of ice wastage is inbroad agreement with contrasted trends in large-scale pre-cipitation and temperature However glaciological (eg ac-cumulation) and meteorological records are needed at highaltitude in the vicinity of glaciers to evaluate more closelythe local influence of atmospheric variables on glacier massbalance and fully understand the reasons behind those con-trasted mass budgets in the PKH

Supplementary material related to this article isavailable online at httpwwwthe-cryospherenet712632013tc-7-1263-2013-supplementpdf

Acknowledgements We thank G Cogley A Gardner T NuimuraT Bolch P Wagnon M Barandun M Hoelzle M Huss and ananonymous referee for their constructive comments that led to aconsiderably improved manuscript We thank the Water and PowerDevelopment Authority (WAPDA) for their hydrological data(processed by P Chevallier) and F Papa for sharing his updatedrun-off data ahead of publication We thank T Bolch for sharinghis glacier outlines from the Everest area We thank the USGSfor allowing free access to their Landsat archive CGIAR-SCI forSRTM C-band data DLR for SRTM X-band data and GLIMS forglacier inventories J Gardelle acknowledges a PhD fellowshipfrom the French Space Agency (CNES) and the French NationalResearch Center (CNRS) E Berthier acknowledges support fromCNES through the TOSCA and ISIS proposal no 397 and fromthe Programme National de Teacuteleacutedeacutetection Spatiale E Berthieracknowledges M Masiokas and R Villalba for welcoming him

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1283

at the IANIGLA (Mendoza Argentina) Y Arnaud acknowledgessupport from French National Research Agency through ANR-09-CEP-005-01PAPRIKA A Kaumlaumlb acknowledges support fromESA through Glaciers_cci (400010177810I-AM) and from theEuropean Research Council under the European Unionrsquos SeventhFramework Programme (FP2007-2013)ERC Grant Agreementn 320816

Edited by I M Howat

The publication of this articleis financed by CNRS-INSU

References

Adler R Huffman G Chang A Ferraro R Xie P JanowiakJ Rudolf B Schneider U Curtis S Bolvin D Gruber ASusskind J Arkin P and Nelkin E The Version-2 GlobalPrecipitation Climatology Project (GPCP) Monthly Precipita-tion Analysis (1979ndashPresent) J Hydrometeorol 4 1147ndash11672003

Archer D R and Fowler H J Spatial and temporal variationsin precipitation in the Upper Indus Basin global teleconnectionsand hydrological implications Hydrol Earth Syst Sci 8 47ndash61doi105194hess-8-47-2004 2004

Arendt A Bolch T Cogley J G Gardner A Hagen J-OHock R Kaser G Pfeffer W T Moholdt G Paul F RadicV Andreassen L Bajracharya S Beedle M Berthier EBhambri R Bliss A Brown I Burgess E Burgess DCawkwell F Chinn T Copland L Davies B de AngelisH Dolgova E Filbert K Forester R Fountain A Frey HGiffen B Glasser N Gurney S Hagg W Hall D Hari-tashya U K Hartmann G Helm C Herreid S Howat IKapustin G Khromova T Kienholz C Koenig M KohlerJ Kriegel D Kutuzov S Lavrentiev I LeBris R Lund JManley W Mayer C Miles E Li X Menounos B Mer-cer A Moelg N Mool P Nosenko G Negrete A NuthC Pettersson R Racoviteanu A Ranzi R Rastner P RauF Rich J Rott H Schneider C Seliverstov Y Sharp MSigurethsson O Stokes C Wheate R Winsvold S WolkenG Wyatt F and Zheltyhina N Randolph Glacier Inventory[v20] A Dataset of Global Glacier Outlines Global Land IceMeasurements from Space Boulder Colorado USA Digital Me-dia httpwwwglimsorgRGIrandolphhtml 2012

Azam M Wagnon P Ramanathan A Vincent C Sharma PArnaud Y Linda A Pottakkal J Chevallier P Singh V andBerthier E From balance to imbalance a shift in the dynamicbehaviour of Chhota Shigri glacier western Himalaya India JGlaciol 58 315ndash324 doi1031892012JoG11J123 2012

Azam M F Wagnon P Vincent C Ramanathan A Linda Aand Singh V B Reconstruction of the annual mass balanceof Chhota Shigri Glacier (Western Himalaya India) since 1969Ann Glaciol in revision 2013

Barrand N and Murray T Multivariate Controls on the In-cidence of Glacier Surging in the Karakoram Himalaya

Arct Antarct Alp Res 38 489ndash498 doi1016571523-0430(2006)38[489MCOTIO]20CO2 2006

Barundun M Huss M Azisov E Gafurov A HoelzleM Merkushkin A Salzmann N and Usubaliev R Re-establishing seasonal mass balance observation at AbramovGlacier Kyrgyzstan from 1968ndash2012 Abstract EGU2013-5668European Geophysical Union Vienna 2013

Benn D Bolch T Hands K Gulley J Luckman ANicholson L Quincey D Thompson S Toumi R andWiseman S Response of debris-covered glaciers in theMount Everest region to recent warming and implicationsfor outburst flood hazards Earth-Sci Rev 114 156ndash174doi101016jearscirev201203008 2012

Berthier E and Vincent C Relative contribution of surface mass-balance and ice-flux changes to the accelerated thinning of Merde Glace French Alps over 1979ndash2008 J Glaciol 58 501ndash512doi1031892012JoG11J083 2012

Berthier E Arnaud Y Baratoux D Vincent C and Reacutemy FRecent rapid thinning of the ldquo Mer de Glace rdquo glacier derivedfrom satellite optical images Geophys Res Lett 31 L17401doi1010292004GL020706 2004

Berthier E Arnaud Y Kumar R Ahmad S Wagnon P andChevallier P Remote sensing estimates of glacier mass balancesin the Himachal Pradesh (Western Himalaya India) RemoteSens Environ 108 327ndash338 doi101016jrse2006110172007

Bhambri R Bolch T Kawishwar P Dobhal D P SrivastavaD and Pratap B Heterogeneity in Glacier response from 1973to 2011 in the Shyok valley Karakoram India The CryosphereDiscuss 6 3049ndash3078 doi105194tcd-6-3049-2012 2012

Bhutiyani M Mass-balance studies on Siachen Glacier in theNubra valley Karakoram Himalaya India J Glaciol 45 112ndash118 1999

Bishop M P Schroder J F and Ward J L SPOT multispec-tral analysis for producing supraglacial debris-load estimates forBatura glacier Pakistan Geocarto Int 10 81ndash90 1995

Bolch T Pieczonka T and Benn D I Multi-decadal mass lossof glaciers in the Everest area (Nepal Himalaya) derived fromstereo imagery The Cryosphere 5 349ndash358 doi105194tc-5-349-2011 2011

Bolch T Kulkarni A Kaumlaumlb A Huggel C Paul F Cogley JFrey H Kargel J Fujita K Scheel M Bajracharya S andStoffel M The state and fate of Himalayan Glaciers Science336 310ndash314 doi101126science1215828 2012

Bollasina M Ming Y and Ramaswamy V Anthropogenicaerosols and the weakening of the South Asian summer mon-soon Science 334 502ndash505 doi101126science12049942011

Bookhagen B and Burbank D Toward a complete Himalayan hy-drological budget spatiotemporal distribution of snowmelt andrainfall and their impact on river discharge J Geophys Res115 F03019 doi1010292009JF001426 2010

Bretherton C Widmann M Dymnikov V Wallace J and BladeacuteI The effective number of spatial degrees of freedom of a time-varying field J Climate 12 1990ndash2009 1999

Copland L Pope S Bishop M Shroder J Clendon PBush A Kamp U Seong Y and Owen L Glacier veloc-ities across the central Karakoram Ann Glaciol 50 41ndash49doi103189172756409789624229 2009

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1273

Fig 6 Same as Fig 2 but for the Spiti Lahaul study site The map of elevation changes contains many voids due to the presence ofthin clouds during the acquisition of the SPOT5 DEM The mass balance of the Chhota Shigri and Bara Shigri glaciers are respectivelyminus039plusmn 018 m we yrminus1 andminus048plusmn 018 m we yrminus1

the published surge inventories in the Pamir and the Karako-ram do not necessarily match the surge-type glaciers labeledon the elevation change maps as these refer only to our ob-servation period (1999ndash2011)

The mass balance of surge-type glaciers (respectivelynon-surge-type glaciers) is +019plusmn 022 m we yrminus1

(+012plusmn 011 m we yrminus1) in the Pamir+009plusmn 030 m we yrminus1 (+009plusmn 015 m we yrminus1) inthe western Karakoram and+010plusmn 025 m we yrminus1

(+012plusmn 012 m we yrminus1) in the eastern Karakoram Theoverall similarity between the mass budgets for surge-typeand non-surge-type glaciers shows that those flow instabili-ties seem not to affect the glacier-wide mass balance at leastover short time periods here 10 yr

5 Discussion

51 SRTM penetration

Alternative estimates of SRTMC-bandpenetration for the PKHregion were obtained by Kaumlaumlb et al (2012) by comparing el-evations of the ICESat altimeter (Zwally et al 2002) with theSRTMC-bandDEM in PKH and extrapolating the 2003ndash2008trend of elevation changes back to the acquisition date ofSRTM The difference between east and west is more dis-tinct in our case with a mean penetration in the Karakoram of34 m (24plusmn 03 m in Kaumlaumlb et al 2012) 24 m in Bhutan and14 m around the Everest area (25plusmn 05 m for a wider areaincluding east Nepal and Bhutan in Kaumlaumlb et al 2012) Thiseastwest gradient is expected given that in February 2000when the SRTM mission was flown the snowfirn thicknesswas probably larger in the western PKH where glaciers areof winter-accumulation type than in the eastern PKH where

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1274 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 7Same as Fig 2 but for the Hindu Kush study site

glaciers are of summer-accumulation type Further studieswould be necessary to investigate the possible reasons be-hind the relatively low SRTMC-bandpenetration for the Pamir(18 m) that do not follow the eastwest gradient mentionedabove

Both our and Kaumlaumlb et alrsquos (2012) estimates agree withintheir error bars and seem thus to provide robust first-orderestimates for the SRTMC-bandradar penetration in the regionHowever it remains to be verified if a penetration depth com-puted at the regional scale is appropriate for a single smallglacier whose altitude distribution is different from the oneof whole study site (eg Abramov Glacier in the far north ofour Pamir study site)

52 Thinning over debris-covered ice

PKH glaciers are heavily debris-covered in their ablation ar-eas because of steep rock walls that surround them and in-tense avalanche activity Twelve percent (12 ) of the glacierarea in our study sites are covered with debris (up to 22 inthe Everest area Table 1) which is in agreement with pre-vious estimates for the Himalaya and the Karakoram 13 in Kaumlaumlb et al (2012) and 10 in Bolch et al (2012) Thedebris layer is expected to modify the ablation of the under-

lying ice It will increase ablation if its thickness is just a fewcentimeters (dominance of the albedo decrease) or decreaseablation it if the layer is thick enough to protect the ice fromsolar radiation (Mattson et al 1993 Mihalcea et al 2006Benn et al 2012 Lejeune et al 2013)

However recent studies have reported similar overall thin-ning rates over debris-covered and debris-free ice in thePKH (Kaumlaumlb et al 2012 Gardelle et al 2012a) or evenhigher lowering rates over debris-covered parts in the Ever-est area (Nuimura et al 2012) Our findings are consistentwith Nuimura et al (2012) for the Everest study site with astronger thinning (rate of elevation change ofminus097 m yrminus1)

in debris-covered areas than over clean ice (rate of elevationchange ofminus057 m yrminus1) In the Pamir Karakoram and SpitiLahaul study sites these rates are similar while in the HinduKush West Nepal Bhutan and Hengduan Shan study sitesthinning is greater over clean ice in agreement with the com-monly assumed protective effect of debris (Fig 4) Thosedifferences between our 9 study sites suggest a complex re-lationship between debris cover and glacier wastage

As local glacier thickness changes are a result of bothmass balance processes and a dynamic component theseunexpected high thinning rates over debris-covered parts

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1275

Fig 8Same as Fig 2 but for the Karakoram east study site Surge-type glaciers in an active surge phase (resp quiescent phase) are identifiedwith a solid triangle (resp solid circle)

are potentially the result of glacier-wide processes Over aglacier tongue the ablation can be enhanced by the pres-ence of steep exposed ice cliffs or supraglacial lakes andponds (Sakai et al 2000 2002) In addition it is also pos-sible that the glacier tongues are mostly covered by thindebris layers such that the albedo effect dominates the in-sulating effect resulting in an increased tongue-wide abla-tion The prevalence of thin debris was recently suggestedfor debris-covered glaciers around the Mount Gongga in thesouth-eastern Tibetan Plateau (Zhang et al 2013b) Finallyice-flow rates could be very low at debris-covered parts asshown by Quincey et al (2007) in the Everest area Kaumlaumlb(2005) for Bhutan and Scherler et al (2011b) in the HinduKush-Karakoram-Himalaya Thus all three factors are likelyto increase the overall thinning under debris despite the un-doubted protective effect of a continuous thick debris coverat a local scale Future investigations combining surface ve-locity fields and maps of elevation changes could allow eval-uating more closely the relative role of ablation and ice fluxesin thinning rates over PKH glacier tongues (eg Berthierand Vincent 2012 Nuth et al 2012) Measurements of the

spatial distribution of debris thickness over the PKH glaciertongues are also needed

Note that in the case of debris-covered tongues the ele-vation change measurement represents the glacier thicknesschange but also the possible debris thickness evolution Thelatter remains poorly known in the PKH or in any othermountain range But based on the debris discharges givenby Bishop et al (1995) for Batura Glacier (Pakistan) 48to 90times 103 m3 yrminus1 the thickening of the debris can be es-timated between 27 and 55 mm yrminus1 which is negligiblegiven the magnitude of the glacier elevation changes

53 Comparison to previous mass balance estimates

Throughout the PKH there is only a single peer-reviewedglaciological record (on Chhota Shigri Glacier) coveringmost of our study period (Wagnon et al 2007 Vincent etal 2013) Our geodetic mass balance estimate for the SpitiLahaul study site has been recently compared to those glacio-logical measurements on Chhota Shigri Glacier (Vincent etal 2013) Mass balance records reported in Yao et al (2012)for the Tibetan Plateau and the surroundings mountain ranges

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1276 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 9Same as Fig 8 but for the Karakoram west study site

only start during glaciological year 20052006 This is whyhere we do not compare our mass balance estimates toglaciological records and limit our comparison to others re-gional estimates obtained using the geodetic or gravimetricmethods

We stress that the comparison of our mass balance mea-surements to published estimates is complicated by thefact that generally they do not cover the same time pe-riod Little is known about the inter-annual variability inthe PKH during the 21st century due to the limited num-ber of long glaciological time series The 9 yr mass balance

record on Chhota Shigri Glacier reveals a relatively highinter-annual variability of the annual mass balance (stan-dard deviationplusmn 074 m we yrminus1 during 2003ndash2011 Vin-cent et al 2013) which is similar to the one obtained be-tween 1968 and 1998 on Abramov Glacier (standard devi-ation 069 m we yrminus1 WGMS 2012) Thus inter-annualvariability can explain part of the difference between two cu-mulative mass balance estimates over 5ndash10 yr even if they areoffset by only one or two years

Based on ICESat repeat track measurements and theSRTM DEM Kaumlaumlb et al (2012) measured a region-wide

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1277

Fig 10Same as Fig 8 but for the Pamir study site

mass balance ofminus021plusmn 005 m we yrminus1 between 2003and 2008 for Hindu Kush-Karakoram-Himalaya glaciersFor that same area (ie excluding the Pamir and theHengduan Shan study sites) we found a mass balance ofminus015plusmn 007 m we yrminus1 Thus our result agrees with Kaumlaumlbet al (2012) within the error bars although our study periodis longer (1999ndash2011) and our measurement principle andextrapolation approach are different In addition we havecalculated updated mass balances from Kaumlaumlb et al (2012)ie averages over 3times 3 degree cells centered over our studysites and find that they are consistent within error barswith our mass balance estimates (Table 5) Similar resultsare found when our mass balances are compared to a spa-tially extended ICESat analysis for 2003ndash2009 (Gardner etal 2013) Mass losses measured with ICESat (Kaumlaumlb et al2012 Gardner et al 2013) tend to be slightly larger thanours This could be due to the difference in time periodsAlso our smaller mass losses could be partly explained if wesystematically underestimated the poorly constrained pene-

tration of the C-band andor X-band radar signal into ice andsnow

At a more local scale Bolch et al (2011) reported a massbalance ofminus079plusmn 052 m yrminus1 between 2002 and 2007 byDEM differencing over a glacier area of 62 km2 includingten glaciers south and west of Mt Everest For these sameglaciers Nuimura et al (2012) found a mean mass balanceof minus045plusmn 060 m yrminus1 during 2000ndash2008 while they com-puted a mass balance ofminus040plusmn 025 m yrminus1 between 1992and 2008 over a glacier area of 183 km2 around Mt EverestIn the present study we found an average mass balance ofminus041plusmn 021 m yrminus1 over the ten glaciers mentioned above(Table 5 Fig S1) and a value ofminus026plusmn 013 m yrminus1 overthe whole Everest study site between 2000 and 2011 Ourvalues are comparable to those of Nuimura et al (2012)Our mass balance is less negative than the 2002ndash2007 massbalance of Bolch et al (2011) but not statistically differentwhen error bars are considered (Fig 12) Indeed Bolch etal (2011) acknowledged some high uncertainties for their

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1278 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Table 4Glacier mass budget of the nine sub-regions to which the results of the study sites (mass balances) have been extrapolated For eachof them the total glacierized area as well as the percentage of the glacier area over which the mass balance was computed (ie study sites)are given The total glacier area is derived from the RGI v20 (Arendt et al 2012) The error for the mass balance in each sub-region is theroot of sum of squares (RSS) of the mass balance error of each study site and the regional extrapolation error The error for the mass budgetin each sub-region includes additionally the error from the total glacier area in the RGI v20

Sub-region Total glacier Measured glacier Mass balance Mass budgetarea (km2) area () (m we yrminus1) (Gt yrminus1)

Hengduan Shan 11584 12 minus033plusmn 014 minus38plusmn 38Bhutan 4021 34 minus022plusmn 013 minus09plusmn 05Everest 6226 23 minus026plusmn 014 minus16plusmn 08West Nepal 6849 13 minus032plusmn 014 minus22plusmn 10Spiti Lahaul 9043 23 minus045plusmn 014 minus41plusmn 13Hindu Kush 6135 13 minus012plusmn 016 minus07plusmn 10Karakoram East

1902428 +011 plusmn 014 +19 plusmn 31

Karakoram West 30 +009 plusmn 018Pamir 9369 34 +014 plusmn 014 +13 plusmn 13

TotalArea- 72251 30 minus014plusmn 008 minus101plusmn 55weighted average

Table 5Comparison of geodetic mass balances between this study and previous published results with overlapping study periods and similargeographic locations Updated figures from Kaumlaumlb et al (2012) are averaged over 3 times 3 cells centered over the study sites of the presentstudy

GlacierSite name This study Bolch et al (2011) Nuimura et al (2012) Kaumlaumlb et al (2012 updated)2002ndash2007 2000ndash2008 2003ndash2008

Study Mass balance Area (km2)a Mass balance Area Mass balance Mass balanceperiod (m we yrminus1) (m we yrminus1) (km2) (m we yrminus1) (m we yrminus1)

Hengduan San 1999ndash2010 minus022 plusmn 014 1303 (55 )Bhutan 1999ndash2010 minus022 plusmn 012 1384 (64 ) minus052 plusmn 016b

Everest

1999ndash2011

minus026 plusmn 013 1461 (58 ) minus039 plusmn 011AX010 minus090plusmn 034 04Changri SharNup minus042plusmn 017 161 (79 ) minus029plusmn 052 130 minus055plusmn 038Khumbu minus051plusmn 019 203 (47 ) minus045plusmn 052 170 minus076plusmn 052Nuptse minus037plusmn 020 50 (74 ) minus040plusmn 053 40 minus034plusmn 027Lhotse Nup minus021plusmn 027 24 (70 ) minus103plusmn 051 19 minus022plusmn 047Lhotse minus043plusmn 018 85 (83 ) minus110plusmn 052 65 minus067plusmn 051Lhotse SharImja minus070plusmn 052 98 (44 ) minus145plusmn 052 107 minus093plusmn 060Amphu Laptse minus046plusmn 034 25 (38 ) minus077plusmn 052 15 minus018plusmn 094Chukhung +044 plusmn 024 42 (47 ) +004 plusmn 054 38 +043 plusmn 081Ama Dablam minus049plusmn 017 36 (54 ) minus056plusmn 052 22 minus056plusmn 073Duwo minus016plusmn 026 19 (18 ) minus196plusmn 053 10 minus068plusmn 074Total Khumbu minus041plusmn 021 744 minus079plusmn 052 617 minus045plusmn 060(10 Glaciers above)

West Nepal 1999ndash2011 minus032 plusmn 013 908 (40 ) minus032 plusmn 012Spiti Lahaul 1999ndash2011 minus045 plusmn 013 2110 (46 ) minus038 plusmn 006Karakoram East 1999ndash2010 +011 plusmn 014 5328 (42 ) minus004 plusmn 004Karakoram West 1999ndash2008 +009 plusmn 018 5434 (45 )Hindu Kush 1999ndash2008 minus012 plusmn 016 793 (80 ) minus020 plusmn 006Pamir 1999ndash2011 +014 plusmn 013 3178 (50 )

a The total glacier area is given in km2 and in parenthesis the of the glacier area actually covered with measurementsb For this cell the ICESat coverage is insufficient and does not sample all glacier elevations

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1279

Fig 11 Elevation changes (SPOT5-SRTM) in the ablation area for clean ice (blue circles) and debris-covered ice (black circles) for eachstudy site For the Karakoram (east and west) and the Pamir study sites only non-surge-type glaciers are considered Standard error of theelevation differences are typically less thanplusmn2 m with each 100 m elevation band (not shown for the sake of clarity) Note elevation changesover separate sections of a glacier cannot be treated as mass changes due to the disregard of glacier dynamics

estimate over this short time period The differences in sur-vey periods and in the glacier outlines may also explainpart of these discrepancies In particular the delimitation ofthe accumulation area is notoriously difficult and Bolch etal (2011) did not include some of the steep high altitudeslopes which we included due to their potential avalanchecontribution For the 10 glaciers mentioned above our massbalances would become 012 m we yrminus1 more negative ifwe used the exact same outlines as Bolch et al (2011)Our positive mass balance for the small Chukhung Glacier(+044plusmn 024 m we yrminus1) is in agreement with Nuimura etal (2012) but enigmatic in a context of regional glacier lossand deserves further investigation If the 2002ndash2007 massbalance estimate by Bolch et al (2011) is excluded all otherrecent mass balance estimates suggest no acceleration of themass loss in the Everest area when compared to the long-term(1970ndash2007) mass balance measured by Bolch et al (2011)in this regionminus032plusmn 008 m weyrminus1

In Bhutan our results indicate a stronger mass loss forsouthern glaciers than for northern ones Interestingly in thisarea Kaumlaumlb (2005) measured high speeds (up to 200 m yrminus1)

for glaciers north of the Himalayan main ridge and ratherlow velocities over southern glacier tongues by using repeat

ASTER data This is consistent with the more negative massbalance that we report for the southern glaciers in Bhutanie for the slow flowing presumably downwasting glaciers

Berthier et al (2007) reported a mass balance betweenminus069 andminus085 m we yrminus1 for an ice-covered area of915 km2 around Chhota Shigri Glacier between 1999 (SRTMDEM) and 2004 (SPOT5-HRG DEM) For the same areawe found a mass balance ofminus045plusmn 013 m we yrminus1 over1999ndash2011 The difference in the study period may explainpart of the differences Also the methodology by Berthieret al (2007) did not explicitly take into account the SRTMradar penetration over glaciers and also corrected an appar-ent elevation-dependent bias according to glacier elevationswhich is now known as a resolution issue and is best cor-rected according to the terrain maximum curvature (Gardelleet al 2012b) More details and discussions about these dif-ferences can be found in Vincent et al (2013)

Importantly the results of the eastern Karakoram studysite confirm the balanced or slightly positive mass budgetreported by Gardelle et al (2012a) in the western Karako-ram over the past decade The two Karakoram study sitesare adjacent with a small overlapping area (sim 1100 km2 ofglaciers) where elevation differences agree within 1 m Both

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1280 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 12 Comparison of geodetic mass balances for 10 glaciers inthe Mount Everest area 1999ndash2010 mass balances measured inthe present study are compared with published estimates during2002ndash2007 (Bolch et al 2011) and during 2000ndash2008 (Nuimuraet al 2012) The 1-to-1 line is drawn The mean difference (plusmn1standard deviation) between (i) our values and Bolch et al (2011)are (047plusmn 058 m we yrminus1) and (ii) our values and Nuimura etal (2012) are (012plusmn 021 m we yrminus1) Part of the differences be-tween estimates may be due to the different periods surveyed andthe different glacier outlines used

sites show similar mass balances over slightly different pe-riods +011plusmn 014 m yrminus1 over 1999ndash2010 in the east and+009plusmn 018 m yrminus1 over 1999ndash2008 in the west This con-firms a ldquoKarakoram anomalyrdquo that was first suggested basedon field observations (Hewitt 2005) We report a mass bal-ance of+010plusmn 013 m yrminus1 for Baltoro Glacier (see loca-tion in Fig 8 size= 304 km2) between 1999 and 2010which is consistent with its gradual speed-up reported since2003 (Quincey et al 2009) However given the completelydifferent methods used in our and the latter study we can-not conclude that this demonstrates a real shift from nega-tive to positive mass balance The mass budget of SiachenGlacier (1145 km2 sometimes referred to as the highestbattlefield in the world) is also balanced or slightly pos-itive (+014plusmn 014 m we yrminus1) Thus the alarming state-ment by Rasul (2012) that this glacier retreated by 59 kmand lost 17 of its mass during 1989ndash2009 is very likelyerroneous Again within error bars our regional mass bal-ance in the Karakoram agrees with the mass balance ofminus003plusmn 004 m yrminus1 found by Kaumlaumlb et al (2012) over 2003ndash2008 and with the mass balance ofminus010plusmn 018 m we yrminus1

(2-sigma error level) obtained by Gardner et al (2013) for aregion including both the Karakoram and the Hindu Kush

In the northernmost part of the Pamir in the Alay moun-tain range the mass balance of Abramov Glacier has beenmeasured in the field for 30 yr between 1968 and 1998(WGMS 2012) with a mean value ofminus046 m yrminus1 Herewe report a mass balance ofminus003plusmn 014 m yrminus1 for this

glacier over 1999ndash2011 Our near zero mass budget estimatefor this glacier disagrees with negative mass balances recon-structed for the same period with a model run using biascorrected NCEPNCAR re-analysis data and calibrated withfield data (Barundun et al 2013) An underestimate of ourSRTM C-Band penetration in the Pamir as a whole and inparticular for the small Abramov Glacier may contribute tothese differences

For the whole Pamir study site our mass budget is bal-anced or slightly positive (+014plusmn 013 m we yrminus1) In theeastern Pamir the mass balance of Muztag Ata Glacier was+025 m we yrminus1 between 2005 and 2010 (Yao et al 2012)Although Muztag Ata Glacier is located outside of our Pamirstudy site and his glaciological records only cover the sec-ond half of our study period this measurement is consistentwith our remotely sensed mass balance Thus the ldquoKarako-ram anomalyrdquo should probably be renamed the ldquoPamir-Karakoram anomalyrdquo

This slightly positive or balanced mass budget measuredin the Pamir seems to contradict previous findings by Heidand Kaumlaumlb (2012) They reported rapidly decreasing glacierspeeds in this region between two snapshots of annual ve-locities in 20002001 and 20092010 an observation thatwould fit with negative mass balances (Span and Kuhn 2003Azam et al 2012) We note though that the relationship be-tween glacier mass balance and ice fluxes is not straightfor-ward and might in particular involve a time delay betweena change in mass balance and the related dynamic response(Heid and Kaumlaumlb 2012) Thus the decrease in speed couldwell be due to negative mass balances before or partiallybefore 1999ndash2011 We acknowledge that this discrepancyneeds to be investigated further in particular by examiningcontinuous changes in annual glacier speed through the firstdecade of the 21st Century Indeed Heid and Kaumlaumlb (2012)measured differences between two annual periods which maynot represent mean decadal velocity changes

Jacob et al (2012) using satellite gravimetry (GRACE)measured a mass budget ofminus5plusmn 6 Gt yrminus1 over 2003ndash2010for the ldquoKarakoram-Himalayardquo their region 8c which cor-responds to our study area excluding the Pamir Karakoramand Hindu Kush sub-regions but including some glaciersnorth of the Himalayan ridge already on the Tibetan PlateauFor this subset of our PKH study area we measured amass budget ofminus126plusmn 42 Gt yrminus1 The two estimates over-lap within their error bars especially if our error is dou-bled to match the two standard error level used by Jacob etal (2012) The estimate of Jacob et al (2012) over our com-plete study region (PKH) ieminus6plusmn 8 Gt yrminus1 (sum of theirregions 8b and 8c) also includes mass changes for the KunlunShan sub-region for which we did not measure glacier massbalances Therefore the comparison with our PKH value(minus101plusmn 55 Gt yrminus1) is not straightforward but suggests areasonable agreement especially if we consider the slightmass gain (15plusmn 17 Gt yrminus1) measured with ICESat in theWest Kunlun (Gardner et al 2013)

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1281

Table 6 Annual average of the glacier seasonal and decadal mass loss contributions to Indus Ganges and Brahmaputra discharges Basinarea and seasonal contributions are given by Kaser et al (2010) glacier area in each basin is derived from the RGI v20 (Arendt et al 2012)Numbers in parentheses indicate the percentage of glacierized area within the river basin Seasonal contributions were modeled by Kaser etal (2010) based on the CRU 20 dataset over 1961ndash1990 assuming a balanced glacier mass budget Glacier decadal mass loss contributionsare given as annual average over 1999ndash2011

River Basin area Glacier area Glacier contribution Mean discharge(km2) (km2) (m3 sminus1) (m3 sminus1)

Seasonally delayed Decadal mass lossKaser et al (2010) (this study)

Indus 1 139 814 25 598 (2 ) 105 103plusmn 72 2146 (Chevallier personalcommunication 2012)

Ganges 1 023 609 11 168 (1 ) 47 103plusmn 49 11739 Papa et al (2012Brahmaputra 527 666 15 296 (3 ) 33 147plusmn 64 19710 updated)

54 Reasons for the heterogeneous pattern of massbalance

The PKH glacier mass balance is slightly negative(minus014plusmn 008 m we yrminus1) but changes are not homoge-neous from one sub-region to another and seem to coincidewith the climatic settings of the mountain range For the east-ern sites (from the Hengduan Shan to West Nepal) which areunder the influence of the Indian and south-east Asian sum-mer monsoons the mass balances are moderately negative(sim minus026 m we yrminus1) In the northwest of the PKH at thePamir and Karakoram sites where glacier climate is char-acterized by the westerlies in winter mass balances are bal-anced or slightly positive (sim +010 m we yrminus1) In betweenthese two influences the glaciers of the Spiti Lahaul sitein a monsoon-arid transition zone experienced the strongestmass loss (minus045plusmn 013 m we yrminus1)

This heterogeneous pattern of mass balances seems to berelated to trends in precipitation throughout the PKH Archerand Fowler (2004) reported an increase in winter precipi-tation over the Upper Indus Basin since 1961 a trend con-firmed by Yao et al (2012) over the Pamir and the Karako-ram based on the analysis of the Global Precipitation Cli-matology Project data set (GPCP Adler et al 2003) Thisprecipitation increase is a source of greater accumulation forglaciers and thus a possible explanation for their slightly pos-itive mass budgets In addition in the Upper Indus Basin themean summer temperature has been decreasing since 1961(Fowler and Archer 2006) which is consistent with the find-ings of Shekhar et al (2010) in the Karakoram who reporteda decrease in both maximum and minimum temperature since1988 These increases in winter precipitation and summertemperature likely translate in greater accumulation and re-duced ablation changes which are consistent with the bal-anced or slightly positive glacier mass budget

On the other hand in the east of the PKH the Indiansummer monsoon has been weakening since the 1950s (Bol-lasina et al 2011) which could have reduced the amount

of snowfall over the eastern study sites and thus resultedin negative mass balances In addition maximum and min-imum temperatures increased since 1988 in the western Hi-malaya (Shekhar et al 2010) and maximum temperaturesincreased in the central Himalaya (Nepal) since the mid-1970s (Shrestha et al 1999) Thus both precipitation andtemperature trends in the Himalaya likely contributed to thenegative mass balances measured over the correspondingstudy sites (from Spiti Lahaul to Hengduan Shan Fig 1)

However gridded data sets such as GPCP are not neces-sarily representative of the climate at high altitude IndeedHewitt (2005) reported a 5-to-10-fold increase in precipi-tation between the glacier front and the accumulation areain the Karakoram that is not captured by reanalysis dataas recently confirmed by Immerzeel et al (2012) Thus di-rect measurements of accumulation on High Mountain Asiaglaciers (eg Azam et al 2013) and high-altitude weatherstations are eagerly needed to validate the large scale grid-ded data set (eg reanalysis) test their ability to describe thespecific climate near to PKH glaciers and assess the relativerole played by temperature and precipitation changes

55 Contribution to water resources

Our glacier mass balance can be averaged over large riverbasins to estimate the mean contribution to river runoffdue to a decade of glacier mass loss We assume therebyonly direct river runoff without loss terms For the threerivers for which we cover the entire glacierized area of theircatchments within our sub-regions (Fig 1) these contribu-tions to the river discharge are equal to 103 m3 sminus1 (Indus)103 m3 sminus1 (Ganges) and 147 m3 sminus1 (Brahmaputra) for theperiod 1999ndash2011 (Table 6)

Rivers of PKH being key water resources measures oftheir discharges are highly sensitive and difficult to access forpolitical reasons Papa et al (2012) built recent time series ofdischarges for Ganges and Brahmaputra by using altimetrymeasurements The locations of the corresponding dischargeestimates in Bangladesh are given in Fig 1 We can therefore

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1282 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

evaluate the relative contribution of glacier decadal mass loss(Table 6) to annual river run-off at 09 for the Gangesat Hardinge (basin area ofsim 1 020 000 km2) and 07 forBrahmaputra at Bahadurabad (basin area ofsim 530 000 km2)between 1999 and 2011 Given the discharge measured at theGuddu station by the Surface Water Hydrology Project of theWater and Power Development Authority (Pakistan see lo-cation on Fig 1) between 1999 and 2003 we can estimatethe contribution of glacier mass loss at 48 for the IndusRiver at this station

The seasonal contributions of glaciers to river run-off (iethe part of the annual precipitation that experiences season-ally delayed release from the glaciers) directly results fromseasonal variations of meteorological quantities in contrastto the glacier mass loss contribution which results from thelong-term climatic change Both runoff components directlyaffect the amount of water available in a river at a given lo-cation and at a given time so that their comparison could beof interest Even if the seasonal timing of the glacier massloss contribution is unclear it could in general peak at asimilar time of the year than the seasonal glacier contribu-tion so that both components rather enlarge each other thancancel Seasonal glacier contributions have been modeled byKaser et al (2010) The latter study assumed glaciers to bein equilibrium with the present climate and thus prescribedglacier mass balances equal to 0 Thus their seasonally de-layed glacier contributions do not include the part due todecadal glacier mass loss and is complementary to our es-timates (Table 6) For the eastern basins under the influenceof the Indian summer monsoon (Ganges and Brahmaputra)the contribution of glacier mass loss is higher than the sea-sonal contribution In other words for the first decade of the21st century the mass loss of glaciers is on average moreimportant for water availability in those two rivers than theirseasonal water storage effect The situation is different for theIndus Basin where the glacier melt season is also the dry sea-son which results in a relatively higher seasonally delayedglacier contribution to the river discharge There the season-ally delayed water release from the glaciers and the decadalglacier mass loss contribute on average equally to the riverdischarge Both contributions are strongly (and equally) de-pendent on the location of the discharge measurement andwill be significantly larger (in relative term) in more heav-ily glacierized and smaller mountain catchments located inthe upper part of these major river basins (Bookhagen andBurbank 2010)

6 Conclusion

In this study we assessed the spatial pattern of glaciermass balance over the PKH by comparing the SRTM andSPOT5 DEMs over 9 well-distributed study sites between1999 and 2011 We found slightly positive or balanced massbudgets in the western part (Pamir Karakoram) moder-ate mass loss in the east (Nepal Bhutan Hengduan Shan)

and the most negative mass balance in the monsoon-aridtransition zone (Spiti Lahaul in the western Himalaya)By extrapolating these values to climatically homogeneoussub-regions we estimate the PKH overall mass balanceto have beenminus014plusmn 008 m we yrminus1 between 1999 and2011 which corresponds to a contribution to sea level riseof 0028plusmn 0015 mm yrminus1 This is about 4 of the globalglacier mass loss estimated by Gardner et al (2013) thoughover a slightly shorter time period (2003ndash2009) The largestsource of uncertainties in those geodetic mass balance esti-mates is the poorly constrained penetration of the C-BandSRTM radar signal into snow and ice

Our technique of DEM differencing allows mapping in de-tail the spatial pattern of glacier elevation changes Thosemaps reveal many surge-type behaviors both in the Pamirand the Karakoram It also appears that the glacier-wide massbalance does not seem to be considerably affected by themass transfer from surges over thesim 10 yr time period stud-ied Similar lowering rates over debris-covered and clean iceare observed for four study sites despite the commonly as-sumed insulating effect of debris covers This suggests forthe scale of entire glacier tongues a spatial mixture of re-duced ablation under intact thick debris covers and enhancedablation by supraglacial lakes or ice cliffs or due to thin de-bris layer and very low glacier velocities in ablation areasthat reduce the ice advection downstream

The regionally heterogeneous pattern of ice wastage is inbroad agreement with contrasted trends in large-scale pre-cipitation and temperature However glaciological (eg ac-cumulation) and meteorological records are needed at highaltitude in the vicinity of glaciers to evaluate more closelythe local influence of atmospheric variables on glacier massbalance and fully understand the reasons behind those con-trasted mass budgets in the PKH

Supplementary material related to this article isavailable online at httpwwwthe-cryospherenet712632013tc-7-1263-2013-supplementpdf

Acknowledgements We thank G Cogley A Gardner T NuimuraT Bolch P Wagnon M Barandun M Hoelzle M Huss and ananonymous referee for their constructive comments that led to aconsiderably improved manuscript We thank the Water and PowerDevelopment Authority (WAPDA) for their hydrological data(processed by P Chevallier) and F Papa for sharing his updatedrun-off data ahead of publication We thank T Bolch for sharinghis glacier outlines from the Everest area We thank the USGSfor allowing free access to their Landsat archive CGIAR-SCI forSRTM C-band data DLR for SRTM X-band data and GLIMS forglacier inventories J Gardelle acknowledges a PhD fellowshipfrom the French Space Agency (CNES) and the French NationalResearch Center (CNRS) E Berthier acknowledges support fromCNES through the TOSCA and ISIS proposal no 397 and fromthe Programme National de Teacuteleacutedeacutetection Spatiale E Berthieracknowledges M Masiokas and R Villalba for welcoming him

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1283

at the IANIGLA (Mendoza Argentina) Y Arnaud acknowledgessupport from French National Research Agency through ANR-09-CEP-005-01PAPRIKA A Kaumlaumlb acknowledges support fromESA through Glaciers_cci (400010177810I-AM) and from theEuropean Research Council under the European Unionrsquos SeventhFramework Programme (FP2007-2013)ERC Grant Agreementn 320816

Edited by I M Howat

The publication of this articleis financed by CNRS-INSU

References

Adler R Huffman G Chang A Ferraro R Xie P JanowiakJ Rudolf B Schneider U Curtis S Bolvin D Gruber ASusskind J Arkin P and Nelkin E The Version-2 GlobalPrecipitation Climatology Project (GPCP) Monthly Precipita-tion Analysis (1979ndashPresent) J Hydrometeorol 4 1147ndash11672003

Archer D R and Fowler H J Spatial and temporal variationsin precipitation in the Upper Indus Basin global teleconnectionsand hydrological implications Hydrol Earth Syst Sci 8 47ndash61doi105194hess-8-47-2004 2004

Arendt A Bolch T Cogley J G Gardner A Hagen J-OHock R Kaser G Pfeffer W T Moholdt G Paul F RadicV Andreassen L Bajracharya S Beedle M Berthier EBhambri R Bliss A Brown I Burgess E Burgess DCawkwell F Chinn T Copland L Davies B de AngelisH Dolgova E Filbert K Forester R Fountain A Frey HGiffen B Glasser N Gurney S Hagg W Hall D Hari-tashya U K Hartmann G Helm C Herreid S Howat IKapustin G Khromova T Kienholz C Koenig M KohlerJ Kriegel D Kutuzov S Lavrentiev I LeBris R Lund JManley W Mayer C Miles E Li X Menounos B Mer-cer A Moelg N Mool P Nosenko G Negrete A NuthC Pettersson R Racoviteanu A Ranzi R Rastner P RauF Rich J Rott H Schneider C Seliverstov Y Sharp MSigurethsson O Stokes C Wheate R Winsvold S WolkenG Wyatt F and Zheltyhina N Randolph Glacier Inventory[v20] A Dataset of Global Glacier Outlines Global Land IceMeasurements from Space Boulder Colorado USA Digital Me-dia httpwwwglimsorgRGIrandolphhtml 2012

Azam M Wagnon P Ramanathan A Vincent C Sharma PArnaud Y Linda A Pottakkal J Chevallier P Singh V andBerthier E From balance to imbalance a shift in the dynamicbehaviour of Chhota Shigri glacier western Himalaya India JGlaciol 58 315ndash324 doi1031892012JoG11J123 2012

Azam M F Wagnon P Vincent C Ramanathan A Linda Aand Singh V B Reconstruction of the annual mass balanceof Chhota Shigri Glacier (Western Himalaya India) since 1969Ann Glaciol in revision 2013

Barrand N and Murray T Multivariate Controls on the In-cidence of Glacier Surging in the Karakoram Himalaya

Arct Antarct Alp Res 38 489ndash498 doi1016571523-0430(2006)38[489MCOTIO]20CO2 2006

Barundun M Huss M Azisov E Gafurov A HoelzleM Merkushkin A Salzmann N and Usubaliev R Re-establishing seasonal mass balance observation at AbramovGlacier Kyrgyzstan from 1968ndash2012 Abstract EGU2013-5668European Geophysical Union Vienna 2013

Benn D Bolch T Hands K Gulley J Luckman ANicholson L Quincey D Thompson S Toumi R andWiseman S Response of debris-covered glaciers in theMount Everest region to recent warming and implicationsfor outburst flood hazards Earth-Sci Rev 114 156ndash174doi101016jearscirev201203008 2012

Berthier E and Vincent C Relative contribution of surface mass-balance and ice-flux changes to the accelerated thinning of Merde Glace French Alps over 1979ndash2008 J Glaciol 58 501ndash512doi1031892012JoG11J083 2012

Berthier E Arnaud Y Baratoux D Vincent C and Reacutemy FRecent rapid thinning of the ldquo Mer de Glace rdquo glacier derivedfrom satellite optical images Geophys Res Lett 31 L17401doi1010292004GL020706 2004

Berthier E Arnaud Y Kumar R Ahmad S Wagnon P andChevallier P Remote sensing estimates of glacier mass balancesin the Himachal Pradesh (Western Himalaya India) RemoteSens Environ 108 327ndash338 doi101016jrse2006110172007

Bhambri R Bolch T Kawishwar P Dobhal D P SrivastavaD and Pratap B Heterogeneity in Glacier response from 1973to 2011 in the Shyok valley Karakoram India The CryosphereDiscuss 6 3049ndash3078 doi105194tcd-6-3049-2012 2012

Bhutiyani M Mass-balance studies on Siachen Glacier in theNubra valley Karakoram Himalaya India J Glaciol 45 112ndash118 1999

Bishop M P Schroder J F and Ward J L SPOT multispec-tral analysis for producing supraglacial debris-load estimates forBatura glacier Pakistan Geocarto Int 10 81ndash90 1995

Bolch T Pieczonka T and Benn D I Multi-decadal mass lossof glaciers in the Everest area (Nepal Himalaya) derived fromstereo imagery The Cryosphere 5 349ndash358 doi105194tc-5-349-2011 2011

Bolch T Kulkarni A Kaumlaumlb A Huggel C Paul F Cogley JFrey H Kargel J Fujita K Scheel M Bajracharya S andStoffel M The state and fate of Himalayan Glaciers Science336 310ndash314 doi101126science1215828 2012

Bollasina M Ming Y and Ramaswamy V Anthropogenicaerosols and the weakening of the South Asian summer mon-soon Science 334 502ndash505 doi101126science12049942011

Bookhagen B and Burbank D Toward a complete Himalayan hy-drological budget spatiotemporal distribution of snowmelt andrainfall and their impact on river discharge J Geophys Res115 F03019 doi1010292009JF001426 2010

Bretherton C Widmann M Dymnikov V Wallace J and BladeacuteI The effective number of spatial degrees of freedom of a time-varying field J Climate 12 1990ndash2009 1999

Copland L Pope S Bishop M Shroder J Clendon PBush A Kamp U Seong Y and Owen L Glacier veloc-ities across the central Karakoram Ann Glaciol 50 41ndash49doi103189172756409789624229 2009

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

1274 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 7Same as Fig 2 but for the Hindu Kush study site

glaciers are of summer-accumulation type Further studieswould be necessary to investigate the possible reasons be-hind the relatively low SRTMC-bandpenetration for the Pamir(18 m) that do not follow the eastwest gradient mentionedabove

Both our and Kaumlaumlb et alrsquos (2012) estimates agree withintheir error bars and seem thus to provide robust first-orderestimates for the SRTMC-bandradar penetration in the regionHowever it remains to be verified if a penetration depth com-puted at the regional scale is appropriate for a single smallglacier whose altitude distribution is different from the oneof whole study site (eg Abramov Glacier in the far north ofour Pamir study site)

52 Thinning over debris-covered ice

PKH glaciers are heavily debris-covered in their ablation ar-eas because of steep rock walls that surround them and in-tense avalanche activity Twelve percent (12 ) of the glacierarea in our study sites are covered with debris (up to 22 inthe Everest area Table 1) which is in agreement with pre-vious estimates for the Himalaya and the Karakoram 13 in Kaumlaumlb et al (2012) and 10 in Bolch et al (2012) Thedebris layer is expected to modify the ablation of the under-

lying ice It will increase ablation if its thickness is just a fewcentimeters (dominance of the albedo decrease) or decreaseablation it if the layer is thick enough to protect the ice fromsolar radiation (Mattson et al 1993 Mihalcea et al 2006Benn et al 2012 Lejeune et al 2013)

However recent studies have reported similar overall thin-ning rates over debris-covered and debris-free ice in thePKH (Kaumlaumlb et al 2012 Gardelle et al 2012a) or evenhigher lowering rates over debris-covered parts in the Ever-est area (Nuimura et al 2012) Our findings are consistentwith Nuimura et al (2012) for the Everest study site with astronger thinning (rate of elevation change ofminus097 m yrminus1)

in debris-covered areas than over clean ice (rate of elevationchange ofminus057 m yrminus1) In the Pamir Karakoram and SpitiLahaul study sites these rates are similar while in the HinduKush West Nepal Bhutan and Hengduan Shan study sitesthinning is greater over clean ice in agreement with the com-monly assumed protective effect of debris (Fig 4) Thosedifferences between our 9 study sites suggest a complex re-lationship between debris cover and glacier wastage

As local glacier thickness changes are a result of bothmass balance processes and a dynamic component theseunexpected high thinning rates over debris-covered parts

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1275

Fig 8Same as Fig 2 but for the Karakoram east study site Surge-type glaciers in an active surge phase (resp quiescent phase) are identifiedwith a solid triangle (resp solid circle)

are potentially the result of glacier-wide processes Over aglacier tongue the ablation can be enhanced by the pres-ence of steep exposed ice cliffs or supraglacial lakes andponds (Sakai et al 2000 2002) In addition it is also pos-sible that the glacier tongues are mostly covered by thindebris layers such that the albedo effect dominates the in-sulating effect resulting in an increased tongue-wide abla-tion The prevalence of thin debris was recently suggestedfor debris-covered glaciers around the Mount Gongga in thesouth-eastern Tibetan Plateau (Zhang et al 2013b) Finallyice-flow rates could be very low at debris-covered parts asshown by Quincey et al (2007) in the Everest area Kaumlaumlb(2005) for Bhutan and Scherler et al (2011b) in the HinduKush-Karakoram-Himalaya Thus all three factors are likelyto increase the overall thinning under debris despite the un-doubted protective effect of a continuous thick debris coverat a local scale Future investigations combining surface ve-locity fields and maps of elevation changes could allow eval-uating more closely the relative role of ablation and ice fluxesin thinning rates over PKH glacier tongues (eg Berthierand Vincent 2012 Nuth et al 2012) Measurements of the

spatial distribution of debris thickness over the PKH glaciertongues are also needed

Note that in the case of debris-covered tongues the ele-vation change measurement represents the glacier thicknesschange but also the possible debris thickness evolution Thelatter remains poorly known in the PKH or in any othermountain range But based on the debris discharges givenby Bishop et al (1995) for Batura Glacier (Pakistan) 48to 90times 103 m3 yrminus1 the thickening of the debris can be es-timated between 27 and 55 mm yrminus1 which is negligiblegiven the magnitude of the glacier elevation changes

53 Comparison to previous mass balance estimates

Throughout the PKH there is only a single peer-reviewedglaciological record (on Chhota Shigri Glacier) coveringmost of our study period (Wagnon et al 2007 Vincent etal 2013) Our geodetic mass balance estimate for the SpitiLahaul study site has been recently compared to those glacio-logical measurements on Chhota Shigri Glacier (Vincent etal 2013) Mass balance records reported in Yao et al (2012)for the Tibetan Plateau and the surroundings mountain ranges

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1276 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 9Same as Fig 8 but for the Karakoram west study site

only start during glaciological year 20052006 This is whyhere we do not compare our mass balance estimates toglaciological records and limit our comparison to others re-gional estimates obtained using the geodetic or gravimetricmethods

We stress that the comparison of our mass balance mea-surements to published estimates is complicated by thefact that generally they do not cover the same time pe-riod Little is known about the inter-annual variability inthe PKH during the 21st century due to the limited num-ber of long glaciological time series The 9 yr mass balance

record on Chhota Shigri Glacier reveals a relatively highinter-annual variability of the annual mass balance (stan-dard deviationplusmn 074 m we yrminus1 during 2003ndash2011 Vin-cent et al 2013) which is similar to the one obtained be-tween 1968 and 1998 on Abramov Glacier (standard devi-ation 069 m we yrminus1 WGMS 2012) Thus inter-annualvariability can explain part of the difference between two cu-mulative mass balance estimates over 5ndash10 yr even if they areoffset by only one or two years

Based on ICESat repeat track measurements and theSRTM DEM Kaumlaumlb et al (2012) measured a region-wide

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1277

Fig 10Same as Fig 8 but for the Pamir study site

mass balance ofminus021plusmn 005 m we yrminus1 between 2003and 2008 for Hindu Kush-Karakoram-Himalaya glaciersFor that same area (ie excluding the Pamir and theHengduan Shan study sites) we found a mass balance ofminus015plusmn 007 m we yrminus1 Thus our result agrees with Kaumlaumlbet al (2012) within the error bars although our study periodis longer (1999ndash2011) and our measurement principle andextrapolation approach are different In addition we havecalculated updated mass balances from Kaumlaumlb et al (2012)ie averages over 3times 3 degree cells centered over our studysites and find that they are consistent within error barswith our mass balance estimates (Table 5) Similar resultsare found when our mass balances are compared to a spa-tially extended ICESat analysis for 2003ndash2009 (Gardner etal 2013) Mass losses measured with ICESat (Kaumlaumlb et al2012 Gardner et al 2013) tend to be slightly larger thanours This could be due to the difference in time periodsAlso our smaller mass losses could be partly explained if wesystematically underestimated the poorly constrained pene-

tration of the C-band andor X-band radar signal into ice andsnow

At a more local scale Bolch et al (2011) reported a massbalance ofminus079plusmn 052 m yrminus1 between 2002 and 2007 byDEM differencing over a glacier area of 62 km2 includingten glaciers south and west of Mt Everest For these sameglaciers Nuimura et al (2012) found a mean mass balanceof minus045plusmn 060 m yrminus1 during 2000ndash2008 while they com-puted a mass balance ofminus040plusmn 025 m yrminus1 between 1992and 2008 over a glacier area of 183 km2 around Mt EverestIn the present study we found an average mass balance ofminus041plusmn 021 m yrminus1 over the ten glaciers mentioned above(Table 5 Fig S1) and a value ofminus026plusmn 013 m yrminus1 overthe whole Everest study site between 2000 and 2011 Ourvalues are comparable to those of Nuimura et al (2012)Our mass balance is less negative than the 2002ndash2007 massbalance of Bolch et al (2011) but not statistically differentwhen error bars are considered (Fig 12) Indeed Bolch etal (2011) acknowledged some high uncertainties for their

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1278 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Table 4Glacier mass budget of the nine sub-regions to which the results of the study sites (mass balances) have been extrapolated For eachof them the total glacierized area as well as the percentage of the glacier area over which the mass balance was computed (ie study sites)are given The total glacier area is derived from the RGI v20 (Arendt et al 2012) The error for the mass balance in each sub-region is theroot of sum of squares (RSS) of the mass balance error of each study site and the regional extrapolation error The error for the mass budgetin each sub-region includes additionally the error from the total glacier area in the RGI v20

Sub-region Total glacier Measured glacier Mass balance Mass budgetarea (km2) area () (m we yrminus1) (Gt yrminus1)

Hengduan Shan 11584 12 minus033plusmn 014 minus38plusmn 38Bhutan 4021 34 minus022plusmn 013 minus09plusmn 05Everest 6226 23 minus026plusmn 014 minus16plusmn 08West Nepal 6849 13 minus032plusmn 014 minus22plusmn 10Spiti Lahaul 9043 23 minus045plusmn 014 minus41plusmn 13Hindu Kush 6135 13 minus012plusmn 016 minus07plusmn 10Karakoram East

1902428 +011 plusmn 014 +19 plusmn 31

Karakoram West 30 +009 plusmn 018Pamir 9369 34 +014 plusmn 014 +13 plusmn 13

TotalArea- 72251 30 minus014plusmn 008 minus101plusmn 55weighted average

Table 5Comparison of geodetic mass balances between this study and previous published results with overlapping study periods and similargeographic locations Updated figures from Kaumlaumlb et al (2012) are averaged over 3 times 3 cells centered over the study sites of the presentstudy

GlacierSite name This study Bolch et al (2011) Nuimura et al (2012) Kaumlaumlb et al (2012 updated)2002ndash2007 2000ndash2008 2003ndash2008

Study Mass balance Area (km2)a Mass balance Area Mass balance Mass balanceperiod (m we yrminus1) (m we yrminus1) (km2) (m we yrminus1) (m we yrminus1)

Hengduan San 1999ndash2010 minus022 plusmn 014 1303 (55 )Bhutan 1999ndash2010 minus022 plusmn 012 1384 (64 ) minus052 plusmn 016b

Everest

1999ndash2011

minus026 plusmn 013 1461 (58 ) minus039 plusmn 011AX010 minus090plusmn 034 04Changri SharNup minus042plusmn 017 161 (79 ) minus029plusmn 052 130 minus055plusmn 038Khumbu minus051plusmn 019 203 (47 ) minus045plusmn 052 170 minus076plusmn 052Nuptse minus037plusmn 020 50 (74 ) minus040plusmn 053 40 minus034plusmn 027Lhotse Nup minus021plusmn 027 24 (70 ) minus103plusmn 051 19 minus022plusmn 047Lhotse minus043plusmn 018 85 (83 ) minus110plusmn 052 65 minus067plusmn 051Lhotse SharImja minus070plusmn 052 98 (44 ) minus145plusmn 052 107 minus093plusmn 060Amphu Laptse minus046plusmn 034 25 (38 ) minus077plusmn 052 15 minus018plusmn 094Chukhung +044 plusmn 024 42 (47 ) +004 plusmn 054 38 +043 plusmn 081Ama Dablam minus049plusmn 017 36 (54 ) minus056plusmn 052 22 minus056plusmn 073Duwo minus016plusmn 026 19 (18 ) minus196plusmn 053 10 minus068plusmn 074Total Khumbu minus041plusmn 021 744 minus079plusmn 052 617 minus045plusmn 060(10 Glaciers above)

West Nepal 1999ndash2011 minus032 plusmn 013 908 (40 ) minus032 plusmn 012Spiti Lahaul 1999ndash2011 minus045 plusmn 013 2110 (46 ) minus038 plusmn 006Karakoram East 1999ndash2010 +011 plusmn 014 5328 (42 ) minus004 plusmn 004Karakoram West 1999ndash2008 +009 plusmn 018 5434 (45 )Hindu Kush 1999ndash2008 minus012 plusmn 016 793 (80 ) minus020 plusmn 006Pamir 1999ndash2011 +014 plusmn 013 3178 (50 )

a The total glacier area is given in km2 and in parenthesis the of the glacier area actually covered with measurementsb For this cell the ICESat coverage is insufficient and does not sample all glacier elevations

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1279

Fig 11 Elevation changes (SPOT5-SRTM) in the ablation area for clean ice (blue circles) and debris-covered ice (black circles) for eachstudy site For the Karakoram (east and west) and the Pamir study sites only non-surge-type glaciers are considered Standard error of theelevation differences are typically less thanplusmn2 m with each 100 m elevation band (not shown for the sake of clarity) Note elevation changesover separate sections of a glacier cannot be treated as mass changes due to the disregard of glacier dynamics

estimate over this short time period The differences in sur-vey periods and in the glacier outlines may also explainpart of these discrepancies In particular the delimitation ofthe accumulation area is notoriously difficult and Bolch etal (2011) did not include some of the steep high altitudeslopes which we included due to their potential avalanchecontribution For the 10 glaciers mentioned above our massbalances would become 012 m we yrminus1 more negative ifwe used the exact same outlines as Bolch et al (2011)Our positive mass balance for the small Chukhung Glacier(+044plusmn 024 m we yrminus1) is in agreement with Nuimura etal (2012) but enigmatic in a context of regional glacier lossand deserves further investigation If the 2002ndash2007 massbalance estimate by Bolch et al (2011) is excluded all otherrecent mass balance estimates suggest no acceleration of themass loss in the Everest area when compared to the long-term(1970ndash2007) mass balance measured by Bolch et al (2011)in this regionminus032plusmn 008 m weyrminus1

In Bhutan our results indicate a stronger mass loss forsouthern glaciers than for northern ones Interestingly in thisarea Kaumlaumlb (2005) measured high speeds (up to 200 m yrminus1)

for glaciers north of the Himalayan main ridge and ratherlow velocities over southern glacier tongues by using repeat

ASTER data This is consistent with the more negative massbalance that we report for the southern glaciers in Bhutanie for the slow flowing presumably downwasting glaciers

Berthier et al (2007) reported a mass balance betweenminus069 andminus085 m we yrminus1 for an ice-covered area of915 km2 around Chhota Shigri Glacier between 1999 (SRTMDEM) and 2004 (SPOT5-HRG DEM) For the same areawe found a mass balance ofminus045plusmn 013 m we yrminus1 over1999ndash2011 The difference in the study period may explainpart of the differences Also the methodology by Berthieret al (2007) did not explicitly take into account the SRTMradar penetration over glaciers and also corrected an appar-ent elevation-dependent bias according to glacier elevationswhich is now known as a resolution issue and is best cor-rected according to the terrain maximum curvature (Gardelleet al 2012b) More details and discussions about these dif-ferences can be found in Vincent et al (2013)

Importantly the results of the eastern Karakoram studysite confirm the balanced or slightly positive mass budgetreported by Gardelle et al (2012a) in the western Karako-ram over the past decade The two Karakoram study sitesare adjacent with a small overlapping area (sim 1100 km2 ofglaciers) where elevation differences agree within 1 m Both

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1280 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 12 Comparison of geodetic mass balances for 10 glaciers inthe Mount Everest area 1999ndash2010 mass balances measured inthe present study are compared with published estimates during2002ndash2007 (Bolch et al 2011) and during 2000ndash2008 (Nuimuraet al 2012) The 1-to-1 line is drawn The mean difference (plusmn1standard deviation) between (i) our values and Bolch et al (2011)are (047plusmn 058 m we yrminus1) and (ii) our values and Nuimura etal (2012) are (012plusmn 021 m we yrminus1) Part of the differences be-tween estimates may be due to the different periods surveyed andthe different glacier outlines used

sites show similar mass balances over slightly different pe-riods +011plusmn 014 m yrminus1 over 1999ndash2010 in the east and+009plusmn 018 m yrminus1 over 1999ndash2008 in the west This con-firms a ldquoKarakoram anomalyrdquo that was first suggested basedon field observations (Hewitt 2005) We report a mass bal-ance of+010plusmn 013 m yrminus1 for Baltoro Glacier (see loca-tion in Fig 8 size= 304 km2) between 1999 and 2010which is consistent with its gradual speed-up reported since2003 (Quincey et al 2009) However given the completelydifferent methods used in our and the latter study we can-not conclude that this demonstrates a real shift from nega-tive to positive mass balance The mass budget of SiachenGlacier (1145 km2 sometimes referred to as the highestbattlefield in the world) is also balanced or slightly pos-itive (+014plusmn 014 m we yrminus1) Thus the alarming state-ment by Rasul (2012) that this glacier retreated by 59 kmand lost 17 of its mass during 1989ndash2009 is very likelyerroneous Again within error bars our regional mass bal-ance in the Karakoram agrees with the mass balance ofminus003plusmn 004 m yrminus1 found by Kaumlaumlb et al (2012) over 2003ndash2008 and with the mass balance ofminus010plusmn 018 m we yrminus1

(2-sigma error level) obtained by Gardner et al (2013) for aregion including both the Karakoram and the Hindu Kush

In the northernmost part of the Pamir in the Alay moun-tain range the mass balance of Abramov Glacier has beenmeasured in the field for 30 yr between 1968 and 1998(WGMS 2012) with a mean value ofminus046 m yrminus1 Herewe report a mass balance ofminus003plusmn 014 m yrminus1 for this

glacier over 1999ndash2011 Our near zero mass budget estimatefor this glacier disagrees with negative mass balances recon-structed for the same period with a model run using biascorrected NCEPNCAR re-analysis data and calibrated withfield data (Barundun et al 2013) An underestimate of ourSRTM C-Band penetration in the Pamir as a whole and inparticular for the small Abramov Glacier may contribute tothese differences

For the whole Pamir study site our mass budget is bal-anced or slightly positive (+014plusmn 013 m we yrminus1) In theeastern Pamir the mass balance of Muztag Ata Glacier was+025 m we yrminus1 between 2005 and 2010 (Yao et al 2012)Although Muztag Ata Glacier is located outside of our Pamirstudy site and his glaciological records only cover the sec-ond half of our study period this measurement is consistentwith our remotely sensed mass balance Thus the ldquoKarako-ram anomalyrdquo should probably be renamed the ldquoPamir-Karakoram anomalyrdquo

This slightly positive or balanced mass budget measuredin the Pamir seems to contradict previous findings by Heidand Kaumlaumlb (2012) They reported rapidly decreasing glacierspeeds in this region between two snapshots of annual ve-locities in 20002001 and 20092010 an observation thatwould fit with negative mass balances (Span and Kuhn 2003Azam et al 2012) We note though that the relationship be-tween glacier mass balance and ice fluxes is not straightfor-ward and might in particular involve a time delay betweena change in mass balance and the related dynamic response(Heid and Kaumlaumlb 2012) Thus the decrease in speed couldwell be due to negative mass balances before or partiallybefore 1999ndash2011 We acknowledge that this discrepancyneeds to be investigated further in particular by examiningcontinuous changes in annual glacier speed through the firstdecade of the 21st Century Indeed Heid and Kaumlaumlb (2012)measured differences between two annual periods which maynot represent mean decadal velocity changes

Jacob et al (2012) using satellite gravimetry (GRACE)measured a mass budget ofminus5plusmn 6 Gt yrminus1 over 2003ndash2010for the ldquoKarakoram-Himalayardquo their region 8c which cor-responds to our study area excluding the Pamir Karakoramand Hindu Kush sub-regions but including some glaciersnorth of the Himalayan ridge already on the Tibetan PlateauFor this subset of our PKH study area we measured amass budget ofminus126plusmn 42 Gt yrminus1 The two estimates over-lap within their error bars especially if our error is dou-bled to match the two standard error level used by Jacob etal (2012) The estimate of Jacob et al (2012) over our com-plete study region (PKH) ieminus6plusmn 8 Gt yrminus1 (sum of theirregions 8b and 8c) also includes mass changes for the KunlunShan sub-region for which we did not measure glacier massbalances Therefore the comparison with our PKH value(minus101plusmn 55 Gt yrminus1) is not straightforward but suggests areasonable agreement especially if we consider the slightmass gain (15plusmn 17 Gt yrminus1) measured with ICESat in theWest Kunlun (Gardner et al 2013)

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1281

Table 6 Annual average of the glacier seasonal and decadal mass loss contributions to Indus Ganges and Brahmaputra discharges Basinarea and seasonal contributions are given by Kaser et al (2010) glacier area in each basin is derived from the RGI v20 (Arendt et al 2012)Numbers in parentheses indicate the percentage of glacierized area within the river basin Seasonal contributions were modeled by Kaser etal (2010) based on the CRU 20 dataset over 1961ndash1990 assuming a balanced glacier mass budget Glacier decadal mass loss contributionsare given as annual average over 1999ndash2011

River Basin area Glacier area Glacier contribution Mean discharge(km2) (km2) (m3 sminus1) (m3 sminus1)

Seasonally delayed Decadal mass lossKaser et al (2010) (this study)

Indus 1 139 814 25 598 (2 ) 105 103plusmn 72 2146 (Chevallier personalcommunication 2012)

Ganges 1 023 609 11 168 (1 ) 47 103plusmn 49 11739 Papa et al (2012Brahmaputra 527 666 15 296 (3 ) 33 147plusmn 64 19710 updated)

54 Reasons for the heterogeneous pattern of massbalance

The PKH glacier mass balance is slightly negative(minus014plusmn 008 m we yrminus1) but changes are not homoge-neous from one sub-region to another and seem to coincidewith the climatic settings of the mountain range For the east-ern sites (from the Hengduan Shan to West Nepal) which areunder the influence of the Indian and south-east Asian sum-mer monsoons the mass balances are moderately negative(sim minus026 m we yrminus1) In the northwest of the PKH at thePamir and Karakoram sites where glacier climate is char-acterized by the westerlies in winter mass balances are bal-anced or slightly positive (sim +010 m we yrminus1) In betweenthese two influences the glaciers of the Spiti Lahaul sitein a monsoon-arid transition zone experienced the strongestmass loss (minus045plusmn 013 m we yrminus1)

This heterogeneous pattern of mass balances seems to berelated to trends in precipitation throughout the PKH Archerand Fowler (2004) reported an increase in winter precipi-tation over the Upper Indus Basin since 1961 a trend con-firmed by Yao et al (2012) over the Pamir and the Karako-ram based on the analysis of the Global Precipitation Cli-matology Project data set (GPCP Adler et al 2003) Thisprecipitation increase is a source of greater accumulation forglaciers and thus a possible explanation for their slightly pos-itive mass budgets In addition in the Upper Indus Basin themean summer temperature has been decreasing since 1961(Fowler and Archer 2006) which is consistent with the find-ings of Shekhar et al (2010) in the Karakoram who reporteda decrease in both maximum and minimum temperature since1988 These increases in winter precipitation and summertemperature likely translate in greater accumulation and re-duced ablation changes which are consistent with the bal-anced or slightly positive glacier mass budget

On the other hand in the east of the PKH the Indiansummer monsoon has been weakening since the 1950s (Bol-lasina et al 2011) which could have reduced the amount

of snowfall over the eastern study sites and thus resultedin negative mass balances In addition maximum and min-imum temperatures increased since 1988 in the western Hi-malaya (Shekhar et al 2010) and maximum temperaturesincreased in the central Himalaya (Nepal) since the mid-1970s (Shrestha et al 1999) Thus both precipitation andtemperature trends in the Himalaya likely contributed to thenegative mass balances measured over the correspondingstudy sites (from Spiti Lahaul to Hengduan Shan Fig 1)

However gridded data sets such as GPCP are not neces-sarily representative of the climate at high altitude IndeedHewitt (2005) reported a 5-to-10-fold increase in precipi-tation between the glacier front and the accumulation areain the Karakoram that is not captured by reanalysis dataas recently confirmed by Immerzeel et al (2012) Thus di-rect measurements of accumulation on High Mountain Asiaglaciers (eg Azam et al 2013) and high-altitude weatherstations are eagerly needed to validate the large scale grid-ded data set (eg reanalysis) test their ability to describe thespecific climate near to PKH glaciers and assess the relativerole played by temperature and precipitation changes

55 Contribution to water resources

Our glacier mass balance can be averaged over large riverbasins to estimate the mean contribution to river runoffdue to a decade of glacier mass loss We assume therebyonly direct river runoff without loss terms For the threerivers for which we cover the entire glacierized area of theircatchments within our sub-regions (Fig 1) these contribu-tions to the river discharge are equal to 103 m3 sminus1 (Indus)103 m3 sminus1 (Ganges) and 147 m3 sminus1 (Brahmaputra) for theperiod 1999ndash2011 (Table 6)

Rivers of PKH being key water resources measures oftheir discharges are highly sensitive and difficult to access forpolitical reasons Papa et al (2012) built recent time series ofdischarges for Ganges and Brahmaputra by using altimetrymeasurements The locations of the corresponding dischargeestimates in Bangladesh are given in Fig 1 We can therefore

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1282 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

evaluate the relative contribution of glacier decadal mass loss(Table 6) to annual river run-off at 09 for the Gangesat Hardinge (basin area ofsim 1 020 000 km2) and 07 forBrahmaputra at Bahadurabad (basin area ofsim 530 000 km2)between 1999 and 2011 Given the discharge measured at theGuddu station by the Surface Water Hydrology Project of theWater and Power Development Authority (Pakistan see lo-cation on Fig 1) between 1999 and 2003 we can estimatethe contribution of glacier mass loss at 48 for the IndusRiver at this station

The seasonal contributions of glaciers to river run-off (iethe part of the annual precipitation that experiences season-ally delayed release from the glaciers) directly results fromseasonal variations of meteorological quantities in contrastto the glacier mass loss contribution which results from thelong-term climatic change Both runoff components directlyaffect the amount of water available in a river at a given lo-cation and at a given time so that their comparison could beof interest Even if the seasonal timing of the glacier massloss contribution is unclear it could in general peak at asimilar time of the year than the seasonal glacier contribu-tion so that both components rather enlarge each other thancancel Seasonal glacier contributions have been modeled byKaser et al (2010) The latter study assumed glaciers to bein equilibrium with the present climate and thus prescribedglacier mass balances equal to 0 Thus their seasonally de-layed glacier contributions do not include the part due todecadal glacier mass loss and is complementary to our es-timates (Table 6) For the eastern basins under the influenceof the Indian summer monsoon (Ganges and Brahmaputra)the contribution of glacier mass loss is higher than the sea-sonal contribution In other words for the first decade of the21st century the mass loss of glaciers is on average moreimportant for water availability in those two rivers than theirseasonal water storage effect The situation is different for theIndus Basin where the glacier melt season is also the dry sea-son which results in a relatively higher seasonally delayedglacier contribution to the river discharge There the season-ally delayed water release from the glaciers and the decadalglacier mass loss contribute on average equally to the riverdischarge Both contributions are strongly (and equally) de-pendent on the location of the discharge measurement andwill be significantly larger (in relative term) in more heav-ily glacierized and smaller mountain catchments located inthe upper part of these major river basins (Bookhagen andBurbank 2010)

6 Conclusion

In this study we assessed the spatial pattern of glaciermass balance over the PKH by comparing the SRTM andSPOT5 DEMs over 9 well-distributed study sites between1999 and 2011 We found slightly positive or balanced massbudgets in the western part (Pamir Karakoram) moder-ate mass loss in the east (Nepal Bhutan Hengduan Shan)

and the most negative mass balance in the monsoon-aridtransition zone (Spiti Lahaul in the western Himalaya)By extrapolating these values to climatically homogeneoussub-regions we estimate the PKH overall mass balanceto have beenminus014plusmn 008 m we yrminus1 between 1999 and2011 which corresponds to a contribution to sea level riseof 0028plusmn 0015 mm yrminus1 This is about 4 of the globalglacier mass loss estimated by Gardner et al (2013) thoughover a slightly shorter time period (2003ndash2009) The largestsource of uncertainties in those geodetic mass balance esti-mates is the poorly constrained penetration of the C-BandSRTM radar signal into snow and ice

Our technique of DEM differencing allows mapping in de-tail the spatial pattern of glacier elevation changes Thosemaps reveal many surge-type behaviors both in the Pamirand the Karakoram It also appears that the glacier-wide massbalance does not seem to be considerably affected by themass transfer from surges over thesim 10 yr time period stud-ied Similar lowering rates over debris-covered and clean iceare observed for four study sites despite the commonly as-sumed insulating effect of debris covers This suggests forthe scale of entire glacier tongues a spatial mixture of re-duced ablation under intact thick debris covers and enhancedablation by supraglacial lakes or ice cliffs or due to thin de-bris layer and very low glacier velocities in ablation areasthat reduce the ice advection downstream

The regionally heterogeneous pattern of ice wastage is inbroad agreement with contrasted trends in large-scale pre-cipitation and temperature However glaciological (eg ac-cumulation) and meteorological records are needed at highaltitude in the vicinity of glaciers to evaluate more closelythe local influence of atmospheric variables on glacier massbalance and fully understand the reasons behind those con-trasted mass budgets in the PKH

Supplementary material related to this article isavailable online at httpwwwthe-cryospherenet712632013tc-7-1263-2013-supplementpdf

Acknowledgements We thank G Cogley A Gardner T NuimuraT Bolch P Wagnon M Barandun M Hoelzle M Huss and ananonymous referee for their constructive comments that led to aconsiderably improved manuscript We thank the Water and PowerDevelopment Authority (WAPDA) for their hydrological data(processed by P Chevallier) and F Papa for sharing his updatedrun-off data ahead of publication We thank T Bolch for sharinghis glacier outlines from the Everest area We thank the USGSfor allowing free access to their Landsat archive CGIAR-SCI forSRTM C-band data DLR for SRTM X-band data and GLIMS forglacier inventories J Gardelle acknowledges a PhD fellowshipfrom the French Space Agency (CNES) and the French NationalResearch Center (CNRS) E Berthier acknowledges support fromCNES through the TOSCA and ISIS proposal no 397 and fromthe Programme National de Teacuteleacutedeacutetection Spatiale E Berthieracknowledges M Masiokas and R Villalba for welcoming him

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1283

at the IANIGLA (Mendoza Argentina) Y Arnaud acknowledgessupport from French National Research Agency through ANR-09-CEP-005-01PAPRIKA A Kaumlaumlb acknowledges support fromESA through Glaciers_cci (400010177810I-AM) and from theEuropean Research Council under the European Unionrsquos SeventhFramework Programme (FP2007-2013)ERC Grant Agreementn 320816

Edited by I M Howat

The publication of this articleis financed by CNRS-INSU

References

Adler R Huffman G Chang A Ferraro R Xie P JanowiakJ Rudolf B Schneider U Curtis S Bolvin D Gruber ASusskind J Arkin P and Nelkin E The Version-2 GlobalPrecipitation Climatology Project (GPCP) Monthly Precipita-tion Analysis (1979ndashPresent) J Hydrometeorol 4 1147ndash11672003

Archer D R and Fowler H J Spatial and temporal variationsin precipitation in the Upper Indus Basin global teleconnectionsand hydrological implications Hydrol Earth Syst Sci 8 47ndash61doi105194hess-8-47-2004 2004

Arendt A Bolch T Cogley J G Gardner A Hagen J-OHock R Kaser G Pfeffer W T Moholdt G Paul F RadicV Andreassen L Bajracharya S Beedle M Berthier EBhambri R Bliss A Brown I Burgess E Burgess DCawkwell F Chinn T Copland L Davies B de AngelisH Dolgova E Filbert K Forester R Fountain A Frey HGiffen B Glasser N Gurney S Hagg W Hall D Hari-tashya U K Hartmann G Helm C Herreid S Howat IKapustin G Khromova T Kienholz C Koenig M KohlerJ Kriegel D Kutuzov S Lavrentiev I LeBris R Lund JManley W Mayer C Miles E Li X Menounos B Mer-cer A Moelg N Mool P Nosenko G Negrete A NuthC Pettersson R Racoviteanu A Ranzi R Rastner P RauF Rich J Rott H Schneider C Seliverstov Y Sharp MSigurethsson O Stokes C Wheate R Winsvold S WolkenG Wyatt F and Zheltyhina N Randolph Glacier Inventory[v20] A Dataset of Global Glacier Outlines Global Land IceMeasurements from Space Boulder Colorado USA Digital Me-dia httpwwwglimsorgRGIrandolphhtml 2012

Azam M Wagnon P Ramanathan A Vincent C Sharma PArnaud Y Linda A Pottakkal J Chevallier P Singh V andBerthier E From balance to imbalance a shift in the dynamicbehaviour of Chhota Shigri glacier western Himalaya India JGlaciol 58 315ndash324 doi1031892012JoG11J123 2012

Azam M F Wagnon P Vincent C Ramanathan A Linda Aand Singh V B Reconstruction of the annual mass balanceof Chhota Shigri Glacier (Western Himalaya India) since 1969Ann Glaciol in revision 2013

Barrand N and Murray T Multivariate Controls on the In-cidence of Glacier Surging in the Karakoram Himalaya

Arct Antarct Alp Res 38 489ndash498 doi1016571523-0430(2006)38[489MCOTIO]20CO2 2006

Barundun M Huss M Azisov E Gafurov A HoelzleM Merkushkin A Salzmann N and Usubaliev R Re-establishing seasonal mass balance observation at AbramovGlacier Kyrgyzstan from 1968ndash2012 Abstract EGU2013-5668European Geophysical Union Vienna 2013

Benn D Bolch T Hands K Gulley J Luckman ANicholson L Quincey D Thompson S Toumi R andWiseman S Response of debris-covered glaciers in theMount Everest region to recent warming and implicationsfor outburst flood hazards Earth-Sci Rev 114 156ndash174doi101016jearscirev201203008 2012

Berthier E and Vincent C Relative contribution of surface mass-balance and ice-flux changes to the accelerated thinning of Merde Glace French Alps over 1979ndash2008 J Glaciol 58 501ndash512doi1031892012JoG11J083 2012

Berthier E Arnaud Y Baratoux D Vincent C and Reacutemy FRecent rapid thinning of the ldquo Mer de Glace rdquo glacier derivedfrom satellite optical images Geophys Res Lett 31 L17401doi1010292004GL020706 2004

Berthier E Arnaud Y Kumar R Ahmad S Wagnon P andChevallier P Remote sensing estimates of glacier mass balancesin the Himachal Pradesh (Western Himalaya India) RemoteSens Environ 108 327ndash338 doi101016jrse2006110172007

Bhambri R Bolch T Kawishwar P Dobhal D P SrivastavaD and Pratap B Heterogeneity in Glacier response from 1973to 2011 in the Shyok valley Karakoram India The CryosphereDiscuss 6 3049ndash3078 doi105194tcd-6-3049-2012 2012

Bhutiyani M Mass-balance studies on Siachen Glacier in theNubra valley Karakoram Himalaya India J Glaciol 45 112ndash118 1999

Bishop M P Schroder J F and Ward J L SPOT multispec-tral analysis for producing supraglacial debris-load estimates forBatura glacier Pakistan Geocarto Int 10 81ndash90 1995

Bolch T Pieczonka T and Benn D I Multi-decadal mass lossof glaciers in the Everest area (Nepal Himalaya) derived fromstereo imagery The Cryosphere 5 349ndash358 doi105194tc-5-349-2011 2011

Bolch T Kulkarni A Kaumlaumlb A Huggel C Paul F Cogley JFrey H Kargel J Fujita K Scheel M Bajracharya S andStoffel M The state and fate of Himalayan Glaciers Science336 310ndash314 doi101126science1215828 2012

Bollasina M Ming Y and Ramaswamy V Anthropogenicaerosols and the weakening of the South Asian summer mon-soon Science 334 502ndash505 doi101126science12049942011

Bookhagen B and Burbank D Toward a complete Himalayan hy-drological budget spatiotemporal distribution of snowmelt andrainfall and their impact on river discharge J Geophys Res115 F03019 doi1010292009JF001426 2010

Bretherton C Widmann M Dymnikov V Wallace J and BladeacuteI The effective number of spatial degrees of freedom of a time-varying field J Climate 12 1990ndash2009 1999

Copland L Pope S Bishop M Shroder J Clendon PBush A Kamp U Seong Y and Owen L Glacier veloc-ities across the central Karakoram Ann Glaciol 50 41ndash49doi103189172756409789624229 2009

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1275

Fig 8Same as Fig 2 but for the Karakoram east study site Surge-type glaciers in an active surge phase (resp quiescent phase) are identifiedwith a solid triangle (resp solid circle)

are potentially the result of glacier-wide processes Over aglacier tongue the ablation can be enhanced by the pres-ence of steep exposed ice cliffs or supraglacial lakes andponds (Sakai et al 2000 2002) In addition it is also pos-sible that the glacier tongues are mostly covered by thindebris layers such that the albedo effect dominates the in-sulating effect resulting in an increased tongue-wide abla-tion The prevalence of thin debris was recently suggestedfor debris-covered glaciers around the Mount Gongga in thesouth-eastern Tibetan Plateau (Zhang et al 2013b) Finallyice-flow rates could be very low at debris-covered parts asshown by Quincey et al (2007) in the Everest area Kaumlaumlb(2005) for Bhutan and Scherler et al (2011b) in the HinduKush-Karakoram-Himalaya Thus all three factors are likelyto increase the overall thinning under debris despite the un-doubted protective effect of a continuous thick debris coverat a local scale Future investigations combining surface ve-locity fields and maps of elevation changes could allow eval-uating more closely the relative role of ablation and ice fluxesin thinning rates over PKH glacier tongues (eg Berthierand Vincent 2012 Nuth et al 2012) Measurements of the

spatial distribution of debris thickness over the PKH glaciertongues are also needed

Note that in the case of debris-covered tongues the ele-vation change measurement represents the glacier thicknesschange but also the possible debris thickness evolution Thelatter remains poorly known in the PKH or in any othermountain range But based on the debris discharges givenby Bishop et al (1995) for Batura Glacier (Pakistan) 48to 90times 103 m3 yrminus1 the thickening of the debris can be es-timated between 27 and 55 mm yrminus1 which is negligiblegiven the magnitude of the glacier elevation changes

53 Comparison to previous mass balance estimates

Throughout the PKH there is only a single peer-reviewedglaciological record (on Chhota Shigri Glacier) coveringmost of our study period (Wagnon et al 2007 Vincent etal 2013) Our geodetic mass balance estimate for the SpitiLahaul study site has been recently compared to those glacio-logical measurements on Chhota Shigri Glacier (Vincent etal 2013) Mass balance records reported in Yao et al (2012)for the Tibetan Plateau and the surroundings mountain ranges

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1276 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 9Same as Fig 8 but for the Karakoram west study site

only start during glaciological year 20052006 This is whyhere we do not compare our mass balance estimates toglaciological records and limit our comparison to others re-gional estimates obtained using the geodetic or gravimetricmethods

We stress that the comparison of our mass balance mea-surements to published estimates is complicated by thefact that generally they do not cover the same time pe-riod Little is known about the inter-annual variability inthe PKH during the 21st century due to the limited num-ber of long glaciological time series The 9 yr mass balance

record on Chhota Shigri Glacier reveals a relatively highinter-annual variability of the annual mass balance (stan-dard deviationplusmn 074 m we yrminus1 during 2003ndash2011 Vin-cent et al 2013) which is similar to the one obtained be-tween 1968 and 1998 on Abramov Glacier (standard devi-ation 069 m we yrminus1 WGMS 2012) Thus inter-annualvariability can explain part of the difference between two cu-mulative mass balance estimates over 5ndash10 yr even if they areoffset by only one or two years

Based on ICESat repeat track measurements and theSRTM DEM Kaumlaumlb et al (2012) measured a region-wide

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1277

Fig 10Same as Fig 8 but for the Pamir study site

mass balance ofminus021plusmn 005 m we yrminus1 between 2003and 2008 for Hindu Kush-Karakoram-Himalaya glaciersFor that same area (ie excluding the Pamir and theHengduan Shan study sites) we found a mass balance ofminus015plusmn 007 m we yrminus1 Thus our result agrees with Kaumlaumlbet al (2012) within the error bars although our study periodis longer (1999ndash2011) and our measurement principle andextrapolation approach are different In addition we havecalculated updated mass balances from Kaumlaumlb et al (2012)ie averages over 3times 3 degree cells centered over our studysites and find that they are consistent within error barswith our mass balance estimates (Table 5) Similar resultsare found when our mass balances are compared to a spa-tially extended ICESat analysis for 2003ndash2009 (Gardner etal 2013) Mass losses measured with ICESat (Kaumlaumlb et al2012 Gardner et al 2013) tend to be slightly larger thanours This could be due to the difference in time periodsAlso our smaller mass losses could be partly explained if wesystematically underestimated the poorly constrained pene-

tration of the C-band andor X-band radar signal into ice andsnow

At a more local scale Bolch et al (2011) reported a massbalance ofminus079plusmn 052 m yrminus1 between 2002 and 2007 byDEM differencing over a glacier area of 62 km2 includingten glaciers south and west of Mt Everest For these sameglaciers Nuimura et al (2012) found a mean mass balanceof minus045plusmn 060 m yrminus1 during 2000ndash2008 while they com-puted a mass balance ofminus040plusmn 025 m yrminus1 between 1992and 2008 over a glacier area of 183 km2 around Mt EverestIn the present study we found an average mass balance ofminus041plusmn 021 m yrminus1 over the ten glaciers mentioned above(Table 5 Fig S1) and a value ofminus026plusmn 013 m yrminus1 overthe whole Everest study site between 2000 and 2011 Ourvalues are comparable to those of Nuimura et al (2012)Our mass balance is less negative than the 2002ndash2007 massbalance of Bolch et al (2011) but not statistically differentwhen error bars are considered (Fig 12) Indeed Bolch etal (2011) acknowledged some high uncertainties for their

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1278 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Table 4Glacier mass budget of the nine sub-regions to which the results of the study sites (mass balances) have been extrapolated For eachof them the total glacierized area as well as the percentage of the glacier area over which the mass balance was computed (ie study sites)are given The total glacier area is derived from the RGI v20 (Arendt et al 2012) The error for the mass balance in each sub-region is theroot of sum of squares (RSS) of the mass balance error of each study site and the regional extrapolation error The error for the mass budgetin each sub-region includes additionally the error from the total glacier area in the RGI v20

Sub-region Total glacier Measured glacier Mass balance Mass budgetarea (km2) area () (m we yrminus1) (Gt yrminus1)

Hengduan Shan 11584 12 minus033plusmn 014 minus38plusmn 38Bhutan 4021 34 minus022plusmn 013 minus09plusmn 05Everest 6226 23 minus026plusmn 014 minus16plusmn 08West Nepal 6849 13 minus032plusmn 014 minus22plusmn 10Spiti Lahaul 9043 23 minus045plusmn 014 minus41plusmn 13Hindu Kush 6135 13 minus012plusmn 016 minus07plusmn 10Karakoram East

1902428 +011 plusmn 014 +19 plusmn 31

Karakoram West 30 +009 plusmn 018Pamir 9369 34 +014 plusmn 014 +13 plusmn 13

TotalArea- 72251 30 minus014plusmn 008 minus101plusmn 55weighted average

Table 5Comparison of geodetic mass balances between this study and previous published results with overlapping study periods and similargeographic locations Updated figures from Kaumlaumlb et al (2012) are averaged over 3 times 3 cells centered over the study sites of the presentstudy

GlacierSite name This study Bolch et al (2011) Nuimura et al (2012) Kaumlaumlb et al (2012 updated)2002ndash2007 2000ndash2008 2003ndash2008

Study Mass balance Area (km2)a Mass balance Area Mass balance Mass balanceperiod (m we yrminus1) (m we yrminus1) (km2) (m we yrminus1) (m we yrminus1)

Hengduan San 1999ndash2010 minus022 plusmn 014 1303 (55 )Bhutan 1999ndash2010 minus022 plusmn 012 1384 (64 ) minus052 plusmn 016b

Everest

1999ndash2011

minus026 plusmn 013 1461 (58 ) minus039 plusmn 011AX010 minus090plusmn 034 04Changri SharNup minus042plusmn 017 161 (79 ) minus029plusmn 052 130 minus055plusmn 038Khumbu minus051plusmn 019 203 (47 ) minus045plusmn 052 170 minus076plusmn 052Nuptse minus037plusmn 020 50 (74 ) minus040plusmn 053 40 minus034plusmn 027Lhotse Nup minus021plusmn 027 24 (70 ) minus103plusmn 051 19 minus022plusmn 047Lhotse minus043plusmn 018 85 (83 ) minus110plusmn 052 65 minus067plusmn 051Lhotse SharImja minus070plusmn 052 98 (44 ) minus145plusmn 052 107 minus093plusmn 060Amphu Laptse minus046plusmn 034 25 (38 ) minus077plusmn 052 15 minus018plusmn 094Chukhung +044 plusmn 024 42 (47 ) +004 plusmn 054 38 +043 plusmn 081Ama Dablam minus049plusmn 017 36 (54 ) minus056plusmn 052 22 minus056plusmn 073Duwo minus016plusmn 026 19 (18 ) minus196plusmn 053 10 minus068plusmn 074Total Khumbu minus041plusmn 021 744 minus079plusmn 052 617 minus045plusmn 060(10 Glaciers above)

West Nepal 1999ndash2011 minus032 plusmn 013 908 (40 ) minus032 plusmn 012Spiti Lahaul 1999ndash2011 minus045 plusmn 013 2110 (46 ) minus038 plusmn 006Karakoram East 1999ndash2010 +011 plusmn 014 5328 (42 ) minus004 plusmn 004Karakoram West 1999ndash2008 +009 plusmn 018 5434 (45 )Hindu Kush 1999ndash2008 minus012 plusmn 016 793 (80 ) minus020 plusmn 006Pamir 1999ndash2011 +014 plusmn 013 3178 (50 )

a The total glacier area is given in km2 and in parenthesis the of the glacier area actually covered with measurementsb For this cell the ICESat coverage is insufficient and does not sample all glacier elevations

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1279

Fig 11 Elevation changes (SPOT5-SRTM) in the ablation area for clean ice (blue circles) and debris-covered ice (black circles) for eachstudy site For the Karakoram (east and west) and the Pamir study sites only non-surge-type glaciers are considered Standard error of theelevation differences are typically less thanplusmn2 m with each 100 m elevation band (not shown for the sake of clarity) Note elevation changesover separate sections of a glacier cannot be treated as mass changes due to the disregard of glacier dynamics

estimate over this short time period The differences in sur-vey periods and in the glacier outlines may also explainpart of these discrepancies In particular the delimitation ofthe accumulation area is notoriously difficult and Bolch etal (2011) did not include some of the steep high altitudeslopes which we included due to their potential avalanchecontribution For the 10 glaciers mentioned above our massbalances would become 012 m we yrminus1 more negative ifwe used the exact same outlines as Bolch et al (2011)Our positive mass balance for the small Chukhung Glacier(+044plusmn 024 m we yrminus1) is in agreement with Nuimura etal (2012) but enigmatic in a context of regional glacier lossand deserves further investigation If the 2002ndash2007 massbalance estimate by Bolch et al (2011) is excluded all otherrecent mass balance estimates suggest no acceleration of themass loss in the Everest area when compared to the long-term(1970ndash2007) mass balance measured by Bolch et al (2011)in this regionminus032plusmn 008 m weyrminus1

In Bhutan our results indicate a stronger mass loss forsouthern glaciers than for northern ones Interestingly in thisarea Kaumlaumlb (2005) measured high speeds (up to 200 m yrminus1)

for glaciers north of the Himalayan main ridge and ratherlow velocities over southern glacier tongues by using repeat

ASTER data This is consistent with the more negative massbalance that we report for the southern glaciers in Bhutanie for the slow flowing presumably downwasting glaciers

Berthier et al (2007) reported a mass balance betweenminus069 andminus085 m we yrminus1 for an ice-covered area of915 km2 around Chhota Shigri Glacier between 1999 (SRTMDEM) and 2004 (SPOT5-HRG DEM) For the same areawe found a mass balance ofminus045plusmn 013 m we yrminus1 over1999ndash2011 The difference in the study period may explainpart of the differences Also the methodology by Berthieret al (2007) did not explicitly take into account the SRTMradar penetration over glaciers and also corrected an appar-ent elevation-dependent bias according to glacier elevationswhich is now known as a resolution issue and is best cor-rected according to the terrain maximum curvature (Gardelleet al 2012b) More details and discussions about these dif-ferences can be found in Vincent et al (2013)

Importantly the results of the eastern Karakoram studysite confirm the balanced or slightly positive mass budgetreported by Gardelle et al (2012a) in the western Karako-ram over the past decade The two Karakoram study sitesare adjacent with a small overlapping area (sim 1100 km2 ofglaciers) where elevation differences agree within 1 m Both

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1280 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 12 Comparison of geodetic mass balances for 10 glaciers inthe Mount Everest area 1999ndash2010 mass balances measured inthe present study are compared with published estimates during2002ndash2007 (Bolch et al 2011) and during 2000ndash2008 (Nuimuraet al 2012) The 1-to-1 line is drawn The mean difference (plusmn1standard deviation) between (i) our values and Bolch et al (2011)are (047plusmn 058 m we yrminus1) and (ii) our values and Nuimura etal (2012) are (012plusmn 021 m we yrminus1) Part of the differences be-tween estimates may be due to the different periods surveyed andthe different glacier outlines used

sites show similar mass balances over slightly different pe-riods +011plusmn 014 m yrminus1 over 1999ndash2010 in the east and+009plusmn 018 m yrminus1 over 1999ndash2008 in the west This con-firms a ldquoKarakoram anomalyrdquo that was first suggested basedon field observations (Hewitt 2005) We report a mass bal-ance of+010plusmn 013 m yrminus1 for Baltoro Glacier (see loca-tion in Fig 8 size= 304 km2) between 1999 and 2010which is consistent with its gradual speed-up reported since2003 (Quincey et al 2009) However given the completelydifferent methods used in our and the latter study we can-not conclude that this demonstrates a real shift from nega-tive to positive mass balance The mass budget of SiachenGlacier (1145 km2 sometimes referred to as the highestbattlefield in the world) is also balanced or slightly pos-itive (+014plusmn 014 m we yrminus1) Thus the alarming state-ment by Rasul (2012) that this glacier retreated by 59 kmand lost 17 of its mass during 1989ndash2009 is very likelyerroneous Again within error bars our regional mass bal-ance in the Karakoram agrees with the mass balance ofminus003plusmn 004 m yrminus1 found by Kaumlaumlb et al (2012) over 2003ndash2008 and with the mass balance ofminus010plusmn 018 m we yrminus1

(2-sigma error level) obtained by Gardner et al (2013) for aregion including both the Karakoram and the Hindu Kush

In the northernmost part of the Pamir in the Alay moun-tain range the mass balance of Abramov Glacier has beenmeasured in the field for 30 yr between 1968 and 1998(WGMS 2012) with a mean value ofminus046 m yrminus1 Herewe report a mass balance ofminus003plusmn 014 m yrminus1 for this

glacier over 1999ndash2011 Our near zero mass budget estimatefor this glacier disagrees with negative mass balances recon-structed for the same period with a model run using biascorrected NCEPNCAR re-analysis data and calibrated withfield data (Barundun et al 2013) An underestimate of ourSRTM C-Band penetration in the Pamir as a whole and inparticular for the small Abramov Glacier may contribute tothese differences

For the whole Pamir study site our mass budget is bal-anced or slightly positive (+014plusmn 013 m we yrminus1) In theeastern Pamir the mass balance of Muztag Ata Glacier was+025 m we yrminus1 between 2005 and 2010 (Yao et al 2012)Although Muztag Ata Glacier is located outside of our Pamirstudy site and his glaciological records only cover the sec-ond half of our study period this measurement is consistentwith our remotely sensed mass balance Thus the ldquoKarako-ram anomalyrdquo should probably be renamed the ldquoPamir-Karakoram anomalyrdquo

This slightly positive or balanced mass budget measuredin the Pamir seems to contradict previous findings by Heidand Kaumlaumlb (2012) They reported rapidly decreasing glacierspeeds in this region between two snapshots of annual ve-locities in 20002001 and 20092010 an observation thatwould fit with negative mass balances (Span and Kuhn 2003Azam et al 2012) We note though that the relationship be-tween glacier mass balance and ice fluxes is not straightfor-ward and might in particular involve a time delay betweena change in mass balance and the related dynamic response(Heid and Kaumlaumlb 2012) Thus the decrease in speed couldwell be due to negative mass balances before or partiallybefore 1999ndash2011 We acknowledge that this discrepancyneeds to be investigated further in particular by examiningcontinuous changes in annual glacier speed through the firstdecade of the 21st Century Indeed Heid and Kaumlaumlb (2012)measured differences between two annual periods which maynot represent mean decadal velocity changes

Jacob et al (2012) using satellite gravimetry (GRACE)measured a mass budget ofminus5plusmn 6 Gt yrminus1 over 2003ndash2010for the ldquoKarakoram-Himalayardquo their region 8c which cor-responds to our study area excluding the Pamir Karakoramand Hindu Kush sub-regions but including some glaciersnorth of the Himalayan ridge already on the Tibetan PlateauFor this subset of our PKH study area we measured amass budget ofminus126plusmn 42 Gt yrminus1 The two estimates over-lap within their error bars especially if our error is dou-bled to match the two standard error level used by Jacob etal (2012) The estimate of Jacob et al (2012) over our com-plete study region (PKH) ieminus6plusmn 8 Gt yrminus1 (sum of theirregions 8b and 8c) also includes mass changes for the KunlunShan sub-region for which we did not measure glacier massbalances Therefore the comparison with our PKH value(minus101plusmn 55 Gt yrminus1) is not straightforward but suggests areasonable agreement especially if we consider the slightmass gain (15plusmn 17 Gt yrminus1) measured with ICESat in theWest Kunlun (Gardner et al 2013)

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1281

Table 6 Annual average of the glacier seasonal and decadal mass loss contributions to Indus Ganges and Brahmaputra discharges Basinarea and seasonal contributions are given by Kaser et al (2010) glacier area in each basin is derived from the RGI v20 (Arendt et al 2012)Numbers in parentheses indicate the percentage of glacierized area within the river basin Seasonal contributions were modeled by Kaser etal (2010) based on the CRU 20 dataset over 1961ndash1990 assuming a balanced glacier mass budget Glacier decadal mass loss contributionsare given as annual average over 1999ndash2011

River Basin area Glacier area Glacier contribution Mean discharge(km2) (km2) (m3 sminus1) (m3 sminus1)

Seasonally delayed Decadal mass lossKaser et al (2010) (this study)

Indus 1 139 814 25 598 (2 ) 105 103plusmn 72 2146 (Chevallier personalcommunication 2012)

Ganges 1 023 609 11 168 (1 ) 47 103plusmn 49 11739 Papa et al (2012Brahmaputra 527 666 15 296 (3 ) 33 147plusmn 64 19710 updated)

54 Reasons for the heterogeneous pattern of massbalance

The PKH glacier mass balance is slightly negative(minus014plusmn 008 m we yrminus1) but changes are not homoge-neous from one sub-region to another and seem to coincidewith the climatic settings of the mountain range For the east-ern sites (from the Hengduan Shan to West Nepal) which areunder the influence of the Indian and south-east Asian sum-mer monsoons the mass balances are moderately negative(sim minus026 m we yrminus1) In the northwest of the PKH at thePamir and Karakoram sites where glacier climate is char-acterized by the westerlies in winter mass balances are bal-anced or slightly positive (sim +010 m we yrminus1) In betweenthese two influences the glaciers of the Spiti Lahaul sitein a monsoon-arid transition zone experienced the strongestmass loss (minus045plusmn 013 m we yrminus1)

This heterogeneous pattern of mass balances seems to berelated to trends in precipitation throughout the PKH Archerand Fowler (2004) reported an increase in winter precipi-tation over the Upper Indus Basin since 1961 a trend con-firmed by Yao et al (2012) over the Pamir and the Karako-ram based on the analysis of the Global Precipitation Cli-matology Project data set (GPCP Adler et al 2003) Thisprecipitation increase is a source of greater accumulation forglaciers and thus a possible explanation for their slightly pos-itive mass budgets In addition in the Upper Indus Basin themean summer temperature has been decreasing since 1961(Fowler and Archer 2006) which is consistent with the find-ings of Shekhar et al (2010) in the Karakoram who reporteda decrease in both maximum and minimum temperature since1988 These increases in winter precipitation and summertemperature likely translate in greater accumulation and re-duced ablation changes which are consistent with the bal-anced or slightly positive glacier mass budget

On the other hand in the east of the PKH the Indiansummer monsoon has been weakening since the 1950s (Bol-lasina et al 2011) which could have reduced the amount

of snowfall over the eastern study sites and thus resultedin negative mass balances In addition maximum and min-imum temperatures increased since 1988 in the western Hi-malaya (Shekhar et al 2010) and maximum temperaturesincreased in the central Himalaya (Nepal) since the mid-1970s (Shrestha et al 1999) Thus both precipitation andtemperature trends in the Himalaya likely contributed to thenegative mass balances measured over the correspondingstudy sites (from Spiti Lahaul to Hengduan Shan Fig 1)

However gridded data sets such as GPCP are not neces-sarily representative of the climate at high altitude IndeedHewitt (2005) reported a 5-to-10-fold increase in precipi-tation between the glacier front and the accumulation areain the Karakoram that is not captured by reanalysis dataas recently confirmed by Immerzeel et al (2012) Thus di-rect measurements of accumulation on High Mountain Asiaglaciers (eg Azam et al 2013) and high-altitude weatherstations are eagerly needed to validate the large scale grid-ded data set (eg reanalysis) test their ability to describe thespecific climate near to PKH glaciers and assess the relativerole played by temperature and precipitation changes

55 Contribution to water resources

Our glacier mass balance can be averaged over large riverbasins to estimate the mean contribution to river runoffdue to a decade of glacier mass loss We assume therebyonly direct river runoff without loss terms For the threerivers for which we cover the entire glacierized area of theircatchments within our sub-regions (Fig 1) these contribu-tions to the river discharge are equal to 103 m3 sminus1 (Indus)103 m3 sminus1 (Ganges) and 147 m3 sminus1 (Brahmaputra) for theperiod 1999ndash2011 (Table 6)

Rivers of PKH being key water resources measures oftheir discharges are highly sensitive and difficult to access forpolitical reasons Papa et al (2012) built recent time series ofdischarges for Ganges and Brahmaputra by using altimetrymeasurements The locations of the corresponding dischargeestimates in Bangladesh are given in Fig 1 We can therefore

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1282 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

evaluate the relative contribution of glacier decadal mass loss(Table 6) to annual river run-off at 09 for the Gangesat Hardinge (basin area ofsim 1 020 000 km2) and 07 forBrahmaputra at Bahadurabad (basin area ofsim 530 000 km2)between 1999 and 2011 Given the discharge measured at theGuddu station by the Surface Water Hydrology Project of theWater and Power Development Authority (Pakistan see lo-cation on Fig 1) between 1999 and 2003 we can estimatethe contribution of glacier mass loss at 48 for the IndusRiver at this station

The seasonal contributions of glaciers to river run-off (iethe part of the annual precipitation that experiences season-ally delayed release from the glaciers) directly results fromseasonal variations of meteorological quantities in contrastto the glacier mass loss contribution which results from thelong-term climatic change Both runoff components directlyaffect the amount of water available in a river at a given lo-cation and at a given time so that their comparison could beof interest Even if the seasonal timing of the glacier massloss contribution is unclear it could in general peak at asimilar time of the year than the seasonal glacier contribu-tion so that both components rather enlarge each other thancancel Seasonal glacier contributions have been modeled byKaser et al (2010) The latter study assumed glaciers to bein equilibrium with the present climate and thus prescribedglacier mass balances equal to 0 Thus their seasonally de-layed glacier contributions do not include the part due todecadal glacier mass loss and is complementary to our es-timates (Table 6) For the eastern basins under the influenceof the Indian summer monsoon (Ganges and Brahmaputra)the contribution of glacier mass loss is higher than the sea-sonal contribution In other words for the first decade of the21st century the mass loss of glaciers is on average moreimportant for water availability in those two rivers than theirseasonal water storage effect The situation is different for theIndus Basin where the glacier melt season is also the dry sea-son which results in a relatively higher seasonally delayedglacier contribution to the river discharge There the season-ally delayed water release from the glaciers and the decadalglacier mass loss contribute on average equally to the riverdischarge Both contributions are strongly (and equally) de-pendent on the location of the discharge measurement andwill be significantly larger (in relative term) in more heav-ily glacierized and smaller mountain catchments located inthe upper part of these major river basins (Bookhagen andBurbank 2010)

6 Conclusion

In this study we assessed the spatial pattern of glaciermass balance over the PKH by comparing the SRTM andSPOT5 DEMs over 9 well-distributed study sites between1999 and 2011 We found slightly positive or balanced massbudgets in the western part (Pamir Karakoram) moder-ate mass loss in the east (Nepal Bhutan Hengduan Shan)

and the most negative mass balance in the monsoon-aridtransition zone (Spiti Lahaul in the western Himalaya)By extrapolating these values to climatically homogeneoussub-regions we estimate the PKH overall mass balanceto have beenminus014plusmn 008 m we yrminus1 between 1999 and2011 which corresponds to a contribution to sea level riseof 0028plusmn 0015 mm yrminus1 This is about 4 of the globalglacier mass loss estimated by Gardner et al (2013) thoughover a slightly shorter time period (2003ndash2009) The largestsource of uncertainties in those geodetic mass balance esti-mates is the poorly constrained penetration of the C-BandSRTM radar signal into snow and ice

Our technique of DEM differencing allows mapping in de-tail the spatial pattern of glacier elevation changes Thosemaps reveal many surge-type behaviors both in the Pamirand the Karakoram It also appears that the glacier-wide massbalance does not seem to be considerably affected by themass transfer from surges over thesim 10 yr time period stud-ied Similar lowering rates over debris-covered and clean iceare observed for four study sites despite the commonly as-sumed insulating effect of debris covers This suggests forthe scale of entire glacier tongues a spatial mixture of re-duced ablation under intact thick debris covers and enhancedablation by supraglacial lakes or ice cliffs or due to thin de-bris layer and very low glacier velocities in ablation areasthat reduce the ice advection downstream

The regionally heterogeneous pattern of ice wastage is inbroad agreement with contrasted trends in large-scale pre-cipitation and temperature However glaciological (eg ac-cumulation) and meteorological records are needed at highaltitude in the vicinity of glaciers to evaluate more closelythe local influence of atmospheric variables on glacier massbalance and fully understand the reasons behind those con-trasted mass budgets in the PKH

Supplementary material related to this article isavailable online at httpwwwthe-cryospherenet712632013tc-7-1263-2013-supplementpdf

Acknowledgements We thank G Cogley A Gardner T NuimuraT Bolch P Wagnon M Barandun M Hoelzle M Huss and ananonymous referee for their constructive comments that led to aconsiderably improved manuscript We thank the Water and PowerDevelopment Authority (WAPDA) for their hydrological data(processed by P Chevallier) and F Papa for sharing his updatedrun-off data ahead of publication We thank T Bolch for sharinghis glacier outlines from the Everest area We thank the USGSfor allowing free access to their Landsat archive CGIAR-SCI forSRTM C-band data DLR for SRTM X-band data and GLIMS forglacier inventories J Gardelle acknowledges a PhD fellowshipfrom the French Space Agency (CNES) and the French NationalResearch Center (CNRS) E Berthier acknowledges support fromCNES through the TOSCA and ISIS proposal no 397 and fromthe Programme National de Teacuteleacutedeacutetection Spatiale E Berthieracknowledges M Masiokas and R Villalba for welcoming him

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1283

at the IANIGLA (Mendoza Argentina) Y Arnaud acknowledgessupport from French National Research Agency through ANR-09-CEP-005-01PAPRIKA A Kaumlaumlb acknowledges support fromESA through Glaciers_cci (400010177810I-AM) and from theEuropean Research Council under the European Unionrsquos SeventhFramework Programme (FP2007-2013)ERC Grant Agreementn 320816

Edited by I M Howat

The publication of this articleis financed by CNRS-INSU

References

Adler R Huffman G Chang A Ferraro R Xie P JanowiakJ Rudolf B Schneider U Curtis S Bolvin D Gruber ASusskind J Arkin P and Nelkin E The Version-2 GlobalPrecipitation Climatology Project (GPCP) Monthly Precipita-tion Analysis (1979ndashPresent) J Hydrometeorol 4 1147ndash11672003

Archer D R and Fowler H J Spatial and temporal variationsin precipitation in the Upper Indus Basin global teleconnectionsand hydrological implications Hydrol Earth Syst Sci 8 47ndash61doi105194hess-8-47-2004 2004

Arendt A Bolch T Cogley J G Gardner A Hagen J-OHock R Kaser G Pfeffer W T Moholdt G Paul F RadicV Andreassen L Bajracharya S Beedle M Berthier EBhambri R Bliss A Brown I Burgess E Burgess DCawkwell F Chinn T Copland L Davies B de AngelisH Dolgova E Filbert K Forester R Fountain A Frey HGiffen B Glasser N Gurney S Hagg W Hall D Hari-tashya U K Hartmann G Helm C Herreid S Howat IKapustin G Khromova T Kienholz C Koenig M KohlerJ Kriegel D Kutuzov S Lavrentiev I LeBris R Lund JManley W Mayer C Miles E Li X Menounos B Mer-cer A Moelg N Mool P Nosenko G Negrete A NuthC Pettersson R Racoviteanu A Ranzi R Rastner P RauF Rich J Rott H Schneider C Seliverstov Y Sharp MSigurethsson O Stokes C Wheate R Winsvold S WolkenG Wyatt F and Zheltyhina N Randolph Glacier Inventory[v20] A Dataset of Global Glacier Outlines Global Land IceMeasurements from Space Boulder Colorado USA Digital Me-dia httpwwwglimsorgRGIrandolphhtml 2012

Azam M Wagnon P Ramanathan A Vincent C Sharma PArnaud Y Linda A Pottakkal J Chevallier P Singh V andBerthier E From balance to imbalance a shift in the dynamicbehaviour of Chhota Shigri glacier western Himalaya India JGlaciol 58 315ndash324 doi1031892012JoG11J123 2012

Azam M F Wagnon P Vincent C Ramanathan A Linda Aand Singh V B Reconstruction of the annual mass balanceof Chhota Shigri Glacier (Western Himalaya India) since 1969Ann Glaciol in revision 2013

Barrand N and Murray T Multivariate Controls on the In-cidence of Glacier Surging in the Karakoram Himalaya

Arct Antarct Alp Res 38 489ndash498 doi1016571523-0430(2006)38[489MCOTIO]20CO2 2006

Barundun M Huss M Azisov E Gafurov A HoelzleM Merkushkin A Salzmann N and Usubaliev R Re-establishing seasonal mass balance observation at AbramovGlacier Kyrgyzstan from 1968ndash2012 Abstract EGU2013-5668European Geophysical Union Vienna 2013

Benn D Bolch T Hands K Gulley J Luckman ANicholson L Quincey D Thompson S Toumi R andWiseman S Response of debris-covered glaciers in theMount Everest region to recent warming and implicationsfor outburst flood hazards Earth-Sci Rev 114 156ndash174doi101016jearscirev201203008 2012

Berthier E and Vincent C Relative contribution of surface mass-balance and ice-flux changes to the accelerated thinning of Merde Glace French Alps over 1979ndash2008 J Glaciol 58 501ndash512doi1031892012JoG11J083 2012

Berthier E Arnaud Y Baratoux D Vincent C and Reacutemy FRecent rapid thinning of the ldquo Mer de Glace rdquo glacier derivedfrom satellite optical images Geophys Res Lett 31 L17401doi1010292004GL020706 2004

Berthier E Arnaud Y Kumar R Ahmad S Wagnon P andChevallier P Remote sensing estimates of glacier mass balancesin the Himachal Pradesh (Western Himalaya India) RemoteSens Environ 108 327ndash338 doi101016jrse2006110172007

Bhambri R Bolch T Kawishwar P Dobhal D P SrivastavaD and Pratap B Heterogeneity in Glacier response from 1973to 2011 in the Shyok valley Karakoram India The CryosphereDiscuss 6 3049ndash3078 doi105194tcd-6-3049-2012 2012

Bhutiyani M Mass-balance studies on Siachen Glacier in theNubra valley Karakoram Himalaya India J Glaciol 45 112ndash118 1999

Bishop M P Schroder J F and Ward J L SPOT multispec-tral analysis for producing supraglacial debris-load estimates forBatura glacier Pakistan Geocarto Int 10 81ndash90 1995

Bolch T Pieczonka T and Benn D I Multi-decadal mass lossof glaciers in the Everest area (Nepal Himalaya) derived fromstereo imagery The Cryosphere 5 349ndash358 doi105194tc-5-349-2011 2011

Bolch T Kulkarni A Kaumlaumlb A Huggel C Paul F Cogley JFrey H Kargel J Fujita K Scheel M Bajracharya S andStoffel M The state and fate of Himalayan Glaciers Science336 310ndash314 doi101126science1215828 2012

Bollasina M Ming Y and Ramaswamy V Anthropogenicaerosols and the weakening of the South Asian summer mon-soon Science 334 502ndash505 doi101126science12049942011

Bookhagen B and Burbank D Toward a complete Himalayan hy-drological budget spatiotemporal distribution of snowmelt andrainfall and their impact on river discharge J Geophys Res115 F03019 doi1010292009JF001426 2010

Bretherton C Widmann M Dymnikov V Wallace J and BladeacuteI The effective number of spatial degrees of freedom of a time-varying field J Climate 12 1990ndash2009 1999

Copland L Pope S Bishop M Shroder J Clendon PBush A Kamp U Seong Y and Owen L Glacier veloc-ities across the central Karakoram Ann Glaciol 50 41ndash49doi103189172756409789624229 2009

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

1276 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 9Same as Fig 8 but for the Karakoram west study site

only start during glaciological year 20052006 This is whyhere we do not compare our mass balance estimates toglaciological records and limit our comparison to others re-gional estimates obtained using the geodetic or gravimetricmethods

We stress that the comparison of our mass balance mea-surements to published estimates is complicated by thefact that generally they do not cover the same time pe-riod Little is known about the inter-annual variability inthe PKH during the 21st century due to the limited num-ber of long glaciological time series The 9 yr mass balance

record on Chhota Shigri Glacier reveals a relatively highinter-annual variability of the annual mass balance (stan-dard deviationplusmn 074 m we yrminus1 during 2003ndash2011 Vin-cent et al 2013) which is similar to the one obtained be-tween 1968 and 1998 on Abramov Glacier (standard devi-ation 069 m we yrminus1 WGMS 2012) Thus inter-annualvariability can explain part of the difference between two cu-mulative mass balance estimates over 5ndash10 yr even if they areoffset by only one or two years

Based on ICESat repeat track measurements and theSRTM DEM Kaumlaumlb et al (2012) measured a region-wide

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1277

Fig 10Same as Fig 8 but for the Pamir study site

mass balance ofminus021plusmn 005 m we yrminus1 between 2003and 2008 for Hindu Kush-Karakoram-Himalaya glaciersFor that same area (ie excluding the Pamir and theHengduan Shan study sites) we found a mass balance ofminus015plusmn 007 m we yrminus1 Thus our result agrees with Kaumlaumlbet al (2012) within the error bars although our study periodis longer (1999ndash2011) and our measurement principle andextrapolation approach are different In addition we havecalculated updated mass balances from Kaumlaumlb et al (2012)ie averages over 3times 3 degree cells centered over our studysites and find that they are consistent within error barswith our mass balance estimates (Table 5) Similar resultsare found when our mass balances are compared to a spa-tially extended ICESat analysis for 2003ndash2009 (Gardner etal 2013) Mass losses measured with ICESat (Kaumlaumlb et al2012 Gardner et al 2013) tend to be slightly larger thanours This could be due to the difference in time periodsAlso our smaller mass losses could be partly explained if wesystematically underestimated the poorly constrained pene-

tration of the C-band andor X-band radar signal into ice andsnow

At a more local scale Bolch et al (2011) reported a massbalance ofminus079plusmn 052 m yrminus1 between 2002 and 2007 byDEM differencing over a glacier area of 62 km2 includingten glaciers south and west of Mt Everest For these sameglaciers Nuimura et al (2012) found a mean mass balanceof minus045plusmn 060 m yrminus1 during 2000ndash2008 while they com-puted a mass balance ofminus040plusmn 025 m yrminus1 between 1992and 2008 over a glacier area of 183 km2 around Mt EverestIn the present study we found an average mass balance ofminus041plusmn 021 m yrminus1 over the ten glaciers mentioned above(Table 5 Fig S1) and a value ofminus026plusmn 013 m yrminus1 overthe whole Everest study site between 2000 and 2011 Ourvalues are comparable to those of Nuimura et al (2012)Our mass balance is less negative than the 2002ndash2007 massbalance of Bolch et al (2011) but not statistically differentwhen error bars are considered (Fig 12) Indeed Bolch etal (2011) acknowledged some high uncertainties for their

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1278 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Table 4Glacier mass budget of the nine sub-regions to which the results of the study sites (mass balances) have been extrapolated For eachof them the total glacierized area as well as the percentage of the glacier area over which the mass balance was computed (ie study sites)are given The total glacier area is derived from the RGI v20 (Arendt et al 2012) The error for the mass balance in each sub-region is theroot of sum of squares (RSS) of the mass balance error of each study site and the regional extrapolation error The error for the mass budgetin each sub-region includes additionally the error from the total glacier area in the RGI v20

Sub-region Total glacier Measured glacier Mass balance Mass budgetarea (km2) area () (m we yrminus1) (Gt yrminus1)

Hengduan Shan 11584 12 minus033plusmn 014 minus38plusmn 38Bhutan 4021 34 minus022plusmn 013 minus09plusmn 05Everest 6226 23 minus026plusmn 014 minus16plusmn 08West Nepal 6849 13 minus032plusmn 014 minus22plusmn 10Spiti Lahaul 9043 23 minus045plusmn 014 minus41plusmn 13Hindu Kush 6135 13 minus012plusmn 016 minus07plusmn 10Karakoram East

1902428 +011 plusmn 014 +19 plusmn 31

Karakoram West 30 +009 plusmn 018Pamir 9369 34 +014 plusmn 014 +13 plusmn 13

TotalArea- 72251 30 minus014plusmn 008 minus101plusmn 55weighted average

Table 5Comparison of geodetic mass balances between this study and previous published results with overlapping study periods and similargeographic locations Updated figures from Kaumlaumlb et al (2012) are averaged over 3 times 3 cells centered over the study sites of the presentstudy

GlacierSite name This study Bolch et al (2011) Nuimura et al (2012) Kaumlaumlb et al (2012 updated)2002ndash2007 2000ndash2008 2003ndash2008

Study Mass balance Area (km2)a Mass balance Area Mass balance Mass balanceperiod (m we yrminus1) (m we yrminus1) (km2) (m we yrminus1) (m we yrminus1)

Hengduan San 1999ndash2010 minus022 plusmn 014 1303 (55 )Bhutan 1999ndash2010 minus022 plusmn 012 1384 (64 ) minus052 plusmn 016b

Everest

1999ndash2011

minus026 plusmn 013 1461 (58 ) minus039 plusmn 011AX010 minus090plusmn 034 04Changri SharNup minus042plusmn 017 161 (79 ) minus029plusmn 052 130 minus055plusmn 038Khumbu minus051plusmn 019 203 (47 ) minus045plusmn 052 170 minus076plusmn 052Nuptse minus037plusmn 020 50 (74 ) minus040plusmn 053 40 minus034plusmn 027Lhotse Nup minus021plusmn 027 24 (70 ) minus103plusmn 051 19 minus022plusmn 047Lhotse minus043plusmn 018 85 (83 ) minus110plusmn 052 65 minus067plusmn 051Lhotse SharImja minus070plusmn 052 98 (44 ) minus145plusmn 052 107 minus093plusmn 060Amphu Laptse minus046plusmn 034 25 (38 ) minus077plusmn 052 15 minus018plusmn 094Chukhung +044 plusmn 024 42 (47 ) +004 plusmn 054 38 +043 plusmn 081Ama Dablam minus049plusmn 017 36 (54 ) minus056plusmn 052 22 minus056plusmn 073Duwo minus016plusmn 026 19 (18 ) minus196plusmn 053 10 minus068plusmn 074Total Khumbu minus041plusmn 021 744 minus079plusmn 052 617 minus045plusmn 060(10 Glaciers above)

West Nepal 1999ndash2011 minus032 plusmn 013 908 (40 ) minus032 plusmn 012Spiti Lahaul 1999ndash2011 minus045 plusmn 013 2110 (46 ) minus038 plusmn 006Karakoram East 1999ndash2010 +011 plusmn 014 5328 (42 ) minus004 plusmn 004Karakoram West 1999ndash2008 +009 plusmn 018 5434 (45 )Hindu Kush 1999ndash2008 minus012 plusmn 016 793 (80 ) minus020 plusmn 006Pamir 1999ndash2011 +014 plusmn 013 3178 (50 )

a The total glacier area is given in km2 and in parenthesis the of the glacier area actually covered with measurementsb For this cell the ICESat coverage is insufficient and does not sample all glacier elevations

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1279

Fig 11 Elevation changes (SPOT5-SRTM) in the ablation area for clean ice (blue circles) and debris-covered ice (black circles) for eachstudy site For the Karakoram (east and west) and the Pamir study sites only non-surge-type glaciers are considered Standard error of theelevation differences are typically less thanplusmn2 m with each 100 m elevation band (not shown for the sake of clarity) Note elevation changesover separate sections of a glacier cannot be treated as mass changes due to the disregard of glacier dynamics

estimate over this short time period The differences in sur-vey periods and in the glacier outlines may also explainpart of these discrepancies In particular the delimitation ofthe accumulation area is notoriously difficult and Bolch etal (2011) did not include some of the steep high altitudeslopes which we included due to their potential avalanchecontribution For the 10 glaciers mentioned above our massbalances would become 012 m we yrminus1 more negative ifwe used the exact same outlines as Bolch et al (2011)Our positive mass balance for the small Chukhung Glacier(+044plusmn 024 m we yrminus1) is in agreement with Nuimura etal (2012) but enigmatic in a context of regional glacier lossand deserves further investigation If the 2002ndash2007 massbalance estimate by Bolch et al (2011) is excluded all otherrecent mass balance estimates suggest no acceleration of themass loss in the Everest area when compared to the long-term(1970ndash2007) mass balance measured by Bolch et al (2011)in this regionminus032plusmn 008 m weyrminus1

In Bhutan our results indicate a stronger mass loss forsouthern glaciers than for northern ones Interestingly in thisarea Kaumlaumlb (2005) measured high speeds (up to 200 m yrminus1)

for glaciers north of the Himalayan main ridge and ratherlow velocities over southern glacier tongues by using repeat

ASTER data This is consistent with the more negative massbalance that we report for the southern glaciers in Bhutanie for the slow flowing presumably downwasting glaciers

Berthier et al (2007) reported a mass balance betweenminus069 andminus085 m we yrminus1 for an ice-covered area of915 km2 around Chhota Shigri Glacier between 1999 (SRTMDEM) and 2004 (SPOT5-HRG DEM) For the same areawe found a mass balance ofminus045plusmn 013 m we yrminus1 over1999ndash2011 The difference in the study period may explainpart of the differences Also the methodology by Berthieret al (2007) did not explicitly take into account the SRTMradar penetration over glaciers and also corrected an appar-ent elevation-dependent bias according to glacier elevationswhich is now known as a resolution issue and is best cor-rected according to the terrain maximum curvature (Gardelleet al 2012b) More details and discussions about these dif-ferences can be found in Vincent et al (2013)

Importantly the results of the eastern Karakoram studysite confirm the balanced or slightly positive mass budgetreported by Gardelle et al (2012a) in the western Karako-ram over the past decade The two Karakoram study sitesare adjacent with a small overlapping area (sim 1100 km2 ofglaciers) where elevation differences agree within 1 m Both

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1280 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 12 Comparison of geodetic mass balances for 10 glaciers inthe Mount Everest area 1999ndash2010 mass balances measured inthe present study are compared with published estimates during2002ndash2007 (Bolch et al 2011) and during 2000ndash2008 (Nuimuraet al 2012) The 1-to-1 line is drawn The mean difference (plusmn1standard deviation) between (i) our values and Bolch et al (2011)are (047plusmn 058 m we yrminus1) and (ii) our values and Nuimura etal (2012) are (012plusmn 021 m we yrminus1) Part of the differences be-tween estimates may be due to the different periods surveyed andthe different glacier outlines used

sites show similar mass balances over slightly different pe-riods +011plusmn 014 m yrminus1 over 1999ndash2010 in the east and+009plusmn 018 m yrminus1 over 1999ndash2008 in the west This con-firms a ldquoKarakoram anomalyrdquo that was first suggested basedon field observations (Hewitt 2005) We report a mass bal-ance of+010plusmn 013 m yrminus1 for Baltoro Glacier (see loca-tion in Fig 8 size= 304 km2) between 1999 and 2010which is consistent with its gradual speed-up reported since2003 (Quincey et al 2009) However given the completelydifferent methods used in our and the latter study we can-not conclude that this demonstrates a real shift from nega-tive to positive mass balance The mass budget of SiachenGlacier (1145 km2 sometimes referred to as the highestbattlefield in the world) is also balanced or slightly pos-itive (+014plusmn 014 m we yrminus1) Thus the alarming state-ment by Rasul (2012) that this glacier retreated by 59 kmand lost 17 of its mass during 1989ndash2009 is very likelyerroneous Again within error bars our regional mass bal-ance in the Karakoram agrees with the mass balance ofminus003plusmn 004 m yrminus1 found by Kaumlaumlb et al (2012) over 2003ndash2008 and with the mass balance ofminus010plusmn 018 m we yrminus1

(2-sigma error level) obtained by Gardner et al (2013) for aregion including both the Karakoram and the Hindu Kush

In the northernmost part of the Pamir in the Alay moun-tain range the mass balance of Abramov Glacier has beenmeasured in the field for 30 yr between 1968 and 1998(WGMS 2012) with a mean value ofminus046 m yrminus1 Herewe report a mass balance ofminus003plusmn 014 m yrminus1 for this

glacier over 1999ndash2011 Our near zero mass budget estimatefor this glacier disagrees with negative mass balances recon-structed for the same period with a model run using biascorrected NCEPNCAR re-analysis data and calibrated withfield data (Barundun et al 2013) An underestimate of ourSRTM C-Band penetration in the Pamir as a whole and inparticular for the small Abramov Glacier may contribute tothese differences

For the whole Pamir study site our mass budget is bal-anced or slightly positive (+014plusmn 013 m we yrminus1) In theeastern Pamir the mass balance of Muztag Ata Glacier was+025 m we yrminus1 between 2005 and 2010 (Yao et al 2012)Although Muztag Ata Glacier is located outside of our Pamirstudy site and his glaciological records only cover the sec-ond half of our study period this measurement is consistentwith our remotely sensed mass balance Thus the ldquoKarako-ram anomalyrdquo should probably be renamed the ldquoPamir-Karakoram anomalyrdquo

This slightly positive or balanced mass budget measuredin the Pamir seems to contradict previous findings by Heidand Kaumlaumlb (2012) They reported rapidly decreasing glacierspeeds in this region between two snapshots of annual ve-locities in 20002001 and 20092010 an observation thatwould fit with negative mass balances (Span and Kuhn 2003Azam et al 2012) We note though that the relationship be-tween glacier mass balance and ice fluxes is not straightfor-ward and might in particular involve a time delay betweena change in mass balance and the related dynamic response(Heid and Kaumlaumlb 2012) Thus the decrease in speed couldwell be due to negative mass balances before or partiallybefore 1999ndash2011 We acknowledge that this discrepancyneeds to be investigated further in particular by examiningcontinuous changes in annual glacier speed through the firstdecade of the 21st Century Indeed Heid and Kaumlaumlb (2012)measured differences between two annual periods which maynot represent mean decadal velocity changes

Jacob et al (2012) using satellite gravimetry (GRACE)measured a mass budget ofminus5plusmn 6 Gt yrminus1 over 2003ndash2010for the ldquoKarakoram-Himalayardquo their region 8c which cor-responds to our study area excluding the Pamir Karakoramand Hindu Kush sub-regions but including some glaciersnorth of the Himalayan ridge already on the Tibetan PlateauFor this subset of our PKH study area we measured amass budget ofminus126plusmn 42 Gt yrminus1 The two estimates over-lap within their error bars especially if our error is dou-bled to match the two standard error level used by Jacob etal (2012) The estimate of Jacob et al (2012) over our com-plete study region (PKH) ieminus6plusmn 8 Gt yrminus1 (sum of theirregions 8b and 8c) also includes mass changes for the KunlunShan sub-region for which we did not measure glacier massbalances Therefore the comparison with our PKH value(minus101plusmn 55 Gt yrminus1) is not straightforward but suggests areasonable agreement especially if we consider the slightmass gain (15plusmn 17 Gt yrminus1) measured with ICESat in theWest Kunlun (Gardner et al 2013)

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1281

Table 6 Annual average of the glacier seasonal and decadal mass loss contributions to Indus Ganges and Brahmaputra discharges Basinarea and seasonal contributions are given by Kaser et al (2010) glacier area in each basin is derived from the RGI v20 (Arendt et al 2012)Numbers in parentheses indicate the percentage of glacierized area within the river basin Seasonal contributions were modeled by Kaser etal (2010) based on the CRU 20 dataset over 1961ndash1990 assuming a balanced glacier mass budget Glacier decadal mass loss contributionsare given as annual average over 1999ndash2011

River Basin area Glacier area Glacier contribution Mean discharge(km2) (km2) (m3 sminus1) (m3 sminus1)

Seasonally delayed Decadal mass lossKaser et al (2010) (this study)

Indus 1 139 814 25 598 (2 ) 105 103plusmn 72 2146 (Chevallier personalcommunication 2012)

Ganges 1 023 609 11 168 (1 ) 47 103plusmn 49 11739 Papa et al (2012Brahmaputra 527 666 15 296 (3 ) 33 147plusmn 64 19710 updated)

54 Reasons for the heterogeneous pattern of massbalance

The PKH glacier mass balance is slightly negative(minus014plusmn 008 m we yrminus1) but changes are not homoge-neous from one sub-region to another and seem to coincidewith the climatic settings of the mountain range For the east-ern sites (from the Hengduan Shan to West Nepal) which areunder the influence of the Indian and south-east Asian sum-mer monsoons the mass balances are moderately negative(sim minus026 m we yrminus1) In the northwest of the PKH at thePamir and Karakoram sites where glacier climate is char-acterized by the westerlies in winter mass balances are bal-anced or slightly positive (sim +010 m we yrminus1) In betweenthese two influences the glaciers of the Spiti Lahaul sitein a monsoon-arid transition zone experienced the strongestmass loss (minus045plusmn 013 m we yrminus1)

This heterogeneous pattern of mass balances seems to berelated to trends in precipitation throughout the PKH Archerand Fowler (2004) reported an increase in winter precipi-tation over the Upper Indus Basin since 1961 a trend con-firmed by Yao et al (2012) over the Pamir and the Karako-ram based on the analysis of the Global Precipitation Cli-matology Project data set (GPCP Adler et al 2003) Thisprecipitation increase is a source of greater accumulation forglaciers and thus a possible explanation for their slightly pos-itive mass budgets In addition in the Upper Indus Basin themean summer temperature has been decreasing since 1961(Fowler and Archer 2006) which is consistent with the find-ings of Shekhar et al (2010) in the Karakoram who reporteda decrease in both maximum and minimum temperature since1988 These increases in winter precipitation and summertemperature likely translate in greater accumulation and re-duced ablation changes which are consistent with the bal-anced or slightly positive glacier mass budget

On the other hand in the east of the PKH the Indiansummer monsoon has been weakening since the 1950s (Bol-lasina et al 2011) which could have reduced the amount

of snowfall over the eastern study sites and thus resultedin negative mass balances In addition maximum and min-imum temperatures increased since 1988 in the western Hi-malaya (Shekhar et al 2010) and maximum temperaturesincreased in the central Himalaya (Nepal) since the mid-1970s (Shrestha et al 1999) Thus both precipitation andtemperature trends in the Himalaya likely contributed to thenegative mass balances measured over the correspondingstudy sites (from Spiti Lahaul to Hengduan Shan Fig 1)

However gridded data sets such as GPCP are not neces-sarily representative of the climate at high altitude IndeedHewitt (2005) reported a 5-to-10-fold increase in precipi-tation between the glacier front and the accumulation areain the Karakoram that is not captured by reanalysis dataas recently confirmed by Immerzeel et al (2012) Thus di-rect measurements of accumulation on High Mountain Asiaglaciers (eg Azam et al 2013) and high-altitude weatherstations are eagerly needed to validate the large scale grid-ded data set (eg reanalysis) test their ability to describe thespecific climate near to PKH glaciers and assess the relativerole played by temperature and precipitation changes

55 Contribution to water resources

Our glacier mass balance can be averaged over large riverbasins to estimate the mean contribution to river runoffdue to a decade of glacier mass loss We assume therebyonly direct river runoff without loss terms For the threerivers for which we cover the entire glacierized area of theircatchments within our sub-regions (Fig 1) these contribu-tions to the river discharge are equal to 103 m3 sminus1 (Indus)103 m3 sminus1 (Ganges) and 147 m3 sminus1 (Brahmaputra) for theperiod 1999ndash2011 (Table 6)

Rivers of PKH being key water resources measures oftheir discharges are highly sensitive and difficult to access forpolitical reasons Papa et al (2012) built recent time series ofdischarges for Ganges and Brahmaputra by using altimetrymeasurements The locations of the corresponding dischargeestimates in Bangladesh are given in Fig 1 We can therefore

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1282 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

evaluate the relative contribution of glacier decadal mass loss(Table 6) to annual river run-off at 09 for the Gangesat Hardinge (basin area ofsim 1 020 000 km2) and 07 forBrahmaputra at Bahadurabad (basin area ofsim 530 000 km2)between 1999 and 2011 Given the discharge measured at theGuddu station by the Surface Water Hydrology Project of theWater and Power Development Authority (Pakistan see lo-cation on Fig 1) between 1999 and 2003 we can estimatethe contribution of glacier mass loss at 48 for the IndusRiver at this station

The seasonal contributions of glaciers to river run-off (iethe part of the annual precipitation that experiences season-ally delayed release from the glaciers) directly results fromseasonal variations of meteorological quantities in contrastto the glacier mass loss contribution which results from thelong-term climatic change Both runoff components directlyaffect the amount of water available in a river at a given lo-cation and at a given time so that their comparison could beof interest Even if the seasonal timing of the glacier massloss contribution is unclear it could in general peak at asimilar time of the year than the seasonal glacier contribu-tion so that both components rather enlarge each other thancancel Seasonal glacier contributions have been modeled byKaser et al (2010) The latter study assumed glaciers to bein equilibrium with the present climate and thus prescribedglacier mass balances equal to 0 Thus their seasonally de-layed glacier contributions do not include the part due todecadal glacier mass loss and is complementary to our es-timates (Table 6) For the eastern basins under the influenceof the Indian summer monsoon (Ganges and Brahmaputra)the contribution of glacier mass loss is higher than the sea-sonal contribution In other words for the first decade of the21st century the mass loss of glaciers is on average moreimportant for water availability in those two rivers than theirseasonal water storage effect The situation is different for theIndus Basin where the glacier melt season is also the dry sea-son which results in a relatively higher seasonally delayedglacier contribution to the river discharge There the season-ally delayed water release from the glaciers and the decadalglacier mass loss contribute on average equally to the riverdischarge Both contributions are strongly (and equally) de-pendent on the location of the discharge measurement andwill be significantly larger (in relative term) in more heav-ily glacierized and smaller mountain catchments located inthe upper part of these major river basins (Bookhagen andBurbank 2010)

6 Conclusion

In this study we assessed the spatial pattern of glaciermass balance over the PKH by comparing the SRTM andSPOT5 DEMs over 9 well-distributed study sites between1999 and 2011 We found slightly positive or balanced massbudgets in the western part (Pamir Karakoram) moder-ate mass loss in the east (Nepal Bhutan Hengduan Shan)

and the most negative mass balance in the monsoon-aridtransition zone (Spiti Lahaul in the western Himalaya)By extrapolating these values to climatically homogeneoussub-regions we estimate the PKH overall mass balanceto have beenminus014plusmn 008 m we yrminus1 between 1999 and2011 which corresponds to a contribution to sea level riseof 0028plusmn 0015 mm yrminus1 This is about 4 of the globalglacier mass loss estimated by Gardner et al (2013) thoughover a slightly shorter time period (2003ndash2009) The largestsource of uncertainties in those geodetic mass balance esti-mates is the poorly constrained penetration of the C-BandSRTM radar signal into snow and ice

Our technique of DEM differencing allows mapping in de-tail the spatial pattern of glacier elevation changes Thosemaps reveal many surge-type behaviors both in the Pamirand the Karakoram It also appears that the glacier-wide massbalance does not seem to be considerably affected by themass transfer from surges over thesim 10 yr time period stud-ied Similar lowering rates over debris-covered and clean iceare observed for four study sites despite the commonly as-sumed insulating effect of debris covers This suggests forthe scale of entire glacier tongues a spatial mixture of re-duced ablation under intact thick debris covers and enhancedablation by supraglacial lakes or ice cliffs or due to thin de-bris layer and very low glacier velocities in ablation areasthat reduce the ice advection downstream

The regionally heterogeneous pattern of ice wastage is inbroad agreement with contrasted trends in large-scale pre-cipitation and temperature However glaciological (eg ac-cumulation) and meteorological records are needed at highaltitude in the vicinity of glaciers to evaluate more closelythe local influence of atmospheric variables on glacier massbalance and fully understand the reasons behind those con-trasted mass budgets in the PKH

Supplementary material related to this article isavailable online at httpwwwthe-cryospherenet712632013tc-7-1263-2013-supplementpdf

Acknowledgements We thank G Cogley A Gardner T NuimuraT Bolch P Wagnon M Barandun M Hoelzle M Huss and ananonymous referee for their constructive comments that led to aconsiderably improved manuscript We thank the Water and PowerDevelopment Authority (WAPDA) for their hydrological data(processed by P Chevallier) and F Papa for sharing his updatedrun-off data ahead of publication We thank T Bolch for sharinghis glacier outlines from the Everest area We thank the USGSfor allowing free access to their Landsat archive CGIAR-SCI forSRTM C-band data DLR for SRTM X-band data and GLIMS forglacier inventories J Gardelle acknowledges a PhD fellowshipfrom the French Space Agency (CNES) and the French NationalResearch Center (CNRS) E Berthier acknowledges support fromCNES through the TOSCA and ISIS proposal no 397 and fromthe Programme National de Teacuteleacutedeacutetection Spatiale E Berthieracknowledges M Masiokas and R Villalba for welcoming him

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1283

at the IANIGLA (Mendoza Argentina) Y Arnaud acknowledgessupport from French National Research Agency through ANR-09-CEP-005-01PAPRIKA A Kaumlaumlb acknowledges support fromESA through Glaciers_cci (400010177810I-AM) and from theEuropean Research Council under the European Unionrsquos SeventhFramework Programme (FP2007-2013)ERC Grant Agreementn 320816

Edited by I M Howat

The publication of this articleis financed by CNRS-INSU

References

Adler R Huffman G Chang A Ferraro R Xie P JanowiakJ Rudolf B Schneider U Curtis S Bolvin D Gruber ASusskind J Arkin P and Nelkin E The Version-2 GlobalPrecipitation Climatology Project (GPCP) Monthly Precipita-tion Analysis (1979ndashPresent) J Hydrometeorol 4 1147ndash11672003

Archer D R and Fowler H J Spatial and temporal variationsin precipitation in the Upper Indus Basin global teleconnectionsand hydrological implications Hydrol Earth Syst Sci 8 47ndash61doi105194hess-8-47-2004 2004

Arendt A Bolch T Cogley J G Gardner A Hagen J-OHock R Kaser G Pfeffer W T Moholdt G Paul F RadicV Andreassen L Bajracharya S Beedle M Berthier EBhambri R Bliss A Brown I Burgess E Burgess DCawkwell F Chinn T Copland L Davies B de AngelisH Dolgova E Filbert K Forester R Fountain A Frey HGiffen B Glasser N Gurney S Hagg W Hall D Hari-tashya U K Hartmann G Helm C Herreid S Howat IKapustin G Khromova T Kienholz C Koenig M KohlerJ Kriegel D Kutuzov S Lavrentiev I LeBris R Lund JManley W Mayer C Miles E Li X Menounos B Mer-cer A Moelg N Mool P Nosenko G Negrete A NuthC Pettersson R Racoviteanu A Ranzi R Rastner P RauF Rich J Rott H Schneider C Seliverstov Y Sharp MSigurethsson O Stokes C Wheate R Winsvold S WolkenG Wyatt F and Zheltyhina N Randolph Glacier Inventory[v20] A Dataset of Global Glacier Outlines Global Land IceMeasurements from Space Boulder Colorado USA Digital Me-dia httpwwwglimsorgRGIrandolphhtml 2012

Azam M Wagnon P Ramanathan A Vincent C Sharma PArnaud Y Linda A Pottakkal J Chevallier P Singh V andBerthier E From balance to imbalance a shift in the dynamicbehaviour of Chhota Shigri glacier western Himalaya India JGlaciol 58 315ndash324 doi1031892012JoG11J123 2012

Azam M F Wagnon P Vincent C Ramanathan A Linda Aand Singh V B Reconstruction of the annual mass balanceof Chhota Shigri Glacier (Western Himalaya India) since 1969Ann Glaciol in revision 2013

Barrand N and Murray T Multivariate Controls on the In-cidence of Glacier Surging in the Karakoram Himalaya

Arct Antarct Alp Res 38 489ndash498 doi1016571523-0430(2006)38[489MCOTIO]20CO2 2006

Barundun M Huss M Azisov E Gafurov A HoelzleM Merkushkin A Salzmann N and Usubaliev R Re-establishing seasonal mass balance observation at AbramovGlacier Kyrgyzstan from 1968ndash2012 Abstract EGU2013-5668European Geophysical Union Vienna 2013

Benn D Bolch T Hands K Gulley J Luckman ANicholson L Quincey D Thompson S Toumi R andWiseman S Response of debris-covered glaciers in theMount Everest region to recent warming and implicationsfor outburst flood hazards Earth-Sci Rev 114 156ndash174doi101016jearscirev201203008 2012

Berthier E and Vincent C Relative contribution of surface mass-balance and ice-flux changes to the accelerated thinning of Merde Glace French Alps over 1979ndash2008 J Glaciol 58 501ndash512doi1031892012JoG11J083 2012

Berthier E Arnaud Y Baratoux D Vincent C and Reacutemy FRecent rapid thinning of the ldquo Mer de Glace rdquo glacier derivedfrom satellite optical images Geophys Res Lett 31 L17401doi1010292004GL020706 2004

Berthier E Arnaud Y Kumar R Ahmad S Wagnon P andChevallier P Remote sensing estimates of glacier mass balancesin the Himachal Pradesh (Western Himalaya India) RemoteSens Environ 108 327ndash338 doi101016jrse2006110172007

Bhambri R Bolch T Kawishwar P Dobhal D P SrivastavaD and Pratap B Heterogeneity in Glacier response from 1973to 2011 in the Shyok valley Karakoram India The CryosphereDiscuss 6 3049ndash3078 doi105194tcd-6-3049-2012 2012

Bhutiyani M Mass-balance studies on Siachen Glacier in theNubra valley Karakoram Himalaya India J Glaciol 45 112ndash118 1999

Bishop M P Schroder J F and Ward J L SPOT multispec-tral analysis for producing supraglacial debris-load estimates forBatura glacier Pakistan Geocarto Int 10 81ndash90 1995

Bolch T Pieczonka T and Benn D I Multi-decadal mass lossof glaciers in the Everest area (Nepal Himalaya) derived fromstereo imagery The Cryosphere 5 349ndash358 doi105194tc-5-349-2011 2011

Bolch T Kulkarni A Kaumlaumlb A Huggel C Paul F Cogley JFrey H Kargel J Fujita K Scheel M Bajracharya S andStoffel M The state and fate of Himalayan Glaciers Science336 310ndash314 doi101126science1215828 2012

Bollasina M Ming Y and Ramaswamy V Anthropogenicaerosols and the weakening of the South Asian summer mon-soon Science 334 502ndash505 doi101126science12049942011

Bookhagen B and Burbank D Toward a complete Himalayan hy-drological budget spatiotemporal distribution of snowmelt andrainfall and their impact on river discharge J Geophys Res115 F03019 doi1010292009JF001426 2010

Bretherton C Widmann M Dymnikov V Wallace J and BladeacuteI The effective number of spatial degrees of freedom of a time-varying field J Climate 12 1990ndash2009 1999

Copland L Pope S Bishop M Shroder J Clendon PBush A Kamp U Seong Y and Owen L Glacier veloc-ities across the central Karakoram Ann Glaciol 50 41ndash49doi103189172756409789624229 2009

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1277

Fig 10Same as Fig 8 but for the Pamir study site

mass balance ofminus021plusmn 005 m we yrminus1 between 2003and 2008 for Hindu Kush-Karakoram-Himalaya glaciersFor that same area (ie excluding the Pamir and theHengduan Shan study sites) we found a mass balance ofminus015plusmn 007 m we yrminus1 Thus our result agrees with Kaumlaumlbet al (2012) within the error bars although our study periodis longer (1999ndash2011) and our measurement principle andextrapolation approach are different In addition we havecalculated updated mass balances from Kaumlaumlb et al (2012)ie averages over 3times 3 degree cells centered over our studysites and find that they are consistent within error barswith our mass balance estimates (Table 5) Similar resultsare found when our mass balances are compared to a spa-tially extended ICESat analysis for 2003ndash2009 (Gardner etal 2013) Mass losses measured with ICESat (Kaumlaumlb et al2012 Gardner et al 2013) tend to be slightly larger thanours This could be due to the difference in time periodsAlso our smaller mass losses could be partly explained if wesystematically underestimated the poorly constrained pene-

tration of the C-band andor X-band radar signal into ice andsnow

At a more local scale Bolch et al (2011) reported a massbalance ofminus079plusmn 052 m yrminus1 between 2002 and 2007 byDEM differencing over a glacier area of 62 km2 includingten glaciers south and west of Mt Everest For these sameglaciers Nuimura et al (2012) found a mean mass balanceof minus045plusmn 060 m yrminus1 during 2000ndash2008 while they com-puted a mass balance ofminus040plusmn 025 m yrminus1 between 1992and 2008 over a glacier area of 183 km2 around Mt EverestIn the present study we found an average mass balance ofminus041plusmn 021 m yrminus1 over the ten glaciers mentioned above(Table 5 Fig S1) and a value ofminus026plusmn 013 m yrminus1 overthe whole Everest study site between 2000 and 2011 Ourvalues are comparable to those of Nuimura et al (2012)Our mass balance is less negative than the 2002ndash2007 massbalance of Bolch et al (2011) but not statistically differentwhen error bars are considered (Fig 12) Indeed Bolch etal (2011) acknowledged some high uncertainties for their

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1278 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Table 4Glacier mass budget of the nine sub-regions to which the results of the study sites (mass balances) have been extrapolated For eachof them the total glacierized area as well as the percentage of the glacier area over which the mass balance was computed (ie study sites)are given The total glacier area is derived from the RGI v20 (Arendt et al 2012) The error for the mass balance in each sub-region is theroot of sum of squares (RSS) of the mass balance error of each study site and the regional extrapolation error The error for the mass budgetin each sub-region includes additionally the error from the total glacier area in the RGI v20

Sub-region Total glacier Measured glacier Mass balance Mass budgetarea (km2) area () (m we yrminus1) (Gt yrminus1)

Hengduan Shan 11584 12 minus033plusmn 014 minus38plusmn 38Bhutan 4021 34 minus022plusmn 013 minus09plusmn 05Everest 6226 23 minus026plusmn 014 minus16plusmn 08West Nepal 6849 13 minus032plusmn 014 minus22plusmn 10Spiti Lahaul 9043 23 minus045plusmn 014 minus41plusmn 13Hindu Kush 6135 13 minus012plusmn 016 minus07plusmn 10Karakoram East

1902428 +011 plusmn 014 +19 plusmn 31

Karakoram West 30 +009 plusmn 018Pamir 9369 34 +014 plusmn 014 +13 plusmn 13

TotalArea- 72251 30 minus014plusmn 008 minus101plusmn 55weighted average

Table 5Comparison of geodetic mass balances between this study and previous published results with overlapping study periods and similargeographic locations Updated figures from Kaumlaumlb et al (2012) are averaged over 3 times 3 cells centered over the study sites of the presentstudy

GlacierSite name This study Bolch et al (2011) Nuimura et al (2012) Kaumlaumlb et al (2012 updated)2002ndash2007 2000ndash2008 2003ndash2008

Study Mass balance Area (km2)a Mass balance Area Mass balance Mass balanceperiod (m we yrminus1) (m we yrminus1) (km2) (m we yrminus1) (m we yrminus1)

Hengduan San 1999ndash2010 minus022 plusmn 014 1303 (55 )Bhutan 1999ndash2010 minus022 plusmn 012 1384 (64 ) minus052 plusmn 016b

Everest

1999ndash2011

minus026 plusmn 013 1461 (58 ) minus039 plusmn 011AX010 minus090plusmn 034 04Changri SharNup minus042plusmn 017 161 (79 ) minus029plusmn 052 130 minus055plusmn 038Khumbu minus051plusmn 019 203 (47 ) minus045plusmn 052 170 minus076plusmn 052Nuptse minus037plusmn 020 50 (74 ) minus040plusmn 053 40 minus034plusmn 027Lhotse Nup minus021plusmn 027 24 (70 ) minus103plusmn 051 19 minus022plusmn 047Lhotse minus043plusmn 018 85 (83 ) minus110plusmn 052 65 minus067plusmn 051Lhotse SharImja minus070plusmn 052 98 (44 ) minus145plusmn 052 107 minus093plusmn 060Amphu Laptse minus046plusmn 034 25 (38 ) minus077plusmn 052 15 minus018plusmn 094Chukhung +044 plusmn 024 42 (47 ) +004 plusmn 054 38 +043 plusmn 081Ama Dablam minus049plusmn 017 36 (54 ) minus056plusmn 052 22 minus056plusmn 073Duwo minus016plusmn 026 19 (18 ) minus196plusmn 053 10 minus068plusmn 074Total Khumbu minus041plusmn 021 744 minus079plusmn 052 617 minus045plusmn 060(10 Glaciers above)

West Nepal 1999ndash2011 minus032 plusmn 013 908 (40 ) minus032 plusmn 012Spiti Lahaul 1999ndash2011 minus045 plusmn 013 2110 (46 ) minus038 plusmn 006Karakoram East 1999ndash2010 +011 plusmn 014 5328 (42 ) minus004 plusmn 004Karakoram West 1999ndash2008 +009 plusmn 018 5434 (45 )Hindu Kush 1999ndash2008 minus012 plusmn 016 793 (80 ) minus020 plusmn 006Pamir 1999ndash2011 +014 plusmn 013 3178 (50 )

a The total glacier area is given in km2 and in parenthesis the of the glacier area actually covered with measurementsb For this cell the ICESat coverage is insufficient and does not sample all glacier elevations

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1279

Fig 11 Elevation changes (SPOT5-SRTM) in the ablation area for clean ice (blue circles) and debris-covered ice (black circles) for eachstudy site For the Karakoram (east and west) and the Pamir study sites only non-surge-type glaciers are considered Standard error of theelevation differences are typically less thanplusmn2 m with each 100 m elevation band (not shown for the sake of clarity) Note elevation changesover separate sections of a glacier cannot be treated as mass changes due to the disregard of glacier dynamics

estimate over this short time period The differences in sur-vey periods and in the glacier outlines may also explainpart of these discrepancies In particular the delimitation ofthe accumulation area is notoriously difficult and Bolch etal (2011) did not include some of the steep high altitudeslopes which we included due to their potential avalanchecontribution For the 10 glaciers mentioned above our massbalances would become 012 m we yrminus1 more negative ifwe used the exact same outlines as Bolch et al (2011)Our positive mass balance for the small Chukhung Glacier(+044plusmn 024 m we yrminus1) is in agreement with Nuimura etal (2012) but enigmatic in a context of regional glacier lossand deserves further investigation If the 2002ndash2007 massbalance estimate by Bolch et al (2011) is excluded all otherrecent mass balance estimates suggest no acceleration of themass loss in the Everest area when compared to the long-term(1970ndash2007) mass balance measured by Bolch et al (2011)in this regionminus032plusmn 008 m weyrminus1

In Bhutan our results indicate a stronger mass loss forsouthern glaciers than for northern ones Interestingly in thisarea Kaumlaumlb (2005) measured high speeds (up to 200 m yrminus1)

for glaciers north of the Himalayan main ridge and ratherlow velocities over southern glacier tongues by using repeat

ASTER data This is consistent with the more negative massbalance that we report for the southern glaciers in Bhutanie for the slow flowing presumably downwasting glaciers

Berthier et al (2007) reported a mass balance betweenminus069 andminus085 m we yrminus1 for an ice-covered area of915 km2 around Chhota Shigri Glacier between 1999 (SRTMDEM) and 2004 (SPOT5-HRG DEM) For the same areawe found a mass balance ofminus045plusmn 013 m we yrminus1 over1999ndash2011 The difference in the study period may explainpart of the differences Also the methodology by Berthieret al (2007) did not explicitly take into account the SRTMradar penetration over glaciers and also corrected an appar-ent elevation-dependent bias according to glacier elevationswhich is now known as a resolution issue and is best cor-rected according to the terrain maximum curvature (Gardelleet al 2012b) More details and discussions about these dif-ferences can be found in Vincent et al (2013)

Importantly the results of the eastern Karakoram studysite confirm the balanced or slightly positive mass budgetreported by Gardelle et al (2012a) in the western Karako-ram over the past decade The two Karakoram study sitesare adjacent with a small overlapping area (sim 1100 km2 ofglaciers) where elevation differences agree within 1 m Both

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1280 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 12 Comparison of geodetic mass balances for 10 glaciers inthe Mount Everest area 1999ndash2010 mass balances measured inthe present study are compared with published estimates during2002ndash2007 (Bolch et al 2011) and during 2000ndash2008 (Nuimuraet al 2012) The 1-to-1 line is drawn The mean difference (plusmn1standard deviation) between (i) our values and Bolch et al (2011)are (047plusmn 058 m we yrminus1) and (ii) our values and Nuimura etal (2012) are (012plusmn 021 m we yrminus1) Part of the differences be-tween estimates may be due to the different periods surveyed andthe different glacier outlines used

sites show similar mass balances over slightly different pe-riods +011plusmn 014 m yrminus1 over 1999ndash2010 in the east and+009plusmn 018 m yrminus1 over 1999ndash2008 in the west This con-firms a ldquoKarakoram anomalyrdquo that was first suggested basedon field observations (Hewitt 2005) We report a mass bal-ance of+010plusmn 013 m yrminus1 for Baltoro Glacier (see loca-tion in Fig 8 size= 304 km2) between 1999 and 2010which is consistent with its gradual speed-up reported since2003 (Quincey et al 2009) However given the completelydifferent methods used in our and the latter study we can-not conclude that this demonstrates a real shift from nega-tive to positive mass balance The mass budget of SiachenGlacier (1145 km2 sometimes referred to as the highestbattlefield in the world) is also balanced or slightly pos-itive (+014plusmn 014 m we yrminus1) Thus the alarming state-ment by Rasul (2012) that this glacier retreated by 59 kmand lost 17 of its mass during 1989ndash2009 is very likelyerroneous Again within error bars our regional mass bal-ance in the Karakoram agrees with the mass balance ofminus003plusmn 004 m yrminus1 found by Kaumlaumlb et al (2012) over 2003ndash2008 and with the mass balance ofminus010plusmn 018 m we yrminus1

(2-sigma error level) obtained by Gardner et al (2013) for aregion including both the Karakoram and the Hindu Kush

In the northernmost part of the Pamir in the Alay moun-tain range the mass balance of Abramov Glacier has beenmeasured in the field for 30 yr between 1968 and 1998(WGMS 2012) with a mean value ofminus046 m yrminus1 Herewe report a mass balance ofminus003plusmn 014 m yrminus1 for this

glacier over 1999ndash2011 Our near zero mass budget estimatefor this glacier disagrees with negative mass balances recon-structed for the same period with a model run using biascorrected NCEPNCAR re-analysis data and calibrated withfield data (Barundun et al 2013) An underestimate of ourSRTM C-Band penetration in the Pamir as a whole and inparticular for the small Abramov Glacier may contribute tothese differences

For the whole Pamir study site our mass budget is bal-anced or slightly positive (+014plusmn 013 m we yrminus1) In theeastern Pamir the mass balance of Muztag Ata Glacier was+025 m we yrminus1 between 2005 and 2010 (Yao et al 2012)Although Muztag Ata Glacier is located outside of our Pamirstudy site and his glaciological records only cover the sec-ond half of our study period this measurement is consistentwith our remotely sensed mass balance Thus the ldquoKarako-ram anomalyrdquo should probably be renamed the ldquoPamir-Karakoram anomalyrdquo

This slightly positive or balanced mass budget measuredin the Pamir seems to contradict previous findings by Heidand Kaumlaumlb (2012) They reported rapidly decreasing glacierspeeds in this region between two snapshots of annual ve-locities in 20002001 and 20092010 an observation thatwould fit with negative mass balances (Span and Kuhn 2003Azam et al 2012) We note though that the relationship be-tween glacier mass balance and ice fluxes is not straightfor-ward and might in particular involve a time delay betweena change in mass balance and the related dynamic response(Heid and Kaumlaumlb 2012) Thus the decrease in speed couldwell be due to negative mass balances before or partiallybefore 1999ndash2011 We acknowledge that this discrepancyneeds to be investigated further in particular by examiningcontinuous changes in annual glacier speed through the firstdecade of the 21st Century Indeed Heid and Kaumlaumlb (2012)measured differences between two annual periods which maynot represent mean decadal velocity changes

Jacob et al (2012) using satellite gravimetry (GRACE)measured a mass budget ofminus5plusmn 6 Gt yrminus1 over 2003ndash2010for the ldquoKarakoram-Himalayardquo their region 8c which cor-responds to our study area excluding the Pamir Karakoramand Hindu Kush sub-regions but including some glaciersnorth of the Himalayan ridge already on the Tibetan PlateauFor this subset of our PKH study area we measured amass budget ofminus126plusmn 42 Gt yrminus1 The two estimates over-lap within their error bars especially if our error is dou-bled to match the two standard error level used by Jacob etal (2012) The estimate of Jacob et al (2012) over our com-plete study region (PKH) ieminus6plusmn 8 Gt yrminus1 (sum of theirregions 8b and 8c) also includes mass changes for the KunlunShan sub-region for which we did not measure glacier massbalances Therefore the comparison with our PKH value(minus101plusmn 55 Gt yrminus1) is not straightforward but suggests areasonable agreement especially if we consider the slightmass gain (15plusmn 17 Gt yrminus1) measured with ICESat in theWest Kunlun (Gardner et al 2013)

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1281

Table 6 Annual average of the glacier seasonal and decadal mass loss contributions to Indus Ganges and Brahmaputra discharges Basinarea and seasonal contributions are given by Kaser et al (2010) glacier area in each basin is derived from the RGI v20 (Arendt et al 2012)Numbers in parentheses indicate the percentage of glacierized area within the river basin Seasonal contributions were modeled by Kaser etal (2010) based on the CRU 20 dataset over 1961ndash1990 assuming a balanced glacier mass budget Glacier decadal mass loss contributionsare given as annual average over 1999ndash2011

River Basin area Glacier area Glacier contribution Mean discharge(km2) (km2) (m3 sminus1) (m3 sminus1)

Seasonally delayed Decadal mass lossKaser et al (2010) (this study)

Indus 1 139 814 25 598 (2 ) 105 103plusmn 72 2146 (Chevallier personalcommunication 2012)

Ganges 1 023 609 11 168 (1 ) 47 103plusmn 49 11739 Papa et al (2012Brahmaputra 527 666 15 296 (3 ) 33 147plusmn 64 19710 updated)

54 Reasons for the heterogeneous pattern of massbalance

The PKH glacier mass balance is slightly negative(minus014plusmn 008 m we yrminus1) but changes are not homoge-neous from one sub-region to another and seem to coincidewith the climatic settings of the mountain range For the east-ern sites (from the Hengduan Shan to West Nepal) which areunder the influence of the Indian and south-east Asian sum-mer monsoons the mass balances are moderately negative(sim minus026 m we yrminus1) In the northwest of the PKH at thePamir and Karakoram sites where glacier climate is char-acterized by the westerlies in winter mass balances are bal-anced or slightly positive (sim +010 m we yrminus1) In betweenthese two influences the glaciers of the Spiti Lahaul sitein a monsoon-arid transition zone experienced the strongestmass loss (minus045plusmn 013 m we yrminus1)

This heterogeneous pattern of mass balances seems to berelated to trends in precipitation throughout the PKH Archerand Fowler (2004) reported an increase in winter precipi-tation over the Upper Indus Basin since 1961 a trend con-firmed by Yao et al (2012) over the Pamir and the Karako-ram based on the analysis of the Global Precipitation Cli-matology Project data set (GPCP Adler et al 2003) Thisprecipitation increase is a source of greater accumulation forglaciers and thus a possible explanation for their slightly pos-itive mass budgets In addition in the Upper Indus Basin themean summer temperature has been decreasing since 1961(Fowler and Archer 2006) which is consistent with the find-ings of Shekhar et al (2010) in the Karakoram who reporteda decrease in both maximum and minimum temperature since1988 These increases in winter precipitation and summertemperature likely translate in greater accumulation and re-duced ablation changes which are consistent with the bal-anced or slightly positive glacier mass budget

On the other hand in the east of the PKH the Indiansummer monsoon has been weakening since the 1950s (Bol-lasina et al 2011) which could have reduced the amount

of snowfall over the eastern study sites and thus resultedin negative mass balances In addition maximum and min-imum temperatures increased since 1988 in the western Hi-malaya (Shekhar et al 2010) and maximum temperaturesincreased in the central Himalaya (Nepal) since the mid-1970s (Shrestha et al 1999) Thus both precipitation andtemperature trends in the Himalaya likely contributed to thenegative mass balances measured over the correspondingstudy sites (from Spiti Lahaul to Hengduan Shan Fig 1)

However gridded data sets such as GPCP are not neces-sarily representative of the climate at high altitude IndeedHewitt (2005) reported a 5-to-10-fold increase in precipi-tation between the glacier front and the accumulation areain the Karakoram that is not captured by reanalysis dataas recently confirmed by Immerzeel et al (2012) Thus di-rect measurements of accumulation on High Mountain Asiaglaciers (eg Azam et al 2013) and high-altitude weatherstations are eagerly needed to validate the large scale grid-ded data set (eg reanalysis) test their ability to describe thespecific climate near to PKH glaciers and assess the relativerole played by temperature and precipitation changes

55 Contribution to water resources

Our glacier mass balance can be averaged over large riverbasins to estimate the mean contribution to river runoffdue to a decade of glacier mass loss We assume therebyonly direct river runoff without loss terms For the threerivers for which we cover the entire glacierized area of theircatchments within our sub-regions (Fig 1) these contribu-tions to the river discharge are equal to 103 m3 sminus1 (Indus)103 m3 sminus1 (Ganges) and 147 m3 sminus1 (Brahmaputra) for theperiod 1999ndash2011 (Table 6)

Rivers of PKH being key water resources measures oftheir discharges are highly sensitive and difficult to access forpolitical reasons Papa et al (2012) built recent time series ofdischarges for Ganges and Brahmaputra by using altimetrymeasurements The locations of the corresponding dischargeestimates in Bangladesh are given in Fig 1 We can therefore

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1282 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

evaluate the relative contribution of glacier decadal mass loss(Table 6) to annual river run-off at 09 for the Gangesat Hardinge (basin area ofsim 1 020 000 km2) and 07 forBrahmaputra at Bahadurabad (basin area ofsim 530 000 km2)between 1999 and 2011 Given the discharge measured at theGuddu station by the Surface Water Hydrology Project of theWater and Power Development Authority (Pakistan see lo-cation on Fig 1) between 1999 and 2003 we can estimatethe contribution of glacier mass loss at 48 for the IndusRiver at this station

The seasonal contributions of glaciers to river run-off (iethe part of the annual precipitation that experiences season-ally delayed release from the glaciers) directly results fromseasonal variations of meteorological quantities in contrastto the glacier mass loss contribution which results from thelong-term climatic change Both runoff components directlyaffect the amount of water available in a river at a given lo-cation and at a given time so that their comparison could beof interest Even if the seasonal timing of the glacier massloss contribution is unclear it could in general peak at asimilar time of the year than the seasonal glacier contribu-tion so that both components rather enlarge each other thancancel Seasonal glacier contributions have been modeled byKaser et al (2010) The latter study assumed glaciers to bein equilibrium with the present climate and thus prescribedglacier mass balances equal to 0 Thus their seasonally de-layed glacier contributions do not include the part due todecadal glacier mass loss and is complementary to our es-timates (Table 6) For the eastern basins under the influenceof the Indian summer monsoon (Ganges and Brahmaputra)the contribution of glacier mass loss is higher than the sea-sonal contribution In other words for the first decade of the21st century the mass loss of glaciers is on average moreimportant for water availability in those two rivers than theirseasonal water storage effect The situation is different for theIndus Basin where the glacier melt season is also the dry sea-son which results in a relatively higher seasonally delayedglacier contribution to the river discharge There the season-ally delayed water release from the glaciers and the decadalglacier mass loss contribute on average equally to the riverdischarge Both contributions are strongly (and equally) de-pendent on the location of the discharge measurement andwill be significantly larger (in relative term) in more heav-ily glacierized and smaller mountain catchments located inthe upper part of these major river basins (Bookhagen andBurbank 2010)

6 Conclusion

In this study we assessed the spatial pattern of glaciermass balance over the PKH by comparing the SRTM andSPOT5 DEMs over 9 well-distributed study sites between1999 and 2011 We found slightly positive or balanced massbudgets in the western part (Pamir Karakoram) moder-ate mass loss in the east (Nepal Bhutan Hengduan Shan)

and the most negative mass balance in the monsoon-aridtransition zone (Spiti Lahaul in the western Himalaya)By extrapolating these values to climatically homogeneoussub-regions we estimate the PKH overall mass balanceto have beenminus014plusmn 008 m we yrminus1 between 1999 and2011 which corresponds to a contribution to sea level riseof 0028plusmn 0015 mm yrminus1 This is about 4 of the globalglacier mass loss estimated by Gardner et al (2013) thoughover a slightly shorter time period (2003ndash2009) The largestsource of uncertainties in those geodetic mass balance esti-mates is the poorly constrained penetration of the C-BandSRTM radar signal into snow and ice

Our technique of DEM differencing allows mapping in de-tail the spatial pattern of glacier elevation changes Thosemaps reveal many surge-type behaviors both in the Pamirand the Karakoram It also appears that the glacier-wide massbalance does not seem to be considerably affected by themass transfer from surges over thesim 10 yr time period stud-ied Similar lowering rates over debris-covered and clean iceare observed for four study sites despite the commonly as-sumed insulating effect of debris covers This suggests forthe scale of entire glacier tongues a spatial mixture of re-duced ablation under intact thick debris covers and enhancedablation by supraglacial lakes or ice cliffs or due to thin de-bris layer and very low glacier velocities in ablation areasthat reduce the ice advection downstream

The regionally heterogeneous pattern of ice wastage is inbroad agreement with contrasted trends in large-scale pre-cipitation and temperature However glaciological (eg ac-cumulation) and meteorological records are needed at highaltitude in the vicinity of glaciers to evaluate more closelythe local influence of atmospheric variables on glacier massbalance and fully understand the reasons behind those con-trasted mass budgets in the PKH

Supplementary material related to this article isavailable online at httpwwwthe-cryospherenet712632013tc-7-1263-2013-supplementpdf

Acknowledgements We thank G Cogley A Gardner T NuimuraT Bolch P Wagnon M Barandun M Hoelzle M Huss and ananonymous referee for their constructive comments that led to aconsiderably improved manuscript We thank the Water and PowerDevelopment Authority (WAPDA) for their hydrological data(processed by P Chevallier) and F Papa for sharing his updatedrun-off data ahead of publication We thank T Bolch for sharinghis glacier outlines from the Everest area We thank the USGSfor allowing free access to their Landsat archive CGIAR-SCI forSRTM C-band data DLR for SRTM X-band data and GLIMS forglacier inventories J Gardelle acknowledges a PhD fellowshipfrom the French Space Agency (CNES) and the French NationalResearch Center (CNRS) E Berthier acknowledges support fromCNES through the TOSCA and ISIS proposal no 397 and fromthe Programme National de Teacuteleacutedeacutetection Spatiale E Berthieracknowledges M Masiokas and R Villalba for welcoming him

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1283

at the IANIGLA (Mendoza Argentina) Y Arnaud acknowledgessupport from French National Research Agency through ANR-09-CEP-005-01PAPRIKA A Kaumlaumlb acknowledges support fromESA through Glaciers_cci (400010177810I-AM) and from theEuropean Research Council under the European Unionrsquos SeventhFramework Programme (FP2007-2013)ERC Grant Agreementn 320816

Edited by I M Howat

The publication of this articleis financed by CNRS-INSU

References

Adler R Huffman G Chang A Ferraro R Xie P JanowiakJ Rudolf B Schneider U Curtis S Bolvin D Gruber ASusskind J Arkin P and Nelkin E The Version-2 GlobalPrecipitation Climatology Project (GPCP) Monthly Precipita-tion Analysis (1979ndashPresent) J Hydrometeorol 4 1147ndash11672003

Archer D R and Fowler H J Spatial and temporal variationsin precipitation in the Upper Indus Basin global teleconnectionsand hydrological implications Hydrol Earth Syst Sci 8 47ndash61doi105194hess-8-47-2004 2004

Arendt A Bolch T Cogley J G Gardner A Hagen J-OHock R Kaser G Pfeffer W T Moholdt G Paul F RadicV Andreassen L Bajracharya S Beedle M Berthier EBhambri R Bliss A Brown I Burgess E Burgess DCawkwell F Chinn T Copland L Davies B de AngelisH Dolgova E Filbert K Forester R Fountain A Frey HGiffen B Glasser N Gurney S Hagg W Hall D Hari-tashya U K Hartmann G Helm C Herreid S Howat IKapustin G Khromova T Kienholz C Koenig M KohlerJ Kriegel D Kutuzov S Lavrentiev I LeBris R Lund JManley W Mayer C Miles E Li X Menounos B Mer-cer A Moelg N Mool P Nosenko G Negrete A NuthC Pettersson R Racoviteanu A Ranzi R Rastner P RauF Rich J Rott H Schneider C Seliverstov Y Sharp MSigurethsson O Stokes C Wheate R Winsvold S WolkenG Wyatt F and Zheltyhina N Randolph Glacier Inventory[v20] A Dataset of Global Glacier Outlines Global Land IceMeasurements from Space Boulder Colorado USA Digital Me-dia httpwwwglimsorgRGIrandolphhtml 2012

Azam M Wagnon P Ramanathan A Vincent C Sharma PArnaud Y Linda A Pottakkal J Chevallier P Singh V andBerthier E From balance to imbalance a shift in the dynamicbehaviour of Chhota Shigri glacier western Himalaya India JGlaciol 58 315ndash324 doi1031892012JoG11J123 2012

Azam M F Wagnon P Vincent C Ramanathan A Linda Aand Singh V B Reconstruction of the annual mass balanceof Chhota Shigri Glacier (Western Himalaya India) since 1969Ann Glaciol in revision 2013

Barrand N and Murray T Multivariate Controls on the In-cidence of Glacier Surging in the Karakoram Himalaya

Arct Antarct Alp Res 38 489ndash498 doi1016571523-0430(2006)38[489MCOTIO]20CO2 2006

Barundun M Huss M Azisov E Gafurov A HoelzleM Merkushkin A Salzmann N and Usubaliev R Re-establishing seasonal mass balance observation at AbramovGlacier Kyrgyzstan from 1968ndash2012 Abstract EGU2013-5668European Geophysical Union Vienna 2013

Benn D Bolch T Hands K Gulley J Luckman ANicholson L Quincey D Thompson S Toumi R andWiseman S Response of debris-covered glaciers in theMount Everest region to recent warming and implicationsfor outburst flood hazards Earth-Sci Rev 114 156ndash174doi101016jearscirev201203008 2012

Berthier E and Vincent C Relative contribution of surface mass-balance and ice-flux changes to the accelerated thinning of Merde Glace French Alps over 1979ndash2008 J Glaciol 58 501ndash512doi1031892012JoG11J083 2012

Berthier E Arnaud Y Baratoux D Vincent C and Reacutemy FRecent rapid thinning of the ldquo Mer de Glace rdquo glacier derivedfrom satellite optical images Geophys Res Lett 31 L17401doi1010292004GL020706 2004

Berthier E Arnaud Y Kumar R Ahmad S Wagnon P andChevallier P Remote sensing estimates of glacier mass balancesin the Himachal Pradesh (Western Himalaya India) RemoteSens Environ 108 327ndash338 doi101016jrse2006110172007

Bhambri R Bolch T Kawishwar P Dobhal D P SrivastavaD and Pratap B Heterogeneity in Glacier response from 1973to 2011 in the Shyok valley Karakoram India The CryosphereDiscuss 6 3049ndash3078 doi105194tcd-6-3049-2012 2012

Bhutiyani M Mass-balance studies on Siachen Glacier in theNubra valley Karakoram Himalaya India J Glaciol 45 112ndash118 1999

Bishop M P Schroder J F and Ward J L SPOT multispec-tral analysis for producing supraglacial debris-load estimates forBatura glacier Pakistan Geocarto Int 10 81ndash90 1995

Bolch T Pieczonka T and Benn D I Multi-decadal mass lossof glaciers in the Everest area (Nepal Himalaya) derived fromstereo imagery The Cryosphere 5 349ndash358 doi105194tc-5-349-2011 2011

Bolch T Kulkarni A Kaumlaumlb A Huggel C Paul F Cogley JFrey H Kargel J Fujita K Scheel M Bajracharya S andStoffel M The state and fate of Himalayan Glaciers Science336 310ndash314 doi101126science1215828 2012

Bollasina M Ming Y and Ramaswamy V Anthropogenicaerosols and the weakening of the South Asian summer mon-soon Science 334 502ndash505 doi101126science12049942011

Bookhagen B and Burbank D Toward a complete Himalayan hy-drological budget spatiotemporal distribution of snowmelt andrainfall and their impact on river discharge J Geophys Res115 F03019 doi1010292009JF001426 2010

Bretherton C Widmann M Dymnikov V Wallace J and BladeacuteI The effective number of spatial degrees of freedom of a time-varying field J Climate 12 1990ndash2009 1999

Copland L Pope S Bishop M Shroder J Clendon PBush A Kamp U Seong Y and Owen L Glacier veloc-ities across the central Karakoram Ann Glaciol 50 41ndash49doi103189172756409789624229 2009

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

1278 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Table 4Glacier mass budget of the nine sub-regions to which the results of the study sites (mass balances) have been extrapolated For eachof them the total glacierized area as well as the percentage of the glacier area over which the mass balance was computed (ie study sites)are given The total glacier area is derived from the RGI v20 (Arendt et al 2012) The error for the mass balance in each sub-region is theroot of sum of squares (RSS) of the mass balance error of each study site and the regional extrapolation error The error for the mass budgetin each sub-region includes additionally the error from the total glacier area in the RGI v20

Sub-region Total glacier Measured glacier Mass balance Mass budgetarea (km2) area () (m we yrminus1) (Gt yrminus1)

Hengduan Shan 11584 12 minus033plusmn 014 minus38plusmn 38Bhutan 4021 34 minus022plusmn 013 minus09plusmn 05Everest 6226 23 minus026plusmn 014 minus16plusmn 08West Nepal 6849 13 minus032plusmn 014 minus22plusmn 10Spiti Lahaul 9043 23 minus045plusmn 014 minus41plusmn 13Hindu Kush 6135 13 minus012plusmn 016 minus07plusmn 10Karakoram East

1902428 +011 plusmn 014 +19 plusmn 31

Karakoram West 30 +009 plusmn 018Pamir 9369 34 +014 plusmn 014 +13 plusmn 13

TotalArea- 72251 30 minus014plusmn 008 minus101plusmn 55weighted average

Table 5Comparison of geodetic mass balances between this study and previous published results with overlapping study periods and similargeographic locations Updated figures from Kaumlaumlb et al (2012) are averaged over 3 times 3 cells centered over the study sites of the presentstudy

GlacierSite name This study Bolch et al (2011) Nuimura et al (2012) Kaumlaumlb et al (2012 updated)2002ndash2007 2000ndash2008 2003ndash2008

Study Mass balance Area (km2)a Mass balance Area Mass balance Mass balanceperiod (m we yrminus1) (m we yrminus1) (km2) (m we yrminus1) (m we yrminus1)

Hengduan San 1999ndash2010 minus022 plusmn 014 1303 (55 )Bhutan 1999ndash2010 minus022 plusmn 012 1384 (64 ) minus052 plusmn 016b

Everest

1999ndash2011

minus026 plusmn 013 1461 (58 ) minus039 plusmn 011AX010 minus090plusmn 034 04Changri SharNup minus042plusmn 017 161 (79 ) minus029plusmn 052 130 minus055plusmn 038Khumbu minus051plusmn 019 203 (47 ) minus045plusmn 052 170 minus076plusmn 052Nuptse minus037plusmn 020 50 (74 ) minus040plusmn 053 40 minus034plusmn 027Lhotse Nup minus021plusmn 027 24 (70 ) minus103plusmn 051 19 minus022plusmn 047Lhotse minus043plusmn 018 85 (83 ) minus110plusmn 052 65 minus067plusmn 051Lhotse SharImja minus070plusmn 052 98 (44 ) minus145plusmn 052 107 minus093plusmn 060Amphu Laptse minus046plusmn 034 25 (38 ) minus077plusmn 052 15 minus018plusmn 094Chukhung +044 plusmn 024 42 (47 ) +004 plusmn 054 38 +043 plusmn 081Ama Dablam minus049plusmn 017 36 (54 ) minus056plusmn 052 22 minus056plusmn 073Duwo minus016plusmn 026 19 (18 ) minus196plusmn 053 10 minus068plusmn 074Total Khumbu minus041plusmn 021 744 minus079plusmn 052 617 minus045plusmn 060(10 Glaciers above)

West Nepal 1999ndash2011 minus032 plusmn 013 908 (40 ) minus032 plusmn 012Spiti Lahaul 1999ndash2011 minus045 plusmn 013 2110 (46 ) minus038 plusmn 006Karakoram East 1999ndash2010 +011 plusmn 014 5328 (42 ) minus004 plusmn 004Karakoram West 1999ndash2008 +009 plusmn 018 5434 (45 )Hindu Kush 1999ndash2008 minus012 plusmn 016 793 (80 ) minus020 plusmn 006Pamir 1999ndash2011 +014 plusmn 013 3178 (50 )

a The total glacier area is given in km2 and in parenthesis the of the glacier area actually covered with measurementsb For this cell the ICESat coverage is insufficient and does not sample all glacier elevations

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1279

Fig 11 Elevation changes (SPOT5-SRTM) in the ablation area for clean ice (blue circles) and debris-covered ice (black circles) for eachstudy site For the Karakoram (east and west) and the Pamir study sites only non-surge-type glaciers are considered Standard error of theelevation differences are typically less thanplusmn2 m with each 100 m elevation band (not shown for the sake of clarity) Note elevation changesover separate sections of a glacier cannot be treated as mass changes due to the disregard of glacier dynamics

estimate over this short time period The differences in sur-vey periods and in the glacier outlines may also explainpart of these discrepancies In particular the delimitation ofthe accumulation area is notoriously difficult and Bolch etal (2011) did not include some of the steep high altitudeslopes which we included due to their potential avalanchecontribution For the 10 glaciers mentioned above our massbalances would become 012 m we yrminus1 more negative ifwe used the exact same outlines as Bolch et al (2011)Our positive mass balance for the small Chukhung Glacier(+044plusmn 024 m we yrminus1) is in agreement with Nuimura etal (2012) but enigmatic in a context of regional glacier lossand deserves further investigation If the 2002ndash2007 massbalance estimate by Bolch et al (2011) is excluded all otherrecent mass balance estimates suggest no acceleration of themass loss in the Everest area when compared to the long-term(1970ndash2007) mass balance measured by Bolch et al (2011)in this regionminus032plusmn 008 m weyrminus1

In Bhutan our results indicate a stronger mass loss forsouthern glaciers than for northern ones Interestingly in thisarea Kaumlaumlb (2005) measured high speeds (up to 200 m yrminus1)

for glaciers north of the Himalayan main ridge and ratherlow velocities over southern glacier tongues by using repeat

ASTER data This is consistent with the more negative massbalance that we report for the southern glaciers in Bhutanie for the slow flowing presumably downwasting glaciers

Berthier et al (2007) reported a mass balance betweenminus069 andminus085 m we yrminus1 for an ice-covered area of915 km2 around Chhota Shigri Glacier between 1999 (SRTMDEM) and 2004 (SPOT5-HRG DEM) For the same areawe found a mass balance ofminus045plusmn 013 m we yrminus1 over1999ndash2011 The difference in the study period may explainpart of the differences Also the methodology by Berthieret al (2007) did not explicitly take into account the SRTMradar penetration over glaciers and also corrected an appar-ent elevation-dependent bias according to glacier elevationswhich is now known as a resolution issue and is best cor-rected according to the terrain maximum curvature (Gardelleet al 2012b) More details and discussions about these dif-ferences can be found in Vincent et al (2013)

Importantly the results of the eastern Karakoram studysite confirm the balanced or slightly positive mass budgetreported by Gardelle et al (2012a) in the western Karako-ram over the past decade The two Karakoram study sitesare adjacent with a small overlapping area (sim 1100 km2 ofglaciers) where elevation differences agree within 1 m Both

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1280 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 12 Comparison of geodetic mass balances for 10 glaciers inthe Mount Everest area 1999ndash2010 mass balances measured inthe present study are compared with published estimates during2002ndash2007 (Bolch et al 2011) and during 2000ndash2008 (Nuimuraet al 2012) The 1-to-1 line is drawn The mean difference (plusmn1standard deviation) between (i) our values and Bolch et al (2011)are (047plusmn 058 m we yrminus1) and (ii) our values and Nuimura etal (2012) are (012plusmn 021 m we yrminus1) Part of the differences be-tween estimates may be due to the different periods surveyed andthe different glacier outlines used

sites show similar mass balances over slightly different pe-riods +011plusmn 014 m yrminus1 over 1999ndash2010 in the east and+009plusmn 018 m yrminus1 over 1999ndash2008 in the west This con-firms a ldquoKarakoram anomalyrdquo that was first suggested basedon field observations (Hewitt 2005) We report a mass bal-ance of+010plusmn 013 m yrminus1 for Baltoro Glacier (see loca-tion in Fig 8 size= 304 km2) between 1999 and 2010which is consistent with its gradual speed-up reported since2003 (Quincey et al 2009) However given the completelydifferent methods used in our and the latter study we can-not conclude that this demonstrates a real shift from nega-tive to positive mass balance The mass budget of SiachenGlacier (1145 km2 sometimes referred to as the highestbattlefield in the world) is also balanced or slightly pos-itive (+014plusmn 014 m we yrminus1) Thus the alarming state-ment by Rasul (2012) that this glacier retreated by 59 kmand lost 17 of its mass during 1989ndash2009 is very likelyerroneous Again within error bars our regional mass bal-ance in the Karakoram agrees with the mass balance ofminus003plusmn 004 m yrminus1 found by Kaumlaumlb et al (2012) over 2003ndash2008 and with the mass balance ofminus010plusmn 018 m we yrminus1

(2-sigma error level) obtained by Gardner et al (2013) for aregion including both the Karakoram and the Hindu Kush

In the northernmost part of the Pamir in the Alay moun-tain range the mass balance of Abramov Glacier has beenmeasured in the field for 30 yr between 1968 and 1998(WGMS 2012) with a mean value ofminus046 m yrminus1 Herewe report a mass balance ofminus003plusmn 014 m yrminus1 for this

glacier over 1999ndash2011 Our near zero mass budget estimatefor this glacier disagrees with negative mass balances recon-structed for the same period with a model run using biascorrected NCEPNCAR re-analysis data and calibrated withfield data (Barundun et al 2013) An underestimate of ourSRTM C-Band penetration in the Pamir as a whole and inparticular for the small Abramov Glacier may contribute tothese differences

For the whole Pamir study site our mass budget is bal-anced or slightly positive (+014plusmn 013 m we yrminus1) In theeastern Pamir the mass balance of Muztag Ata Glacier was+025 m we yrminus1 between 2005 and 2010 (Yao et al 2012)Although Muztag Ata Glacier is located outside of our Pamirstudy site and his glaciological records only cover the sec-ond half of our study period this measurement is consistentwith our remotely sensed mass balance Thus the ldquoKarako-ram anomalyrdquo should probably be renamed the ldquoPamir-Karakoram anomalyrdquo

This slightly positive or balanced mass budget measuredin the Pamir seems to contradict previous findings by Heidand Kaumlaumlb (2012) They reported rapidly decreasing glacierspeeds in this region between two snapshots of annual ve-locities in 20002001 and 20092010 an observation thatwould fit with negative mass balances (Span and Kuhn 2003Azam et al 2012) We note though that the relationship be-tween glacier mass balance and ice fluxes is not straightfor-ward and might in particular involve a time delay betweena change in mass balance and the related dynamic response(Heid and Kaumlaumlb 2012) Thus the decrease in speed couldwell be due to negative mass balances before or partiallybefore 1999ndash2011 We acknowledge that this discrepancyneeds to be investigated further in particular by examiningcontinuous changes in annual glacier speed through the firstdecade of the 21st Century Indeed Heid and Kaumlaumlb (2012)measured differences between two annual periods which maynot represent mean decadal velocity changes

Jacob et al (2012) using satellite gravimetry (GRACE)measured a mass budget ofminus5plusmn 6 Gt yrminus1 over 2003ndash2010for the ldquoKarakoram-Himalayardquo their region 8c which cor-responds to our study area excluding the Pamir Karakoramand Hindu Kush sub-regions but including some glaciersnorth of the Himalayan ridge already on the Tibetan PlateauFor this subset of our PKH study area we measured amass budget ofminus126plusmn 42 Gt yrminus1 The two estimates over-lap within their error bars especially if our error is dou-bled to match the two standard error level used by Jacob etal (2012) The estimate of Jacob et al (2012) over our com-plete study region (PKH) ieminus6plusmn 8 Gt yrminus1 (sum of theirregions 8b and 8c) also includes mass changes for the KunlunShan sub-region for which we did not measure glacier massbalances Therefore the comparison with our PKH value(minus101plusmn 55 Gt yrminus1) is not straightforward but suggests areasonable agreement especially if we consider the slightmass gain (15plusmn 17 Gt yrminus1) measured with ICESat in theWest Kunlun (Gardner et al 2013)

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1281

Table 6 Annual average of the glacier seasonal and decadal mass loss contributions to Indus Ganges and Brahmaputra discharges Basinarea and seasonal contributions are given by Kaser et al (2010) glacier area in each basin is derived from the RGI v20 (Arendt et al 2012)Numbers in parentheses indicate the percentage of glacierized area within the river basin Seasonal contributions were modeled by Kaser etal (2010) based on the CRU 20 dataset over 1961ndash1990 assuming a balanced glacier mass budget Glacier decadal mass loss contributionsare given as annual average over 1999ndash2011

River Basin area Glacier area Glacier contribution Mean discharge(km2) (km2) (m3 sminus1) (m3 sminus1)

Seasonally delayed Decadal mass lossKaser et al (2010) (this study)

Indus 1 139 814 25 598 (2 ) 105 103plusmn 72 2146 (Chevallier personalcommunication 2012)

Ganges 1 023 609 11 168 (1 ) 47 103plusmn 49 11739 Papa et al (2012Brahmaputra 527 666 15 296 (3 ) 33 147plusmn 64 19710 updated)

54 Reasons for the heterogeneous pattern of massbalance

The PKH glacier mass balance is slightly negative(minus014plusmn 008 m we yrminus1) but changes are not homoge-neous from one sub-region to another and seem to coincidewith the climatic settings of the mountain range For the east-ern sites (from the Hengduan Shan to West Nepal) which areunder the influence of the Indian and south-east Asian sum-mer monsoons the mass balances are moderately negative(sim minus026 m we yrminus1) In the northwest of the PKH at thePamir and Karakoram sites where glacier climate is char-acterized by the westerlies in winter mass balances are bal-anced or slightly positive (sim +010 m we yrminus1) In betweenthese two influences the glaciers of the Spiti Lahaul sitein a monsoon-arid transition zone experienced the strongestmass loss (minus045plusmn 013 m we yrminus1)

This heterogeneous pattern of mass balances seems to berelated to trends in precipitation throughout the PKH Archerand Fowler (2004) reported an increase in winter precipi-tation over the Upper Indus Basin since 1961 a trend con-firmed by Yao et al (2012) over the Pamir and the Karako-ram based on the analysis of the Global Precipitation Cli-matology Project data set (GPCP Adler et al 2003) Thisprecipitation increase is a source of greater accumulation forglaciers and thus a possible explanation for their slightly pos-itive mass budgets In addition in the Upper Indus Basin themean summer temperature has been decreasing since 1961(Fowler and Archer 2006) which is consistent with the find-ings of Shekhar et al (2010) in the Karakoram who reporteda decrease in both maximum and minimum temperature since1988 These increases in winter precipitation and summertemperature likely translate in greater accumulation and re-duced ablation changes which are consistent with the bal-anced or slightly positive glacier mass budget

On the other hand in the east of the PKH the Indiansummer monsoon has been weakening since the 1950s (Bol-lasina et al 2011) which could have reduced the amount

of snowfall over the eastern study sites and thus resultedin negative mass balances In addition maximum and min-imum temperatures increased since 1988 in the western Hi-malaya (Shekhar et al 2010) and maximum temperaturesincreased in the central Himalaya (Nepal) since the mid-1970s (Shrestha et al 1999) Thus both precipitation andtemperature trends in the Himalaya likely contributed to thenegative mass balances measured over the correspondingstudy sites (from Spiti Lahaul to Hengduan Shan Fig 1)

However gridded data sets such as GPCP are not neces-sarily representative of the climate at high altitude IndeedHewitt (2005) reported a 5-to-10-fold increase in precipi-tation between the glacier front and the accumulation areain the Karakoram that is not captured by reanalysis dataas recently confirmed by Immerzeel et al (2012) Thus di-rect measurements of accumulation on High Mountain Asiaglaciers (eg Azam et al 2013) and high-altitude weatherstations are eagerly needed to validate the large scale grid-ded data set (eg reanalysis) test their ability to describe thespecific climate near to PKH glaciers and assess the relativerole played by temperature and precipitation changes

55 Contribution to water resources

Our glacier mass balance can be averaged over large riverbasins to estimate the mean contribution to river runoffdue to a decade of glacier mass loss We assume therebyonly direct river runoff without loss terms For the threerivers for which we cover the entire glacierized area of theircatchments within our sub-regions (Fig 1) these contribu-tions to the river discharge are equal to 103 m3 sminus1 (Indus)103 m3 sminus1 (Ganges) and 147 m3 sminus1 (Brahmaputra) for theperiod 1999ndash2011 (Table 6)

Rivers of PKH being key water resources measures oftheir discharges are highly sensitive and difficult to access forpolitical reasons Papa et al (2012) built recent time series ofdischarges for Ganges and Brahmaputra by using altimetrymeasurements The locations of the corresponding dischargeestimates in Bangladesh are given in Fig 1 We can therefore

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1282 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

evaluate the relative contribution of glacier decadal mass loss(Table 6) to annual river run-off at 09 for the Gangesat Hardinge (basin area ofsim 1 020 000 km2) and 07 forBrahmaputra at Bahadurabad (basin area ofsim 530 000 km2)between 1999 and 2011 Given the discharge measured at theGuddu station by the Surface Water Hydrology Project of theWater and Power Development Authority (Pakistan see lo-cation on Fig 1) between 1999 and 2003 we can estimatethe contribution of glacier mass loss at 48 for the IndusRiver at this station

The seasonal contributions of glaciers to river run-off (iethe part of the annual precipitation that experiences season-ally delayed release from the glaciers) directly results fromseasonal variations of meteorological quantities in contrastto the glacier mass loss contribution which results from thelong-term climatic change Both runoff components directlyaffect the amount of water available in a river at a given lo-cation and at a given time so that their comparison could beof interest Even if the seasonal timing of the glacier massloss contribution is unclear it could in general peak at asimilar time of the year than the seasonal glacier contribu-tion so that both components rather enlarge each other thancancel Seasonal glacier contributions have been modeled byKaser et al (2010) The latter study assumed glaciers to bein equilibrium with the present climate and thus prescribedglacier mass balances equal to 0 Thus their seasonally de-layed glacier contributions do not include the part due todecadal glacier mass loss and is complementary to our es-timates (Table 6) For the eastern basins under the influenceof the Indian summer monsoon (Ganges and Brahmaputra)the contribution of glacier mass loss is higher than the sea-sonal contribution In other words for the first decade of the21st century the mass loss of glaciers is on average moreimportant for water availability in those two rivers than theirseasonal water storage effect The situation is different for theIndus Basin where the glacier melt season is also the dry sea-son which results in a relatively higher seasonally delayedglacier contribution to the river discharge There the season-ally delayed water release from the glaciers and the decadalglacier mass loss contribute on average equally to the riverdischarge Both contributions are strongly (and equally) de-pendent on the location of the discharge measurement andwill be significantly larger (in relative term) in more heav-ily glacierized and smaller mountain catchments located inthe upper part of these major river basins (Bookhagen andBurbank 2010)

6 Conclusion

In this study we assessed the spatial pattern of glaciermass balance over the PKH by comparing the SRTM andSPOT5 DEMs over 9 well-distributed study sites between1999 and 2011 We found slightly positive or balanced massbudgets in the western part (Pamir Karakoram) moder-ate mass loss in the east (Nepal Bhutan Hengduan Shan)

and the most negative mass balance in the monsoon-aridtransition zone (Spiti Lahaul in the western Himalaya)By extrapolating these values to climatically homogeneoussub-regions we estimate the PKH overall mass balanceto have beenminus014plusmn 008 m we yrminus1 between 1999 and2011 which corresponds to a contribution to sea level riseof 0028plusmn 0015 mm yrminus1 This is about 4 of the globalglacier mass loss estimated by Gardner et al (2013) thoughover a slightly shorter time period (2003ndash2009) The largestsource of uncertainties in those geodetic mass balance esti-mates is the poorly constrained penetration of the C-BandSRTM radar signal into snow and ice

Our technique of DEM differencing allows mapping in de-tail the spatial pattern of glacier elevation changes Thosemaps reveal many surge-type behaviors both in the Pamirand the Karakoram It also appears that the glacier-wide massbalance does not seem to be considerably affected by themass transfer from surges over thesim 10 yr time period stud-ied Similar lowering rates over debris-covered and clean iceare observed for four study sites despite the commonly as-sumed insulating effect of debris covers This suggests forthe scale of entire glacier tongues a spatial mixture of re-duced ablation under intact thick debris covers and enhancedablation by supraglacial lakes or ice cliffs or due to thin de-bris layer and very low glacier velocities in ablation areasthat reduce the ice advection downstream

The regionally heterogeneous pattern of ice wastage is inbroad agreement with contrasted trends in large-scale pre-cipitation and temperature However glaciological (eg ac-cumulation) and meteorological records are needed at highaltitude in the vicinity of glaciers to evaluate more closelythe local influence of atmospheric variables on glacier massbalance and fully understand the reasons behind those con-trasted mass budgets in the PKH

Supplementary material related to this article isavailable online at httpwwwthe-cryospherenet712632013tc-7-1263-2013-supplementpdf

Acknowledgements We thank G Cogley A Gardner T NuimuraT Bolch P Wagnon M Barandun M Hoelzle M Huss and ananonymous referee for their constructive comments that led to aconsiderably improved manuscript We thank the Water and PowerDevelopment Authority (WAPDA) for their hydrological data(processed by P Chevallier) and F Papa for sharing his updatedrun-off data ahead of publication We thank T Bolch for sharinghis glacier outlines from the Everest area We thank the USGSfor allowing free access to their Landsat archive CGIAR-SCI forSRTM C-band data DLR for SRTM X-band data and GLIMS forglacier inventories J Gardelle acknowledges a PhD fellowshipfrom the French Space Agency (CNES) and the French NationalResearch Center (CNRS) E Berthier acknowledges support fromCNES through the TOSCA and ISIS proposal no 397 and fromthe Programme National de Teacuteleacutedeacutetection Spatiale E Berthieracknowledges M Masiokas and R Villalba for welcoming him

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1283

at the IANIGLA (Mendoza Argentina) Y Arnaud acknowledgessupport from French National Research Agency through ANR-09-CEP-005-01PAPRIKA A Kaumlaumlb acknowledges support fromESA through Glaciers_cci (400010177810I-AM) and from theEuropean Research Council under the European Unionrsquos SeventhFramework Programme (FP2007-2013)ERC Grant Agreementn 320816

Edited by I M Howat

The publication of this articleis financed by CNRS-INSU

References

Adler R Huffman G Chang A Ferraro R Xie P JanowiakJ Rudolf B Schneider U Curtis S Bolvin D Gruber ASusskind J Arkin P and Nelkin E The Version-2 GlobalPrecipitation Climatology Project (GPCP) Monthly Precipita-tion Analysis (1979ndashPresent) J Hydrometeorol 4 1147ndash11672003

Archer D R and Fowler H J Spatial and temporal variationsin precipitation in the Upper Indus Basin global teleconnectionsand hydrological implications Hydrol Earth Syst Sci 8 47ndash61doi105194hess-8-47-2004 2004

Arendt A Bolch T Cogley J G Gardner A Hagen J-OHock R Kaser G Pfeffer W T Moholdt G Paul F RadicV Andreassen L Bajracharya S Beedle M Berthier EBhambri R Bliss A Brown I Burgess E Burgess DCawkwell F Chinn T Copland L Davies B de AngelisH Dolgova E Filbert K Forester R Fountain A Frey HGiffen B Glasser N Gurney S Hagg W Hall D Hari-tashya U K Hartmann G Helm C Herreid S Howat IKapustin G Khromova T Kienholz C Koenig M KohlerJ Kriegel D Kutuzov S Lavrentiev I LeBris R Lund JManley W Mayer C Miles E Li X Menounos B Mer-cer A Moelg N Mool P Nosenko G Negrete A NuthC Pettersson R Racoviteanu A Ranzi R Rastner P RauF Rich J Rott H Schneider C Seliverstov Y Sharp MSigurethsson O Stokes C Wheate R Winsvold S WolkenG Wyatt F and Zheltyhina N Randolph Glacier Inventory[v20] A Dataset of Global Glacier Outlines Global Land IceMeasurements from Space Boulder Colorado USA Digital Me-dia httpwwwglimsorgRGIrandolphhtml 2012

Azam M Wagnon P Ramanathan A Vincent C Sharma PArnaud Y Linda A Pottakkal J Chevallier P Singh V andBerthier E From balance to imbalance a shift in the dynamicbehaviour of Chhota Shigri glacier western Himalaya India JGlaciol 58 315ndash324 doi1031892012JoG11J123 2012

Azam M F Wagnon P Vincent C Ramanathan A Linda Aand Singh V B Reconstruction of the annual mass balanceof Chhota Shigri Glacier (Western Himalaya India) since 1969Ann Glaciol in revision 2013

Barrand N and Murray T Multivariate Controls on the In-cidence of Glacier Surging in the Karakoram Himalaya

Arct Antarct Alp Res 38 489ndash498 doi1016571523-0430(2006)38[489MCOTIO]20CO2 2006

Barundun M Huss M Azisov E Gafurov A HoelzleM Merkushkin A Salzmann N and Usubaliev R Re-establishing seasonal mass balance observation at AbramovGlacier Kyrgyzstan from 1968ndash2012 Abstract EGU2013-5668European Geophysical Union Vienna 2013

Benn D Bolch T Hands K Gulley J Luckman ANicholson L Quincey D Thompson S Toumi R andWiseman S Response of debris-covered glaciers in theMount Everest region to recent warming and implicationsfor outburst flood hazards Earth-Sci Rev 114 156ndash174doi101016jearscirev201203008 2012

Berthier E and Vincent C Relative contribution of surface mass-balance and ice-flux changes to the accelerated thinning of Merde Glace French Alps over 1979ndash2008 J Glaciol 58 501ndash512doi1031892012JoG11J083 2012

Berthier E Arnaud Y Baratoux D Vincent C and Reacutemy FRecent rapid thinning of the ldquo Mer de Glace rdquo glacier derivedfrom satellite optical images Geophys Res Lett 31 L17401doi1010292004GL020706 2004

Berthier E Arnaud Y Kumar R Ahmad S Wagnon P andChevallier P Remote sensing estimates of glacier mass balancesin the Himachal Pradesh (Western Himalaya India) RemoteSens Environ 108 327ndash338 doi101016jrse2006110172007

Bhambri R Bolch T Kawishwar P Dobhal D P SrivastavaD and Pratap B Heterogeneity in Glacier response from 1973to 2011 in the Shyok valley Karakoram India The CryosphereDiscuss 6 3049ndash3078 doi105194tcd-6-3049-2012 2012

Bhutiyani M Mass-balance studies on Siachen Glacier in theNubra valley Karakoram Himalaya India J Glaciol 45 112ndash118 1999

Bishop M P Schroder J F and Ward J L SPOT multispec-tral analysis for producing supraglacial debris-load estimates forBatura glacier Pakistan Geocarto Int 10 81ndash90 1995

Bolch T Pieczonka T and Benn D I Multi-decadal mass lossof glaciers in the Everest area (Nepal Himalaya) derived fromstereo imagery The Cryosphere 5 349ndash358 doi105194tc-5-349-2011 2011

Bolch T Kulkarni A Kaumlaumlb A Huggel C Paul F Cogley JFrey H Kargel J Fujita K Scheel M Bajracharya S andStoffel M The state and fate of Himalayan Glaciers Science336 310ndash314 doi101126science1215828 2012

Bollasina M Ming Y and Ramaswamy V Anthropogenicaerosols and the weakening of the South Asian summer mon-soon Science 334 502ndash505 doi101126science12049942011

Bookhagen B and Burbank D Toward a complete Himalayan hy-drological budget spatiotemporal distribution of snowmelt andrainfall and their impact on river discharge J Geophys Res115 F03019 doi1010292009JF001426 2010

Bretherton C Widmann M Dymnikov V Wallace J and BladeacuteI The effective number of spatial degrees of freedom of a time-varying field J Climate 12 1990ndash2009 1999

Copland L Pope S Bishop M Shroder J Clendon PBush A Kamp U Seong Y and Owen L Glacier veloc-ities across the central Karakoram Ann Glaciol 50 41ndash49doi103189172756409789624229 2009

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1279

Fig 11 Elevation changes (SPOT5-SRTM) in the ablation area for clean ice (blue circles) and debris-covered ice (black circles) for eachstudy site For the Karakoram (east and west) and the Pamir study sites only non-surge-type glaciers are considered Standard error of theelevation differences are typically less thanplusmn2 m with each 100 m elevation band (not shown for the sake of clarity) Note elevation changesover separate sections of a glacier cannot be treated as mass changes due to the disregard of glacier dynamics

estimate over this short time period The differences in sur-vey periods and in the glacier outlines may also explainpart of these discrepancies In particular the delimitation ofthe accumulation area is notoriously difficult and Bolch etal (2011) did not include some of the steep high altitudeslopes which we included due to their potential avalanchecontribution For the 10 glaciers mentioned above our massbalances would become 012 m we yrminus1 more negative ifwe used the exact same outlines as Bolch et al (2011)Our positive mass balance for the small Chukhung Glacier(+044plusmn 024 m we yrminus1) is in agreement with Nuimura etal (2012) but enigmatic in a context of regional glacier lossand deserves further investigation If the 2002ndash2007 massbalance estimate by Bolch et al (2011) is excluded all otherrecent mass balance estimates suggest no acceleration of themass loss in the Everest area when compared to the long-term(1970ndash2007) mass balance measured by Bolch et al (2011)in this regionminus032plusmn 008 m weyrminus1

In Bhutan our results indicate a stronger mass loss forsouthern glaciers than for northern ones Interestingly in thisarea Kaumlaumlb (2005) measured high speeds (up to 200 m yrminus1)

for glaciers north of the Himalayan main ridge and ratherlow velocities over southern glacier tongues by using repeat

ASTER data This is consistent with the more negative massbalance that we report for the southern glaciers in Bhutanie for the slow flowing presumably downwasting glaciers

Berthier et al (2007) reported a mass balance betweenminus069 andminus085 m we yrminus1 for an ice-covered area of915 km2 around Chhota Shigri Glacier between 1999 (SRTMDEM) and 2004 (SPOT5-HRG DEM) For the same areawe found a mass balance ofminus045plusmn 013 m we yrminus1 over1999ndash2011 The difference in the study period may explainpart of the differences Also the methodology by Berthieret al (2007) did not explicitly take into account the SRTMradar penetration over glaciers and also corrected an appar-ent elevation-dependent bias according to glacier elevationswhich is now known as a resolution issue and is best cor-rected according to the terrain maximum curvature (Gardelleet al 2012b) More details and discussions about these dif-ferences can be found in Vincent et al (2013)

Importantly the results of the eastern Karakoram studysite confirm the balanced or slightly positive mass budgetreported by Gardelle et al (2012a) in the western Karako-ram over the past decade The two Karakoram study sitesare adjacent with a small overlapping area (sim 1100 km2 ofglaciers) where elevation differences agree within 1 m Both

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1280 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 12 Comparison of geodetic mass balances for 10 glaciers inthe Mount Everest area 1999ndash2010 mass balances measured inthe present study are compared with published estimates during2002ndash2007 (Bolch et al 2011) and during 2000ndash2008 (Nuimuraet al 2012) The 1-to-1 line is drawn The mean difference (plusmn1standard deviation) between (i) our values and Bolch et al (2011)are (047plusmn 058 m we yrminus1) and (ii) our values and Nuimura etal (2012) are (012plusmn 021 m we yrminus1) Part of the differences be-tween estimates may be due to the different periods surveyed andthe different glacier outlines used

sites show similar mass balances over slightly different pe-riods +011plusmn 014 m yrminus1 over 1999ndash2010 in the east and+009plusmn 018 m yrminus1 over 1999ndash2008 in the west This con-firms a ldquoKarakoram anomalyrdquo that was first suggested basedon field observations (Hewitt 2005) We report a mass bal-ance of+010plusmn 013 m yrminus1 for Baltoro Glacier (see loca-tion in Fig 8 size= 304 km2) between 1999 and 2010which is consistent with its gradual speed-up reported since2003 (Quincey et al 2009) However given the completelydifferent methods used in our and the latter study we can-not conclude that this demonstrates a real shift from nega-tive to positive mass balance The mass budget of SiachenGlacier (1145 km2 sometimes referred to as the highestbattlefield in the world) is also balanced or slightly pos-itive (+014plusmn 014 m we yrminus1) Thus the alarming state-ment by Rasul (2012) that this glacier retreated by 59 kmand lost 17 of its mass during 1989ndash2009 is very likelyerroneous Again within error bars our regional mass bal-ance in the Karakoram agrees with the mass balance ofminus003plusmn 004 m yrminus1 found by Kaumlaumlb et al (2012) over 2003ndash2008 and with the mass balance ofminus010plusmn 018 m we yrminus1

(2-sigma error level) obtained by Gardner et al (2013) for aregion including both the Karakoram and the Hindu Kush

In the northernmost part of the Pamir in the Alay moun-tain range the mass balance of Abramov Glacier has beenmeasured in the field for 30 yr between 1968 and 1998(WGMS 2012) with a mean value ofminus046 m yrminus1 Herewe report a mass balance ofminus003plusmn 014 m yrminus1 for this

glacier over 1999ndash2011 Our near zero mass budget estimatefor this glacier disagrees with negative mass balances recon-structed for the same period with a model run using biascorrected NCEPNCAR re-analysis data and calibrated withfield data (Barundun et al 2013) An underestimate of ourSRTM C-Band penetration in the Pamir as a whole and inparticular for the small Abramov Glacier may contribute tothese differences

For the whole Pamir study site our mass budget is bal-anced or slightly positive (+014plusmn 013 m we yrminus1) In theeastern Pamir the mass balance of Muztag Ata Glacier was+025 m we yrminus1 between 2005 and 2010 (Yao et al 2012)Although Muztag Ata Glacier is located outside of our Pamirstudy site and his glaciological records only cover the sec-ond half of our study period this measurement is consistentwith our remotely sensed mass balance Thus the ldquoKarako-ram anomalyrdquo should probably be renamed the ldquoPamir-Karakoram anomalyrdquo

This slightly positive or balanced mass budget measuredin the Pamir seems to contradict previous findings by Heidand Kaumlaumlb (2012) They reported rapidly decreasing glacierspeeds in this region between two snapshots of annual ve-locities in 20002001 and 20092010 an observation thatwould fit with negative mass balances (Span and Kuhn 2003Azam et al 2012) We note though that the relationship be-tween glacier mass balance and ice fluxes is not straightfor-ward and might in particular involve a time delay betweena change in mass balance and the related dynamic response(Heid and Kaumlaumlb 2012) Thus the decrease in speed couldwell be due to negative mass balances before or partiallybefore 1999ndash2011 We acknowledge that this discrepancyneeds to be investigated further in particular by examiningcontinuous changes in annual glacier speed through the firstdecade of the 21st Century Indeed Heid and Kaumlaumlb (2012)measured differences between two annual periods which maynot represent mean decadal velocity changes

Jacob et al (2012) using satellite gravimetry (GRACE)measured a mass budget ofminus5plusmn 6 Gt yrminus1 over 2003ndash2010for the ldquoKarakoram-Himalayardquo their region 8c which cor-responds to our study area excluding the Pamir Karakoramand Hindu Kush sub-regions but including some glaciersnorth of the Himalayan ridge already on the Tibetan PlateauFor this subset of our PKH study area we measured amass budget ofminus126plusmn 42 Gt yrminus1 The two estimates over-lap within their error bars especially if our error is dou-bled to match the two standard error level used by Jacob etal (2012) The estimate of Jacob et al (2012) over our com-plete study region (PKH) ieminus6plusmn 8 Gt yrminus1 (sum of theirregions 8b and 8c) also includes mass changes for the KunlunShan sub-region for which we did not measure glacier massbalances Therefore the comparison with our PKH value(minus101plusmn 55 Gt yrminus1) is not straightforward but suggests areasonable agreement especially if we consider the slightmass gain (15plusmn 17 Gt yrminus1) measured with ICESat in theWest Kunlun (Gardner et al 2013)

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1281

Table 6 Annual average of the glacier seasonal and decadal mass loss contributions to Indus Ganges and Brahmaputra discharges Basinarea and seasonal contributions are given by Kaser et al (2010) glacier area in each basin is derived from the RGI v20 (Arendt et al 2012)Numbers in parentheses indicate the percentage of glacierized area within the river basin Seasonal contributions were modeled by Kaser etal (2010) based on the CRU 20 dataset over 1961ndash1990 assuming a balanced glacier mass budget Glacier decadal mass loss contributionsare given as annual average over 1999ndash2011

River Basin area Glacier area Glacier contribution Mean discharge(km2) (km2) (m3 sminus1) (m3 sminus1)

Seasonally delayed Decadal mass lossKaser et al (2010) (this study)

Indus 1 139 814 25 598 (2 ) 105 103plusmn 72 2146 (Chevallier personalcommunication 2012)

Ganges 1 023 609 11 168 (1 ) 47 103plusmn 49 11739 Papa et al (2012Brahmaputra 527 666 15 296 (3 ) 33 147plusmn 64 19710 updated)

54 Reasons for the heterogeneous pattern of massbalance

The PKH glacier mass balance is slightly negative(minus014plusmn 008 m we yrminus1) but changes are not homoge-neous from one sub-region to another and seem to coincidewith the climatic settings of the mountain range For the east-ern sites (from the Hengduan Shan to West Nepal) which areunder the influence of the Indian and south-east Asian sum-mer monsoons the mass balances are moderately negative(sim minus026 m we yrminus1) In the northwest of the PKH at thePamir and Karakoram sites where glacier climate is char-acterized by the westerlies in winter mass balances are bal-anced or slightly positive (sim +010 m we yrminus1) In betweenthese two influences the glaciers of the Spiti Lahaul sitein a monsoon-arid transition zone experienced the strongestmass loss (minus045plusmn 013 m we yrminus1)

This heterogeneous pattern of mass balances seems to berelated to trends in precipitation throughout the PKH Archerand Fowler (2004) reported an increase in winter precipi-tation over the Upper Indus Basin since 1961 a trend con-firmed by Yao et al (2012) over the Pamir and the Karako-ram based on the analysis of the Global Precipitation Cli-matology Project data set (GPCP Adler et al 2003) Thisprecipitation increase is a source of greater accumulation forglaciers and thus a possible explanation for their slightly pos-itive mass budgets In addition in the Upper Indus Basin themean summer temperature has been decreasing since 1961(Fowler and Archer 2006) which is consistent with the find-ings of Shekhar et al (2010) in the Karakoram who reporteda decrease in both maximum and minimum temperature since1988 These increases in winter precipitation and summertemperature likely translate in greater accumulation and re-duced ablation changes which are consistent with the bal-anced or slightly positive glacier mass budget

On the other hand in the east of the PKH the Indiansummer monsoon has been weakening since the 1950s (Bol-lasina et al 2011) which could have reduced the amount

of snowfall over the eastern study sites and thus resultedin negative mass balances In addition maximum and min-imum temperatures increased since 1988 in the western Hi-malaya (Shekhar et al 2010) and maximum temperaturesincreased in the central Himalaya (Nepal) since the mid-1970s (Shrestha et al 1999) Thus both precipitation andtemperature trends in the Himalaya likely contributed to thenegative mass balances measured over the correspondingstudy sites (from Spiti Lahaul to Hengduan Shan Fig 1)

However gridded data sets such as GPCP are not neces-sarily representative of the climate at high altitude IndeedHewitt (2005) reported a 5-to-10-fold increase in precipi-tation between the glacier front and the accumulation areain the Karakoram that is not captured by reanalysis dataas recently confirmed by Immerzeel et al (2012) Thus di-rect measurements of accumulation on High Mountain Asiaglaciers (eg Azam et al 2013) and high-altitude weatherstations are eagerly needed to validate the large scale grid-ded data set (eg reanalysis) test their ability to describe thespecific climate near to PKH glaciers and assess the relativerole played by temperature and precipitation changes

55 Contribution to water resources

Our glacier mass balance can be averaged over large riverbasins to estimate the mean contribution to river runoffdue to a decade of glacier mass loss We assume therebyonly direct river runoff without loss terms For the threerivers for which we cover the entire glacierized area of theircatchments within our sub-regions (Fig 1) these contribu-tions to the river discharge are equal to 103 m3 sminus1 (Indus)103 m3 sminus1 (Ganges) and 147 m3 sminus1 (Brahmaputra) for theperiod 1999ndash2011 (Table 6)

Rivers of PKH being key water resources measures oftheir discharges are highly sensitive and difficult to access forpolitical reasons Papa et al (2012) built recent time series ofdischarges for Ganges and Brahmaputra by using altimetrymeasurements The locations of the corresponding dischargeestimates in Bangladesh are given in Fig 1 We can therefore

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1282 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

evaluate the relative contribution of glacier decadal mass loss(Table 6) to annual river run-off at 09 for the Gangesat Hardinge (basin area ofsim 1 020 000 km2) and 07 forBrahmaputra at Bahadurabad (basin area ofsim 530 000 km2)between 1999 and 2011 Given the discharge measured at theGuddu station by the Surface Water Hydrology Project of theWater and Power Development Authority (Pakistan see lo-cation on Fig 1) between 1999 and 2003 we can estimatethe contribution of glacier mass loss at 48 for the IndusRiver at this station

The seasonal contributions of glaciers to river run-off (iethe part of the annual precipitation that experiences season-ally delayed release from the glaciers) directly results fromseasonal variations of meteorological quantities in contrastto the glacier mass loss contribution which results from thelong-term climatic change Both runoff components directlyaffect the amount of water available in a river at a given lo-cation and at a given time so that their comparison could beof interest Even if the seasonal timing of the glacier massloss contribution is unclear it could in general peak at asimilar time of the year than the seasonal glacier contribu-tion so that both components rather enlarge each other thancancel Seasonal glacier contributions have been modeled byKaser et al (2010) The latter study assumed glaciers to bein equilibrium with the present climate and thus prescribedglacier mass balances equal to 0 Thus their seasonally de-layed glacier contributions do not include the part due todecadal glacier mass loss and is complementary to our es-timates (Table 6) For the eastern basins under the influenceof the Indian summer monsoon (Ganges and Brahmaputra)the contribution of glacier mass loss is higher than the sea-sonal contribution In other words for the first decade of the21st century the mass loss of glaciers is on average moreimportant for water availability in those two rivers than theirseasonal water storage effect The situation is different for theIndus Basin where the glacier melt season is also the dry sea-son which results in a relatively higher seasonally delayedglacier contribution to the river discharge There the season-ally delayed water release from the glaciers and the decadalglacier mass loss contribute on average equally to the riverdischarge Both contributions are strongly (and equally) de-pendent on the location of the discharge measurement andwill be significantly larger (in relative term) in more heav-ily glacierized and smaller mountain catchments located inthe upper part of these major river basins (Bookhagen andBurbank 2010)

6 Conclusion

In this study we assessed the spatial pattern of glaciermass balance over the PKH by comparing the SRTM andSPOT5 DEMs over 9 well-distributed study sites between1999 and 2011 We found slightly positive or balanced massbudgets in the western part (Pamir Karakoram) moder-ate mass loss in the east (Nepal Bhutan Hengduan Shan)

and the most negative mass balance in the monsoon-aridtransition zone (Spiti Lahaul in the western Himalaya)By extrapolating these values to climatically homogeneoussub-regions we estimate the PKH overall mass balanceto have beenminus014plusmn 008 m we yrminus1 between 1999 and2011 which corresponds to a contribution to sea level riseof 0028plusmn 0015 mm yrminus1 This is about 4 of the globalglacier mass loss estimated by Gardner et al (2013) thoughover a slightly shorter time period (2003ndash2009) The largestsource of uncertainties in those geodetic mass balance esti-mates is the poorly constrained penetration of the C-BandSRTM radar signal into snow and ice

Our technique of DEM differencing allows mapping in de-tail the spatial pattern of glacier elevation changes Thosemaps reveal many surge-type behaviors both in the Pamirand the Karakoram It also appears that the glacier-wide massbalance does not seem to be considerably affected by themass transfer from surges over thesim 10 yr time period stud-ied Similar lowering rates over debris-covered and clean iceare observed for four study sites despite the commonly as-sumed insulating effect of debris covers This suggests forthe scale of entire glacier tongues a spatial mixture of re-duced ablation under intact thick debris covers and enhancedablation by supraglacial lakes or ice cliffs or due to thin de-bris layer and very low glacier velocities in ablation areasthat reduce the ice advection downstream

The regionally heterogeneous pattern of ice wastage is inbroad agreement with contrasted trends in large-scale pre-cipitation and temperature However glaciological (eg ac-cumulation) and meteorological records are needed at highaltitude in the vicinity of glaciers to evaluate more closelythe local influence of atmospheric variables on glacier massbalance and fully understand the reasons behind those con-trasted mass budgets in the PKH

Supplementary material related to this article isavailable online at httpwwwthe-cryospherenet712632013tc-7-1263-2013-supplementpdf

Acknowledgements We thank G Cogley A Gardner T NuimuraT Bolch P Wagnon M Barandun M Hoelzle M Huss and ananonymous referee for their constructive comments that led to aconsiderably improved manuscript We thank the Water and PowerDevelopment Authority (WAPDA) for their hydrological data(processed by P Chevallier) and F Papa for sharing his updatedrun-off data ahead of publication We thank T Bolch for sharinghis glacier outlines from the Everest area We thank the USGSfor allowing free access to their Landsat archive CGIAR-SCI forSRTM C-band data DLR for SRTM X-band data and GLIMS forglacier inventories J Gardelle acknowledges a PhD fellowshipfrom the French Space Agency (CNES) and the French NationalResearch Center (CNRS) E Berthier acknowledges support fromCNES through the TOSCA and ISIS proposal no 397 and fromthe Programme National de Teacuteleacutedeacutetection Spatiale E Berthieracknowledges M Masiokas and R Villalba for welcoming him

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1283

at the IANIGLA (Mendoza Argentina) Y Arnaud acknowledgessupport from French National Research Agency through ANR-09-CEP-005-01PAPRIKA A Kaumlaumlb acknowledges support fromESA through Glaciers_cci (400010177810I-AM) and from theEuropean Research Council under the European Unionrsquos SeventhFramework Programme (FP2007-2013)ERC Grant Agreementn 320816

Edited by I M Howat

The publication of this articleis financed by CNRS-INSU

References

Adler R Huffman G Chang A Ferraro R Xie P JanowiakJ Rudolf B Schneider U Curtis S Bolvin D Gruber ASusskind J Arkin P and Nelkin E The Version-2 GlobalPrecipitation Climatology Project (GPCP) Monthly Precipita-tion Analysis (1979ndashPresent) J Hydrometeorol 4 1147ndash11672003

Archer D R and Fowler H J Spatial and temporal variationsin precipitation in the Upper Indus Basin global teleconnectionsand hydrological implications Hydrol Earth Syst Sci 8 47ndash61doi105194hess-8-47-2004 2004

Arendt A Bolch T Cogley J G Gardner A Hagen J-OHock R Kaser G Pfeffer W T Moholdt G Paul F RadicV Andreassen L Bajracharya S Beedle M Berthier EBhambri R Bliss A Brown I Burgess E Burgess DCawkwell F Chinn T Copland L Davies B de AngelisH Dolgova E Filbert K Forester R Fountain A Frey HGiffen B Glasser N Gurney S Hagg W Hall D Hari-tashya U K Hartmann G Helm C Herreid S Howat IKapustin G Khromova T Kienholz C Koenig M KohlerJ Kriegel D Kutuzov S Lavrentiev I LeBris R Lund JManley W Mayer C Miles E Li X Menounos B Mer-cer A Moelg N Mool P Nosenko G Negrete A NuthC Pettersson R Racoviteanu A Ranzi R Rastner P RauF Rich J Rott H Schneider C Seliverstov Y Sharp MSigurethsson O Stokes C Wheate R Winsvold S WolkenG Wyatt F and Zheltyhina N Randolph Glacier Inventory[v20] A Dataset of Global Glacier Outlines Global Land IceMeasurements from Space Boulder Colorado USA Digital Me-dia httpwwwglimsorgRGIrandolphhtml 2012

Azam M Wagnon P Ramanathan A Vincent C Sharma PArnaud Y Linda A Pottakkal J Chevallier P Singh V andBerthier E From balance to imbalance a shift in the dynamicbehaviour of Chhota Shigri glacier western Himalaya India JGlaciol 58 315ndash324 doi1031892012JoG11J123 2012

Azam M F Wagnon P Vincent C Ramanathan A Linda Aand Singh V B Reconstruction of the annual mass balanceof Chhota Shigri Glacier (Western Himalaya India) since 1969Ann Glaciol in revision 2013

Barrand N and Murray T Multivariate Controls on the In-cidence of Glacier Surging in the Karakoram Himalaya

Arct Antarct Alp Res 38 489ndash498 doi1016571523-0430(2006)38[489MCOTIO]20CO2 2006

Barundun M Huss M Azisov E Gafurov A HoelzleM Merkushkin A Salzmann N and Usubaliev R Re-establishing seasonal mass balance observation at AbramovGlacier Kyrgyzstan from 1968ndash2012 Abstract EGU2013-5668European Geophysical Union Vienna 2013

Benn D Bolch T Hands K Gulley J Luckman ANicholson L Quincey D Thompson S Toumi R andWiseman S Response of debris-covered glaciers in theMount Everest region to recent warming and implicationsfor outburst flood hazards Earth-Sci Rev 114 156ndash174doi101016jearscirev201203008 2012

Berthier E and Vincent C Relative contribution of surface mass-balance and ice-flux changes to the accelerated thinning of Merde Glace French Alps over 1979ndash2008 J Glaciol 58 501ndash512doi1031892012JoG11J083 2012

Berthier E Arnaud Y Baratoux D Vincent C and Reacutemy FRecent rapid thinning of the ldquo Mer de Glace rdquo glacier derivedfrom satellite optical images Geophys Res Lett 31 L17401doi1010292004GL020706 2004

Berthier E Arnaud Y Kumar R Ahmad S Wagnon P andChevallier P Remote sensing estimates of glacier mass balancesin the Himachal Pradesh (Western Himalaya India) RemoteSens Environ 108 327ndash338 doi101016jrse2006110172007

Bhambri R Bolch T Kawishwar P Dobhal D P SrivastavaD and Pratap B Heterogeneity in Glacier response from 1973to 2011 in the Shyok valley Karakoram India The CryosphereDiscuss 6 3049ndash3078 doi105194tcd-6-3049-2012 2012

Bhutiyani M Mass-balance studies on Siachen Glacier in theNubra valley Karakoram Himalaya India J Glaciol 45 112ndash118 1999

Bishop M P Schroder J F and Ward J L SPOT multispec-tral analysis for producing supraglacial debris-load estimates forBatura glacier Pakistan Geocarto Int 10 81ndash90 1995

Bolch T Pieczonka T and Benn D I Multi-decadal mass lossof glaciers in the Everest area (Nepal Himalaya) derived fromstereo imagery The Cryosphere 5 349ndash358 doi105194tc-5-349-2011 2011

Bolch T Kulkarni A Kaumlaumlb A Huggel C Paul F Cogley JFrey H Kargel J Fujita K Scheel M Bajracharya S andStoffel M The state and fate of Himalayan Glaciers Science336 310ndash314 doi101126science1215828 2012

Bollasina M Ming Y and Ramaswamy V Anthropogenicaerosols and the weakening of the South Asian summer mon-soon Science 334 502ndash505 doi101126science12049942011

Bookhagen B and Burbank D Toward a complete Himalayan hy-drological budget spatiotemporal distribution of snowmelt andrainfall and their impact on river discharge J Geophys Res115 F03019 doi1010292009JF001426 2010

Bretherton C Widmann M Dymnikov V Wallace J and BladeacuteI The effective number of spatial degrees of freedom of a time-varying field J Climate 12 1990ndash2009 1999

Copland L Pope S Bishop M Shroder J Clendon PBush A Kamp U Seong Y and Owen L Glacier veloc-ities across the central Karakoram Ann Glaciol 50 41ndash49doi103189172756409789624229 2009

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

1280 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Fig 12 Comparison of geodetic mass balances for 10 glaciers inthe Mount Everest area 1999ndash2010 mass balances measured inthe present study are compared with published estimates during2002ndash2007 (Bolch et al 2011) and during 2000ndash2008 (Nuimuraet al 2012) The 1-to-1 line is drawn The mean difference (plusmn1standard deviation) between (i) our values and Bolch et al (2011)are (047plusmn 058 m we yrminus1) and (ii) our values and Nuimura etal (2012) are (012plusmn 021 m we yrminus1) Part of the differences be-tween estimates may be due to the different periods surveyed andthe different glacier outlines used

sites show similar mass balances over slightly different pe-riods +011plusmn 014 m yrminus1 over 1999ndash2010 in the east and+009plusmn 018 m yrminus1 over 1999ndash2008 in the west This con-firms a ldquoKarakoram anomalyrdquo that was first suggested basedon field observations (Hewitt 2005) We report a mass bal-ance of+010plusmn 013 m yrminus1 for Baltoro Glacier (see loca-tion in Fig 8 size= 304 km2) between 1999 and 2010which is consistent with its gradual speed-up reported since2003 (Quincey et al 2009) However given the completelydifferent methods used in our and the latter study we can-not conclude that this demonstrates a real shift from nega-tive to positive mass balance The mass budget of SiachenGlacier (1145 km2 sometimes referred to as the highestbattlefield in the world) is also balanced or slightly pos-itive (+014plusmn 014 m we yrminus1) Thus the alarming state-ment by Rasul (2012) that this glacier retreated by 59 kmand lost 17 of its mass during 1989ndash2009 is very likelyerroneous Again within error bars our regional mass bal-ance in the Karakoram agrees with the mass balance ofminus003plusmn 004 m yrminus1 found by Kaumlaumlb et al (2012) over 2003ndash2008 and with the mass balance ofminus010plusmn 018 m we yrminus1

(2-sigma error level) obtained by Gardner et al (2013) for aregion including both the Karakoram and the Hindu Kush

In the northernmost part of the Pamir in the Alay moun-tain range the mass balance of Abramov Glacier has beenmeasured in the field for 30 yr between 1968 and 1998(WGMS 2012) with a mean value ofminus046 m yrminus1 Herewe report a mass balance ofminus003plusmn 014 m yrminus1 for this

glacier over 1999ndash2011 Our near zero mass budget estimatefor this glacier disagrees with negative mass balances recon-structed for the same period with a model run using biascorrected NCEPNCAR re-analysis data and calibrated withfield data (Barundun et al 2013) An underestimate of ourSRTM C-Band penetration in the Pamir as a whole and inparticular for the small Abramov Glacier may contribute tothese differences

For the whole Pamir study site our mass budget is bal-anced or slightly positive (+014plusmn 013 m we yrminus1) In theeastern Pamir the mass balance of Muztag Ata Glacier was+025 m we yrminus1 between 2005 and 2010 (Yao et al 2012)Although Muztag Ata Glacier is located outside of our Pamirstudy site and his glaciological records only cover the sec-ond half of our study period this measurement is consistentwith our remotely sensed mass balance Thus the ldquoKarako-ram anomalyrdquo should probably be renamed the ldquoPamir-Karakoram anomalyrdquo

This slightly positive or balanced mass budget measuredin the Pamir seems to contradict previous findings by Heidand Kaumlaumlb (2012) They reported rapidly decreasing glacierspeeds in this region between two snapshots of annual ve-locities in 20002001 and 20092010 an observation thatwould fit with negative mass balances (Span and Kuhn 2003Azam et al 2012) We note though that the relationship be-tween glacier mass balance and ice fluxes is not straightfor-ward and might in particular involve a time delay betweena change in mass balance and the related dynamic response(Heid and Kaumlaumlb 2012) Thus the decrease in speed couldwell be due to negative mass balances before or partiallybefore 1999ndash2011 We acknowledge that this discrepancyneeds to be investigated further in particular by examiningcontinuous changes in annual glacier speed through the firstdecade of the 21st Century Indeed Heid and Kaumlaumlb (2012)measured differences between two annual periods which maynot represent mean decadal velocity changes

Jacob et al (2012) using satellite gravimetry (GRACE)measured a mass budget ofminus5plusmn 6 Gt yrminus1 over 2003ndash2010for the ldquoKarakoram-Himalayardquo their region 8c which cor-responds to our study area excluding the Pamir Karakoramand Hindu Kush sub-regions but including some glaciersnorth of the Himalayan ridge already on the Tibetan PlateauFor this subset of our PKH study area we measured amass budget ofminus126plusmn 42 Gt yrminus1 The two estimates over-lap within their error bars especially if our error is dou-bled to match the two standard error level used by Jacob etal (2012) The estimate of Jacob et al (2012) over our com-plete study region (PKH) ieminus6plusmn 8 Gt yrminus1 (sum of theirregions 8b and 8c) also includes mass changes for the KunlunShan sub-region for which we did not measure glacier massbalances Therefore the comparison with our PKH value(minus101plusmn 55 Gt yrminus1) is not straightforward but suggests areasonable agreement especially if we consider the slightmass gain (15plusmn 17 Gt yrminus1) measured with ICESat in theWest Kunlun (Gardner et al 2013)

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1281

Table 6 Annual average of the glacier seasonal and decadal mass loss contributions to Indus Ganges and Brahmaputra discharges Basinarea and seasonal contributions are given by Kaser et al (2010) glacier area in each basin is derived from the RGI v20 (Arendt et al 2012)Numbers in parentheses indicate the percentage of glacierized area within the river basin Seasonal contributions were modeled by Kaser etal (2010) based on the CRU 20 dataset over 1961ndash1990 assuming a balanced glacier mass budget Glacier decadal mass loss contributionsare given as annual average over 1999ndash2011

River Basin area Glacier area Glacier contribution Mean discharge(km2) (km2) (m3 sminus1) (m3 sminus1)

Seasonally delayed Decadal mass lossKaser et al (2010) (this study)

Indus 1 139 814 25 598 (2 ) 105 103plusmn 72 2146 (Chevallier personalcommunication 2012)

Ganges 1 023 609 11 168 (1 ) 47 103plusmn 49 11739 Papa et al (2012Brahmaputra 527 666 15 296 (3 ) 33 147plusmn 64 19710 updated)

54 Reasons for the heterogeneous pattern of massbalance

The PKH glacier mass balance is slightly negative(minus014plusmn 008 m we yrminus1) but changes are not homoge-neous from one sub-region to another and seem to coincidewith the climatic settings of the mountain range For the east-ern sites (from the Hengduan Shan to West Nepal) which areunder the influence of the Indian and south-east Asian sum-mer monsoons the mass balances are moderately negative(sim minus026 m we yrminus1) In the northwest of the PKH at thePamir and Karakoram sites where glacier climate is char-acterized by the westerlies in winter mass balances are bal-anced or slightly positive (sim +010 m we yrminus1) In betweenthese two influences the glaciers of the Spiti Lahaul sitein a monsoon-arid transition zone experienced the strongestmass loss (minus045plusmn 013 m we yrminus1)

This heterogeneous pattern of mass balances seems to berelated to trends in precipitation throughout the PKH Archerand Fowler (2004) reported an increase in winter precipi-tation over the Upper Indus Basin since 1961 a trend con-firmed by Yao et al (2012) over the Pamir and the Karako-ram based on the analysis of the Global Precipitation Cli-matology Project data set (GPCP Adler et al 2003) Thisprecipitation increase is a source of greater accumulation forglaciers and thus a possible explanation for their slightly pos-itive mass budgets In addition in the Upper Indus Basin themean summer temperature has been decreasing since 1961(Fowler and Archer 2006) which is consistent with the find-ings of Shekhar et al (2010) in the Karakoram who reporteda decrease in both maximum and minimum temperature since1988 These increases in winter precipitation and summertemperature likely translate in greater accumulation and re-duced ablation changes which are consistent with the bal-anced or slightly positive glacier mass budget

On the other hand in the east of the PKH the Indiansummer monsoon has been weakening since the 1950s (Bol-lasina et al 2011) which could have reduced the amount

of snowfall over the eastern study sites and thus resultedin negative mass balances In addition maximum and min-imum temperatures increased since 1988 in the western Hi-malaya (Shekhar et al 2010) and maximum temperaturesincreased in the central Himalaya (Nepal) since the mid-1970s (Shrestha et al 1999) Thus both precipitation andtemperature trends in the Himalaya likely contributed to thenegative mass balances measured over the correspondingstudy sites (from Spiti Lahaul to Hengduan Shan Fig 1)

However gridded data sets such as GPCP are not neces-sarily representative of the climate at high altitude IndeedHewitt (2005) reported a 5-to-10-fold increase in precipi-tation between the glacier front and the accumulation areain the Karakoram that is not captured by reanalysis dataas recently confirmed by Immerzeel et al (2012) Thus di-rect measurements of accumulation on High Mountain Asiaglaciers (eg Azam et al 2013) and high-altitude weatherstations are eagerly needed to validate the large scale grid-ded data set (eg reanalysis) test their ability to describe thespecific climate near to PKH glaciers and assess the relativerole played by temperature and precipitation changes

55 Contribution to water resources

Our glacier mass balance can be averaged over large riverbasins to estimate the mean contribution to river runoffdue to a decade of glacier mass loss We assume therebyonly direct river runoff without loss terms For the threerivers for which we cover the entire glacierized area of theircatchments within our sub-regions (Fig 1) these contribu-tions to the river discharge are equal to 103 m3 sminus1 (Indus)103 m3 sminus1 (Ganges) and 147 m3 sminus1 (Brahmaputra) for theperiod 1999ndash2011 (Table 6)

Rivers of PKH being key water resources measures oftheir discharges are highly sensitive and difficult to access forpolitical reasons Papa et al (2012) built recent time series ofdischarges for Ganges and Brahmaputra by using altimetrymeasurements The locations of the corresponding dischargeestimates in Bangladesh are given in Fig 1 We can therefore

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1282 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

evaluate the relative contribution of glacier decadal mass loss(Table 6) to annual river run-off at 09 for the Gangesat Hardinge (basin area ofsim 1 020 000 km2) and 07 forBrahmaputra at Bahadurabad (basin area ofsim 530 000 km2)between 1999 and 2011 Given the discharge measured at theGuddu station by the Surface Water Hydrology Project of theWater and Power Development Authority (Pakistan see lo-cation on Fig 1) between 1999 and 2003 we can estimatethe contribution of glacier mass loss at 48 for the IndusRiver at this station

The seasonal contributions of glaciers to river run-off (iethe part of the annual precipitation that experiences season-ally delayed release from the glaciers) directly results fromseasonal variations of meteorological quantities in contrastto the glacier mass loss contribution which results from thelong-term climatic change Both runoff components directlyaffect the amount of water available in a river at a given lo-cation and at a given time so that their comparison could beof interest Even if the seasonal timing of the glacier massloss contribution is unclear it could in general peak at asimilar time of the year than the seasonal glacier contribu-tion so that both components rather enlarge each other thancancel Seasonal glacier contributions have been modeled byKaser et al (2010) The latter study assumed glaciers to bein equilibrium with the present climate and thus prescribedglacier mass balances equal to 0 Thus their seasonally de-layed glacier contributions do not include the part due todecadal glacier mass loss and is complementary to our es-timates (Table 6) For the eastern basins under the influenceof the Indian summer monsoon (Ganges and Brahmaputra)the contribution of glacier mass loss is higher than the sea-sonal contribution In other words for the first decade of the21st century the mass loss of glaciers is on average moreimportant for water availability in those two rivers than theirseasonal water storage effect The situation is different for theIndus Basin where the glacier melt season is also the dry sea-son which results in a relatively higher seasonally delayedglacier contribution to the river discharge There the season-ally delayed water release from the glaciers and the decadalglacier mass loss contribute on average equally to the riverdischarge Both contributions are strongly (and equally) de-pendent on the location of the discharge measurement andwill be significantly larger (in relative term) in more heav-ily glacierized and smaller mountain catchments located inthe upper part of these major river basins (Bookhagen andBurbank 2010)

6 Conclusion

In this study we assessed the spatial pattern of glaciermass balance over the PKH by comparing the SRTM andSPOT5 DEMs over 9 well-distributed study sites between1999 and 2011 We found slightly positive or balanced massbudgets in the western part (Pamir Karakoram) moder-ate mass loss in the east (Nepal Bhutan Hengduan Shan)

and the most negative mass balance in the monsoon-aridtransition zone (Spiti Lahaul in the western Himalaya)By extrapolating these values to climatically homogeneoussub-regions we estimate the PKH overall mass balanceto have beenminus014plusmn 008 m we yrminus1 between 1999 and2011 which corresponds to a contribution to sea level riseof 0028plusmn 0015 mm yrminus1 This is about 4 of the globalglacier mass loss estimated by Gardner et al (2013) thoughover a slightly shorter time period (2003ndash2009) The largestsource of uncertainties in those geodetic mass balance esti-mates is the poorly constrained penetration of the C-BandSRTM radar signal into snow and ice

Our technique of DEM differencing allows mapping in de-tail the spatial pattern of glacier elevation changes Thosemaps reveal many surge-type behaviors both in the Pamirand the Karakoram It also appears that the glacier-wide massbalance does not seem to be considerably affected by themass transfer from surges over thesim 10 yr time period stud-ied Similar lowering rates over debris-covered and clean iceare observed for four study sites despite the commonly as-sumed insulating effect of debris covers This suggests forthe scale of entire glacier tongues a spatial mixture of re-duced ablation under intact thick debris covers and enhancedablation by supraglacial lakes or ice cliffs or due to thin de-bris layer and very low glacier velocities in ablation areasthat reduce the ice advection downstream

The regionally heterogeneous pattern of ice wastage is inbroad agreement with contrasted trends in large-scale pre-cipitation and temperature However glaciological (eg ac-cumulation) and meteorological records are needed at highaltitude in the vicinity of glaciers to evaluate more closelythe local influence of atmospheric variables on glacier massbalance and fully understand the reasons behind those con-trasted mass budgets in the PKH

Supplementary material related to this article isavailable online at httpwwwthe-cryospherenet712632013tc-7-1263-2013-supplementpdf

Acknowledgements We thank G Cogley A Gardner T NuimuraT Bolch P Wagnon M Barandun M Hoelzle M Huss and ananonymous referee for their constructive comments that led to aconsiderably improved manuscript We thank the Water and PowerDevelopment Authority (WAPDA) for their hydrological data(processed by P Chevallier) and F Papa for sharing his updatedrun-off data ahead of publication We thank T Bolch for sharinghis glacier outlines from the Everest area We thank the USGSfor allowing free access to their Landsat archive CGIAR-SCI forSRTM C-band data DLR for SRTM X-band data and GLIMS forglacier inventories J Gardelle acknowledges a PhD fellowshipfrom the French Space Agency (CNES) and the French NationalResearch Center (CNRS) E Berthier acknowledges support fromCNES through the TOSCA and ISIS proposal no 397 and fromthe Programme National de Teacuteleacutedeacutetection Spatiale E Berthieracknowledges M Masiokas and R Villalba for welcoming him

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1283

at the IANIGLA (Mendoza Argentina) Y Arnaud acknowledgessupport from French National Research Agency through ANR-09-CEP-005-01PAPRIKA A Kaumlaumlb acknowledges support fromESA through Glaciers_cci (400010177810I-AM) and from theEuropean Research Council under the European Unionrsquos SeventhFramework Programme (FP2007-2013)ERC Grant Agreementn 320816

Edited by I M Howat

The publication of this articleis financed by CNRS-INSU

References

Adler R Huffman G Chang A Ferraro R Xie P JanowiakJ Rudolf B Schneider U Curtis S Bolvin D Gruber ASusskind J Arkin P and Nelkin E The Version-2 GlobalPrecipitation Climatology Project (GPCP) Monthly Precipita-tion Analysis (1979ndashPresent) J Hydrometeorol 4 1147ndash11672003

Archer D R and Fowler H J Spatial and temporal variationsin precipitation in the Upper Indus Basin global teleconnectionsand hydrological implications Hydrol Earth Syst Sci 8 47ndash61doi105194hess-8-47-2004 2004

Arendt A Bolch T Cogley J G Gardner A Hagen J-OHock R Kaser G Pfeffer W T Moholdt G Paul F RadicV Andreassen L Bajracharya S Beedle M Berthier EBhambri R Bliss A Brown I Burgess E Burgess DCawkwell F Chinn T Copland L Davies B de AngelisH Dolgova E Filbert K Forester R Fountain A Frey HGiffen B Glasser N Gurney S Hagg W Hall D Hari-tashya U K Hartmann G Helm C Herreid S Howat IKapustin G Khromova T Kienholz C Koenig M KohlerJ Kriegel D Kutuzov S Lavrentiev I LeBris R Lund JManley W Mayer C Miles E Li X Menounos B Mer-cer A Moelg N Mool P Nosenko G Negrete A NuthC Pettersson R Racoviteanu A Ranzi R Rastner P RauF Rich J Rott H Schneider C Seliverstov Y Sharp MSigurethsson O Stokes C Wheate R Winsvold S WolkenG Wyatt F and Zheltyhina N Randolph Glacier Inventory[v20] A Dataset of Global Glacier Outlines Global Land IceMeasurements from Space Boulder Colorado USA Digital Me-dia httpwwwglimsorgRGIrandolphhtml 2012

Azam M Wagnon P Ramanathan A Vincent C Sharma PArnaud Y Linda A Pottakkal J Chevallier P Singh V andBerthier E From balance to imbalance a shift in the dynamicbehaviour of Chhota Shigri glacier western Himalaya India JGlaciol 58 315ndash324 doi1031892012JoG11J123 2012

Azam M F Wagnon P Vincent C Ramanathan A Linda Aand Singh V B Reconstruction of the annual mass balanceof Chhota Shigri Glacier (Western Himalaya India) since 1969Ann Glaciol in revision 2013

Barrand N and Murray T Multivariate Controls on the In-cidence of Glacier Surging in the Karakoram Himalaya

Arct Antarct Alp Res 38 489ndash498 doi1016571523-0430(2006)38[489MCOTIO]20CO2 2006

Barundun M Huss M Azisov E Gafurov A HoelzleM Merkushkin A Salzmann N and Usubaliev R Re-establishing seasonal mass balance observation at AbramovGlacier Kyrgyzstan from 1968ndash2012 Abstract EGU2013-5668European Geophysical Union Vienna 2013

Benn D Bolch T Hands K Gulley J Luckman ANicholson L Quincey D Thompson S Toumi R andWiseman S Response of debris-covered glaciers in theMount Everest region to recent warming and implicationsfor outburst flood hazards Earth-Sci Rev 114 156ndash174doi101016jearscirev201203008 2012

Berthier E and Vincent C Relative contribution of surface mass-balance and ice-flux changes to the accelerated thinning of Merde Glace French Alps over 1979ndash2008 J Glaciol 58 501ndash512doi1031892012JoG11J083 2012

Berthier E Arnaud Y Baratoux D Vincent C and Reacutemy FRecent rapid thinning of the ldquo Mer de Glace rdquo glacier derivedfrom satellite optical images Geophys Res Lett 31 L17401doi1010292004GL020706 2004

Berthier E Arnaud Y Kumar R Ahmad S Wagnon P andChevallier P Remote sensing estimates of glacier mass balancesin the Himachal Pradesh (Western Himalaya India) RemoteSens Environ 108 327ndash338 doi101016jrse2006110172007

Bhambri R Bolch T Kawishwar P Dobhal D P SrivastavaD and Pratap B Heterogeneity in Glacier response from 1973to 2011 in the Shyok valley Karakoram India The CryosphereDiscuss 6 3049ndash3078 doi105194tcd-6-3049-2012 2012

Bhutiyani M Mass-balance studies on Siachen Glacier in theNubra valley Karakoram Himalaya India J Glaciol 45 112ndash118 1999

Bishop M P Schroder J F and Ward J L SPOT multispec-tral analysis for producing supraglacial debris-load estimates forBatura glacier Pakistan Geocarto Int 10 81ndash90 1995

Bolch T Pieczonka T and Benn D I Multi-decadal mass lossof glaciers in the Everest area (Nepal Himalaya) derived fromstereo imagery The Cryosphere 5 349ndash358 doi105194tc-5-349-2011 2011

Bolch T Kulkarni A Kaumlaumlb A Huggel C Paul F Cogley JFrey H Kargel J Fujita K Scheel M Bajracharya S andStoffel M The state and fate of Himalayan Glaciers Science336 310ndash314 doi101126science1215828 2012

Bollasina M Ming Y and Ramaswamy V Anthropogenicaerosols and the weakening of the South Asian summer mon-soon Science 334 502ndash505 doi101126science12049942011

Bookhagen B and Burbank D Toward a complete Himalayan hy-drological budget spatiotemporal distribution of snowmelt andrainfall and their impact on river discharge J Geophys Res115 F03019 doi1010292009JF001426 2010

Bretherton C Widmann M Dymnikov V Wallace J and BladeacuteI The effective number of spatial degrees of freedom of a time-varying field J Climate 12 1990ndash2009 1999

Copland L Pope S Bishop M Shroder J Clendon PBush A Kamp U Seong Y and Owen L Glacier veloc-ities across the central Karakoram Ann Glaciol 50 41ndash49doi103189172756409789624229 2009

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1281

Table 6 Annual average of the glacier seasonal and decadal mass loss contributions to Indus Ganges and Brahmaputra discharges Basinarea and seasonal contributions are given by Kaser et al (2010) glacier area in each basin is derived from the RGI v20 (Arendt et al 2012)Numbers in parentheses indicate the percentage of glacierized area within the river basin Seasonal contributions were modeled by Kaser etal (2010) based on the CRU 20 dataset over 1961ndash1990 assuming a balanced glacier mass budget Glacier decadal mass loss contributionsare given as annual average over 1999ndash2011

River Basin area Glacier area Glacier contribution Mean discharge(km2) (km2) (m3 sminus1) (m3 sminus1)

Seasonally delayed Decadal mass lossKaser et al (2010) (this study)

Indus 1 139 814 25 598 (2 ) 105 103plusmn 72 2146 (Chevallier personalcommunication 2012)

Ganges 1 023 609 11 168 (1 ) 47 103plusmn 49 11739 Papa et al (2012Brahmaputra 527 666 15 296 (3 ) 33 147plusmn 64 19710 updated)

54 Reasons for the heterogeneous pattern of massbalance

The PKH glacier mass balance is slightly negative(minus014plusmn 008 m we yrminus1) but changes are not homoge-neous from one sub-region to another and seem to coincidewith the climatic settings of the mountain range For the east-ern sites (from the Hengduan Shan to West Nepal) which areunder the influence of the Indian and south-east Asian sum-mer monsoons the mass balances are moderately negative(sim minus026 m we yrminus1) In the northwest of the PKH at thePamir and Karakoram sites where glacier climate is char-acterized by the westerlies in winter mass balances are bal-anced or slightly positive (sim +010 m we yrminus1) In betweenthese two influences the glaciers of the Spiti Lahaul sitein a monsoon-arid transition zone experienced the strongestmass loss (minus045plusmn 013 m we yrminus1)

This heterogeneous pattern of mass balances seems to berelated to trends in precipitation throughout the PKH Archerand Fowler (2004) reported an increase in winter precipi-tation over the Upper Indus Basin since 1961 a trend con-firmed by Yao et al (2012) over the Pamir and the Karako-ram based on the analysis of the Global Precipitation Cli-matology Project data set (GPCP Adler et al 2003) Thisprecipitation increase is a source of greater accumulation forglaciers and thus a possible explanation for their slightly pos-itive mass budgets In addition in the Upper Indus Basin themean summer temperature has been decreasing since 1961(Fowler and Archer 2006) which is consistent with the find-ings of Shekhar et al (2010) in the Karakoram who reporteda decrease in both maximum and minimum temperature since1988 These increases in winter precipitation and summertemperature likely translate in greater accumulation and re-duced ablation changes which are consistent with the bal-anced or slightly positive glacier mass budget

On the other hand in the east of the PKH the Indiansummer monsoon has been weakening since the 1950s (Bol-lasina et al 2011) which could have reduced the amount

of snowfall over the eastern study sites and thus resultedin negative mass balances In addition maximum and min-imum temperatures increased since 1988 in the western Hi-malaya (Shekhar et al 2010) and maximum temperaturesincreased in the central Himalaya (Nepal) since the mid-1970s (Shrestha et al 1999) Thus both precipitation andtemperature trends in the Himalaya likely contributed to thenegative mass balances measured over the correspondingstudy sites (from Spiti Lahaul to Hengduan Shan Fig 1)

However gridded data sets such as GPCP are not neces-sarily representative of the climate at high altitude IndeedHewitt (2005) reported a 5-to-10-fold increase in precipi-tation between the glacier front and the accumulation areain the Karakoram that is not captured by reanalysis dataas recently confirmed by Immerzeel et al (2012) Thus di-rect measurements of accumulation on High Mountain Asiaglaciers (eg Azam et al 2013) and high-altitude weatherstations are eagerly needed to validate the large scale grid-ded data set (eg reanalysis) test their ability to describe thespecific climate near to PKH glaciers and assess the relativerole played by temperature and precipitation changes

55 Contribution to water resources

Our glacier mass balance can be averaged over large riverbasins to estimate the mean contribution to river runoffdue to a decade of glacier mass loss We assume therebyonly direct river runoff without loss terms For the threerivers for which we cover the entire glacierized area of theircatchments within our sub-regions (Fig 1) these contribu-tions to the river discharge are equal to 103 m3 sminus1 (Indus)103 m3 sminus1 (Ganges) and 147 m3 sminus1 (Brahmaputra) for theperiod 1999ndash2011 (Table 6)

Rivers of PKH being key water resources measures oftheir discharges are highly sensitive and difficult to access forpolitical reasons Papa et al (2012) built recent time series ofdischarges for Ganges and Brahmaputra by using altimetrymeasurements The locations of the corresponding dischargeestimates in Bangladesh are given in Fig 1 We can therefore

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1282 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

evaluate the relative contribution of glacier decadal mass loss(Table 6) to annual river run-off at 09 for the Gangesat Hardinge (basin area ofsim 1 020 000 km2) and 07 forBrahmaputra at Bahadurabad (basin area ofsim 530 000 km2)between 1999 and 2011 Given the discharge measured at theGuddu station by the Surface Water Hydrology Project of theWater and Power Development Authority (Pakistan see lo-cation on Fig 1) between 1999 and 2003 we can estimatethe contribution of glacier mass loss at 48 for the IndusRiver at this station

The seasonal contributions of glaciers to river run-off (iethe part of the annual precipitation that experiences season-ally delayed release from the glaciers) directly results fromseasonal variations of meteorological quantities in contrastto the glacier mass loss contribution which results from thelong-term climatic change Both runoff components directlyaffect the amount of water available in a river at a given lo-cation and at a given time so that their comparison could beof interest Even if the seasonal timing of the glacier massloss contribution is unclear it could in general peak at asimilar time of the year than the seasonal glacier contribu-tion so that both components rather enlarge each other thancancel Seasonal glacier contributions have been modeled byKaser et al (2010) The latter study assumed glaciers to bein equilibrium with the present climate and thus prescribedglacier mass balances equal to 0 Thus their seasonally de-layed glacier contributions do not include the part due todecadal glacier mass loss and is complementary to our es-timates (Table 6) For the eastern basins under the influenceof the Indian summer monsoon (Ganges and Brahmaputra)the contribution of glacier mass loss is higher than the sea-sonal contribution In other words for the first decade of the21st century the mass loss of glaciers is on average moreimportant for water availability in those two rivers than theirseasonal water storage effect The situation is different for theIndus Basin where the glacier melt season is also the dry sea-son which results in a relatively higher seasonally delayedglacier contribution to the river discharge There the season-ally delayed water release from the glaciers and the decadalglacier mass loss contribute on average equally to the riverdischarge Both contributions are strongly (and equally) de-pendent on the location of the discharge measurement andwill be significantly larger (in relative term) in more heav-ily glacierized and smaller mountain catchments located inthe upper part of these major river basins (Bookhagen andBurbank 2010)

6 Conclusion

In this study we assessed the spatial pattern of glaciermass balance over the PKH by comparing the SRTM andSPOT5 DEMs over 9 well-distributed study sites between1999 and 2011 We found slightly positive or balanced massbudgets in the western part (Pamir Karakoram) moder-ate mass loss in the east (Nepal Bhutan Hengduan Shan)

and the most negative mass balance in the monsoon-aridtransition zone (Spiti Lahaul in the western Himalaya)By extrapolating these values to climatically homogeneoussub-regions we estimate the PKH overall mass balanceto have beenminus014plusmn 008 m we yrminus1 between 1999 and2011 which corresponds to a contribution to sea level riseof 0028plusmn 0015 mm yrminus1 This is about 4 of the globalglacier mass loss estimated by Gardner et al (2013) thoughover a slightly shorter time period (2003ndash2009) The largestsource of uncertainties in those geodetic mass balance esti-mates is the poorly constrained penetration of the C-BandSRTM radar signal into snow and ice

Our technique of DEM differencing allows mapping in de-tail the spatial pattern of glacier elevation changes Thosemaps reveal many surge-type behaviors both in the Pamirand the Karakoram It also appears that the glacier-wide massbalance does not seem to be considerably affected by themass transfer from surges over thesim 10 yr time period stud-ied Similar lowering rates over debris-covered and clean iceare observed for four study sites despite the commonly as-sumed insulating effect of debris covers This suggests forthe scale of entire glacier tongues a spatial mixture of re-duced ablation under intact thick debris covers and enhancedablation by supraglacial lakes or ice cliffs or due to thin de-bris layer and very low glacier velocities in ablation areasthat reduce the ice advection downstream

The regionally heterogeneous pattern of ice wastage is inbroad agreement with contrasted trends in large-scale pre-cipitation and temperature However glaciological (eg ac-cumulation) and meteorological records are needed at highaltitude in the vicinity of glaciers to evaluate more closelythe local influence of atmospheric variables on glacier massbalance and fully understand the reasons behind those con-trasted mass budgets in the PKH

Supplementary material related to this article isavailable online at httpwwwthe-cryospherenet712632013tc-7-1263-2013-supplementpdf

Acknowledgements We thank G Cogley A Gardner T NuimuraT Bolch P Wagnon M Barandun M Hoelzle M Huss and ananonymous referee for their constructive comments that led to aconsiderably improved manuscript We thank the Water and PowerDevelopment Authority (WAPDA) for their hydrological data(processed by P Chevallier) and F Papa for sharing his updatedrun-off data ahead of publication We thank T Bolch for sharinghis glacier outlines from the Everest area We thank the USGSfor allowing free access to their Landsat archive CGIAR-SCI forSRTM C-band data DLR for SRTM X-band data and GLIMS forglacier inventories J Gardelle acknowledges a PhD fellowshipfrom the French Space Agency (CNES) and the French NationalResearch Center (CNRS) E Berthier acknowledges support fromCNES through the TOSCA and ISIS proposal no 397 and fromthe Programme National de Teacuteleacutedeacutetection Spatiale E Berthieracknowledges M Masiokas and R Villalba for welcoming him

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1283

at the IANIGLA (Mendoza Argentina) Y Arnaud acknowledgessupport from French National Research Agency through ANR-09-CEP-005-01PAPRIKA A Kaumlaumlb acknowledges support fromESA through Glaciers_cci (400010177810I-AM) and from theEuropean Research Council under the European Unionrsquos SeventhFramework Programme (FP2007-2013)ERC Grant Agreementn 320816

Edited by I M Howat

The publication of this articleis financed by CNRS-INSU

References

Adler R Huffman G Chang A Ferraro R Xie P JanowiakJ Rudolf B Schneider U Curtis S Bolvin D Gruber ASusskind J Arkin P and Nelkin E The Version-2 GlobalPrecipitation Climatology Project (GPCP) Monthly Precipita-tion Analysis (1979ndashPresent) J Hydrometeorol 4 1147ndash11672003

Archer D R and Fowler H J Spatial and temporal variationsin precipitation in the Upper Indus Basin global teleconnectionsand hydrological implications Hydrol Earth Syst Sci 8 47ndash61doi105194hess-8-47-2004 2004

Arendt A Bolch T Cogley J G Gardner A Hagen J-OHock R Kaser G Pfeffer W T Moholdt G Paul F RadicV Andreassen L Bajracharya S Beedle M Berthier EBhambri R Bliss A Brown I Burgess E Burgess DCawkwell F Chinn T Copland L Davies B de AngelisH Dolgova E Filbert K Forester R Fountain A Frey HGiffen B Glasser N Gurney S Hagg W Hall D Hari-tashya U K Hartmann G Helm C Herreid S Howat IKapustin G Khromova T Kienholz C Koenig M KohlerJ Kriegel D Kutuzov S Lavrentiev I LeBris R Lund JManley W Mayer C Miles E Li X Menounos B Mer-cer A Moelg N Mool P Nosenko G Negrete A NuthC Pettersson R Racoviteanu A Ranzi R Rastner P RauF Rich J Rott H Schneider C Seliverstov Y Sharp MSigurethsson O Stokes C Wheate R Winsvold S WolkenG Wyatt F and Zheltyhina N Randolph Glacier Inventory[v20] A Dataset of Global Glacier Outlines Global Land IceMeasurements from Space Boulder Colorado USA Digital Me-dia httpwwwglimsorgRGIrandolphhtml 2012

Azam M Wagnon P Ramanathan A Vincent C Sharma PArnaud Y Linda A Pottakkal J Chevallier P Singh V andBerthier E From balance to imbalance a shift in the dynamicbehaviour of Chhota Shigri glacier western Himalaya India JGlaciol 58 315ndash324 doi1031892012JoG11J123 2012

Azam M F Wagnon P Vincent C Ramanathan A Linda Aand Singh V B Reconstruction of the annual mass balanceof Chhota Shigri Glacier (Western Himalaya India) since 1969Ann Glaciol in revision 2013

Barrand N and Murray T Multivariate Controls on the In-cidence of Glacier Surging in the Karakoram Himalaya

Arct Antarct Alp Res 38 489ndash498 doi1016571523-0430(2006)38[489MCOTIO]20CO2 2006

Barundun M Huss M Azisov E Gafurov A HoelzleM Merkushkin A Salzmann N and Usubaliev R Re-establishing seasonal mass balance observation at AbramovGlacier Kyrgyzstan from 1968ndash2012 Abstract EGU2013-5668European Geophysical Union Vienna 2013

Benn D Bolch T Hands K Gulley J Luckman ANicholson L Quincey D Thompson S Toumi R andWiseman S Response of debris-covered glaciers in theMount Everest region to recent warming and implicationsfor outburst flood hazards Earth-Sci Rev 114 156ndash174doi101016jearscirev201203008 2012

Berthier E and Vincent C Relative contribution of surface mass-balance and ice-flux changes to the accelerated thinning of Merde Glace French Alps over 1979ndash2008 J Glaciol 58 501ndash512doi1031892012JoG11J083 2012

Berthier E Arnaud Y Baratoux D Vincent C and Reacutemy FRecent rapid thinning of the ldquo Mer de Glace rdquo glacier derivedfrom satellite optical images Geophys Res Lett 31 L17401doi1010292004GL020706 2004

Berthier E Arnaud Y Kumar R Ahmad S Wagnon P andChevallier P Remote sensing estimates of glacier mass balancesin the Himachal Pradesh (Western Himalaya India) RemoteSens Environ 108 327ndash338 doi101016jrse2006110172007

Bhambri R Bolch T Kawishwar P Dobhal D P SrivastavaD and Pratap B Heterogeneity in Glacier response from 1973to 2011 in the Shyok valley Karakoram India The CryosphereDiscuss 6 3049ndash3078 doi105194tcd-6-3049-2012 2012

Bhutiyani M Mass-balance studies on Siachen Glacier in theNubra valley Karakoram Himalaya India J Glaciol 45 112ndash118 1999

Bishop M P Schroder J F and Ward J L SPOT multispec-tral analysis for producing supraglacial debris-load estimates forBatura glacier Pakistan Geocarto Int 10 81ndash90 1995

Bolch T Pieczonka T and Benn D I Multi-decadal mass lossof glaciers in the Everest area (Nepal Himalaya) derived fromstereo imagery The Cryosphere 5 349ndash358 doi105194tc-5-349-2011 2011

Bolch T Kulkarni A Kaumlaumlb A Huggel C Paul F Cogley JFrey H Kargel J Fujita K Scheel M Bajracharya S andStoffel M The state and fate of Himalayan Glaciers Science336 310ndash314 doi101126science1215828 2012

Bollasina M Ming Y and Ramaswamy V Anthropogenicaerosols and the weakening of the South Asian summer mon-soon Science 334 502ndash505 doi101126science12049942011

Bookhagen B and Burbank D Toward a complete Himalayan hy-drological budget spatiotemporal distribution of snowmelt andrainfall and their impact on river discharge J Geophys Res115 F03019 doi1010292009JF001426 2010

Bretherton C Widmann M Dymnikov V Wallace J and BladeacuteI The effective number of spatial degrees of freedom of a time-varying field J Climate 12 1990ndash2009 1999

Copland L Pope S Bishop M Shroder J Clendon PBush A Kamp U Seong Y and Owen L Glacier veloc-ities across the central Karakoram Ann Glaciol 50 41ndash49doi103189172756409789624229 2009

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

1282 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

evaluate the relative contribution of glacier decadal mass loss(Table 6) to annual river run-off at 09 for the Gangesat Hardinge (basin area ofsim 1 020 000 km2) and 07 forBrahmaputra at Bahadurabad (basin area ofsim 530 000 km2)between 1999 and 2011 Given the discharge measured at theGuddu station by the Surface Water Hydrology Project of theWater and Power Development Authority (Pakistan see lo-cation on Fig 1) between 1999 and 2003 we can estimatethe contribution of glacier mass loss at 48 for the IndusRiver at this station

The seasonal contributions of glaciers to river run-off (iethe part of the annual precipitation that experiences season-ally delayed release from the glaciers) directly results fromseasonal variations of meteorological quantities in contrastto the glacier mass loss contribution which results from thelong-term climatic change Both runoff components directlyaffect the amount of water available in a river at a given lo-cation and at a given time so that their comparison could beof interest Even if the seasonal timing of the glacier massloss contribution is unclear it could in general peak at asimilar time of the year than the seasonal glacier contribu-tion so that both components rather enlarge each other thancancel Seasonal glacier contributions have been modeled byKaser et al (2010) The latter study assumed glaciers to bein equilibrium with the present climate and thus prescribedglacier mass balances equal to 0 Thus their seasonally de-layed glacier contributions do not include the part due todecadal glacier mass loss and is complementary to our es-timates (Table 6) For the eastern basins under the influenceof the Indian summer monsoon (Ganges and Brahmaputra)the contribution of glacier mass loss is higher than the sea-sonal contribution In other words for the first decade of the21st century the mass loss of glaciers is on average moreimportant for water availability in those two rivers than theirseasonal water storage effect The situation is different for theIndus Basin where the glacier melt season is also the dry sea-son which results in a relatively higher seasonally delayedglacier contribution to the river discharge There the season-ally delayed water release from the glaciers and the decadalglacier mass loss contribute on average equally to the riverdischarge Both contributions are strongly (and equally) de-pendent on the location of the discharge measurement andwill be significantly larger (in relative term) in more heav-ily glacierized and smaller mountain catchments located inthe upper part of these major river basins (Bookhagen andBurbank 2010)

6 Conclusion

In this study we assessed the spatial pattern of glaciermass balance over the PKH by comparing the SRTM andSPOT5 DEMs over 9 well-distributed study sites between1999 and 2011 We found slightly positive or balanced massbudgets in the western part (Pamir Karakoram) moder-ate mass loss in the east (Nepal Bhutan Hengduan Shan)

and the most negative mass balance in the monsoon-aridtransition zone (Spiti Lahaul in the western Himalaya)By extrapolating these values to climatically homogeneoussub-regions we estimate the PKH overall mass balanceto have beenminus014plusmn 008 m we yrminus1 between 1999 and2011 which corresponds to a contribution to sea level riseof 0028plusmn 0015 mm yrminus1 This is about 4 of the globalglacier mass loss estimated by Gardner et al (2013) thoughover a slightly shorter time period (2003ndash2009) The largestsource of uncertainties in those geodetic mass balance esti-mates is the poorly constrained penetration of the C-BandSRTM radar signal into snow and ice

Our technique of DEM differencing allows mapping in de-tail the spatial pattern of glacier elevation changes Thosemaps reveal many surge-type behaviors both in the Pamirand the Karakoram It also appears that the glacier-wide massbalance does not seem to be considerably affected by themass transfer from surges over thesim 10 yr time period stud-ied Similar lowering rates over debris-covered and clean iceare observed for four study sites despite the commonly as-sumed insulating effect of debris covers This suggests forthe scale of entire glacier tongues a spatial mixture of re-duced ablation under intact thick debris covers and enhancedablation by supraglacial lakes or ice cliffs or due to thin de-bris layer and very low glacier velocities in ablation areasthat reduce the ice advection downstream

The regionally heterogeneous pattern of ice wastage is inbroad agreement with contrasted trends in large-scale pre-cipitation and temperature However glaciological (eg ac-cumulation) and meteorological records are needed at highaltitude in the vicinity of glaciers to evaluate more closelythe local influence of atmospheric variables on glacier massbalance and fully understand the reasons behind those con-trasted mass budgets in the PKH

Supplementary material related to this article isavailable online at httpwwwthe-cryospherenet712632013tc-7-1263-2013-supplementpdf

Acknowledgements We thank G Cogley A Gardner T NuimuraT Bolch P Wagnon M Barandun M Hoelzle M Huss and ananonymous referee for their constructive comments that led to aconsiderably improved manuscript We thank the Water and PowerDevelopment Authority (WAPDA) for their hydrological data(processed by P Chevallier) and F Papa for sharing his updatedrun-off data ahead of publication We thank T Bolch for sharinghis glacier outlines from the Everest area We thank the USGSfor allowing free access to their Landsat archive CGIAR-SCI forSRTM C-band data DLR for SRTM X-band data and GLIMS forglacier inventories J Gardelle acknowledges a PhD fellowshipfrom the French Space Agency (CNES) and the French NationalResearch Center (CNRS) E Berthier acknowledges support fromCNES through the TOSCA and ISIS proposal no 397 and fromthe Programme National de Teacuteleacutedeacutetection Spatiale E Berthieracknowledges M Masiokas and R Villalba for welcoming him

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1283

at the IANIGLA (Mendoza Argentina) Y Arnaud acknowledgessupport from French National Research Agency through ANR-09-CEP-005-01PAPRIKA A Kaumlaumlb acknowledges support fromESA through Glaciers_cci (400010177810I-AM) and from theEuropean Research Council under the European Unionrsquos SeventhFramework Programme (FP2007-2013)ERC Grant Agreementn 320816

Edited by I M Howat

The publication of this articleis financed by CNRS-INSU

References

Adler R Huffman G Chang A Ferraro R Xie P JanowiakJ Rudolf B Schneider U Curtis S Bolvin D Gruber ASusskind J Arkin P and Nelkin E The Version-2 GlobalPrecipitation Climatology Project (GPCP) Monthly Precipita-tion Analysis (1979ndashPresent) J Hydrometeorol 4 1147ndash11672003

Archer D R and Fowler H J Spatial and temporal variationsin precipitation in the Upper Indus Basin global teleconnectionsand hydrological implications Hydrol Earth Syst Sci 8 47ndash61doi105194hess-8-47-2004 2004

Arendt A Bolch T Cogley J G Gardner A Hagen J-OHock R Kaser G Pfeffer W T Moholdt G Paul F RadicV Andreassen L Bajracharya S Beedle M Berthier EBhambri R Bliss A Brown I Burgess E Burgess DCawkwell F Chinn T Copland L Davies B de AngelisH Dolgova E Filbert K Forester R Fountain A Frey HGiffen B Glasser N Gurney S Hagg W Hall D Hari-tashya U K Hartmann G Helm C Herreid S Howat IKapustin G Khromova T Kienholz C Koenig M KohlerJ Kriegel D Kutuzov S Lavrentiev I LeBris R Lund JManley W Mayer C Miles E Li X Menounos B Mer-cer A Moelg N Mool P Nosenko G Negrete A NuthC Pettersson R Racoviteanu A Ranzi R Rastner P RauF Rich J Rott H Schneider C Seliverstov Y Sharp MSigurethsson O Stokes C Wheate R Winsvold S WolkenG Wyatt F and Zheltyhina N Randolph Glacier Inventory[v20] A Dataset of Global Glacier Outlines Global Land IceMeasurements from Space Boulder Colorado USA Digital Me-dia httpwwwglimsorgRGIrandolphhtml 2012

Azam M Wagnon P Ramanathan A Vincent C Sharma PArnaud Y Linda A Pottakkal J Chevallier P Singh V andBerthier E From balance to imbalance a shift in the dynamicbehaviour of Chhota Shigri glacier western Himalaya India JGlaciol 58 315ndash324 doi1031892012JoG11J123 2012

Azam M F Wagnon P Vincent C Ramanathan A Linda Aand Singh V B Reconstruction of the annual mass balanceof Chhota Shigri Glacier (Western Himalaya India) since 1969Ann Glaciol in revision 2013

Barrand N and Murray T Multivariate Controls on the In-cidence of Glacier Surging in the Karakoram Himalaya

Arct Antarct Alp Res 38 489ndash498 doi1016571523-0430(2006)38[489MCOTIO]20CO2 2006

Barundun M Huss M Azisov E Gafurov A HoelzleM Merkushkin A Salzmann N and Usubaliev R Re-establishing seasonal mass balance observation at AbramovGlacier Kyrgyzstan from 1968ndash2012 Abstract EGU2013-5668European Geophysical Union Vienna 2013

Benn D Bolch T Hands K Gulley J Luckman ANicholson L Quincey D Thompson S Toumi R andWiseman S Response of debris-covered glaciers in theMount Everest region to recent warming and implicationsfor outburst flood hazards Earth-Sci Rev 114 156ndash174doi101016jearscirev201203008 2012

Berthier E and Vincent C Relative contribution of surface mass-balance and ice-flux changes to the accelerated thinning of Merde Glace French Alps over 1979ndash2008 J Glaciol 58 501ndash512doi1031892012JoG11J083 2012

Berthier E Arnaud Y Baratoux D Vincent C and Reacutemy FRecent rapid thinning of the ldquo Mer de Glace rdquo glacier derivedfrom satellite optical images Geophys Res Lett 31 L17401doi1010292004GL020706 2004

Berthier E Arnaud Y Kumar R Ahmad S Wagnon P andChevallier P Remote sensing estimates of glacier mass balancesin the Himachal Pradesh (Western Himalaya India) RemoteSens Environ 108 327ndash338 doi101016jrse2006110172007

Bhambri R Bolch T Kawishwar P Dobhal D P SrivastavaD and Pratap B Heterogeneity in Glacier response from 1973to 2011 in the Shyok valley Karakoram India The CryosphereDiscuss 6 3049ndash3078 doi105194tcd-6-3049-2012 2012

Bhutiyani M Mass-balance studies on Siachen Glacier in theNubra valley Karakoram Himalaya India J Glaciol 45 112ndash118 1999

Bishop M P Schroder J F and Ward J L SPOT multispec-tral analysis for producing supraglacial debris-load estimates forBatura glacier Pakistan Geocarto Int 10 81ndash90 1995

Bolch T Pieczonka T and Benn D I Multi-decadal mass lossof glaciers in the Everest area (Nepal Himalaya) derived fromstereo imagery The Cryosphere 5 349ndash358 doi105194tc-5-349-2011 2011

Bolch T Kulkarni A Kaumlaumlb A Huggel C Paul F Cogley JFrey H Kargel J Fujita K Scheel M Bajracharya S andStoffel M The state and fate of Himalayan Glaciers Science336 310ndash314 doi101126science1215828 2012

Bollasina M Ming Y and Ramaswamy V Anthropogenicaerosols and the weakening of the South Asian summer mon-soon Science 334 502ndash505 doi101126science12049942011

Bookhagen B and Burbank D Toward a complete Himalayan hy-drological budget spatiotemporal distribution of snowmelt andrainfall and their impact on river discharge J Geophys Res115 F03019 doi1010292009JF001426 2010

Bretherton C Widmann M Dymnikov V Wallace J and BladeacuteI The effective number of spatial degrees of freedom of a time-varying field J Climate 12 1990ndash2009 1999

Copland L Pope S Bishop M Shroder J Clendon PBush A Kamp U Seong Y and Owen L Glacier veloc-ities across the central Karakoram Ann Glaciol 50 41ndash49doi103189172756409789624229 2009

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1283

at the IANIGLA (Mendoza Argentina) Y Arnaud acknowledgessupport from French National Research Agency through ANR-09-CEP-005-01PAPRIKA A Kaumlaumlb acknowledges support fromESA through Glaciers_cci (400010177810I-AM) and from theEuropean Research Council under the European Unionrsquos SeventhFramework Programme (FP2007-2013)ERC Grant Agreementn 320816

Edited by I M Howat

The publication of this articleis financed by CNRS-INSU

References

Adler R Huffman G Chang A Ferraro R Xie P JanowiakJ Rudolf B Schneider U Curtis S Bolvin D Gruber ASusskind J Arkin P and Nelkin E The Version-2 GlobalPrecipitation Climatology Project (GPCP) Monthly Precipita-tion Analysis (1979ndashPresent) J Hydrometeorol 4 1147ndash11672003

Archer D R and Fowler H J Spatial and temporal variationsin precipitation in the Upper Indus Basin global teleconnectionsand hydrological implications Hydrol Earth Syst Sci 8 47ndash61doi105194hess-8-47-2004 2004

Arendt A Bolch T Cogley J G Gardner A Hagen J-OHock R Kaser G Pfeffer W T Moholdt G Paul F RadicV Andreassen L Bajracharya S Beedle M Berthier EBhambri R Bliss A Brown I Burgess E Burgess DCawkwell F Chinn T Copland L Davies B de AngelisH Dolgova E Filbert K Forester R Fountain A Frey HGiffen B Glasser N Gurney S Hagg W Hall D Hari-tashya U K Hartmann G Helm C Herreid S Howat IKapustin G Khromova T Kienholz C Koenig M KohlerJ Kriegel D Kutuzov S Lavrentiev I LeBris R Lund JManley W Mayer C Miles E Li X Menounos B Mer-cer A Moelg N Mool P Nosenko G Negrete A NuthC Pettersson R Racoviteanu A Ranzi R Rastner P RauF Rich J Rott H Schneider C Seliverstov Y Sharp MSigurethsson O Stokes C Wheate R Winsvold S WolkenG Wyatt F and Zheltyhina N Randolph Glacier Inventory[v20] A Dataset of Global Glacier Outlines Global Land IceMeasurements from Space Boulder Colorado USA Digital Me-dia httpwwwglimsorgRGIrandolphhtml 2012

Azam M Wagnon P Ramanathan A Vincent C Sharma PArnaud Y Linda A Pottakkal J Chevallier P Singh V andBerthier E From balance to imbalance a shift in the dynamicbehaviour of Chhota Shigri glacier western Himalaya India JGlaciol 58 315ndash324 doi1031892012JoG11J123 2012

Azam M F Wagnon P Vincent C Ramanathan A Linda Aand Singh V B Reconstruction of the annual mass balanceof Chhota Shigri Glacier (Western Himalaya India) since 1969Ann Glaciol in revision 2013

Barrand N and Murray T Multivariate Controls on the In-cidence of Glacier Surging in the Karakoram Himalaya

Arct Antarct Alp Res 38 489ndash498 doi1016571523-0430(2006)38[489MCOTIO]20CO2 2006

Barundun M Huss M Azisov E Gafurov A HoelzleM Merkushkin A Salzmann N and Usubaliev R Re-establishing seasonal mass balance observation at AbramovGlacier Kyrgyzstan from 1968ndash2012 Abstract EGU2013-5668European Geophysical Union Vienna 2013

Benn D Bolch T Hands K Gulley J Luckman ANicholson L Quincey D Thompson S Toumi R andWiseman S Response of debris-covered glaciers in theMount Everest region to recent warming and implicationsfor outburst flood hazards Earth-Sci Rev 114 156ndash174doi101016jearscirev201203008 2012

Berthier E and Vincent C Relative contribution of surface mass-balance and ice-flux changes to the accelerated thinning of Merde Glace French Alps over 1979ndash2008 J Glaciol 58 501ndash512doi1031892012JoG11J083 2012

Berthier E Arnaud Y Baratoux D Vincent C and Reacutemy FRecent rapid thinning of the ldquo Mer de Glace rdquo glacier derivedfrom satellite optical images Geophys Res Lett 31 L17401doi1010292004GL020706 2004

Berthier E Arnaud Y Kumar R Ahmad S Wagnon P andChevallier P Remote sensing estimates of glacier mass balancesin the Himachal Pradesh (Western Himalaya India) RemoteSens Environ 108 327ndash338 doi101016jrse2006110172007

Bhambri R Bolch T Kawishwar P Dobhal D P SrivastavaD and Pratap B Heterogeneity in Glacier response from 1973to 2011 in the Shyok valley Karakoram India The CryosphereDiscuss 6 3049ndash3078 doi105194tcd-6-3049-2012 2012

Bhutiyani M Mass-balance studies on Siachen Glacier in theNubra valley Karakoram Himalaya India J Glaciol 45 112ndash118 1999

Bishop M P Schroder J F and Ward J L SPOT multispec-tral analysis for producing supraglacial debris-load estimates forBatura glacier Pakistan Geocarto Int 10 81ndash90 1995

Bolch T Pieczonka T and Benn D I Multi-decadal mass lossof glaciers in the Everest area (Nepal Himalaya) derived fromstereo imagery The Cryosphere 5 349ndash358 doi105194tc-5-349-2011 2011

Bolch T Kulkarni A Kaumlaumlb A Huggel C Paul F Cogley JFrey H Kargel J Fujita K Scheel M Bajracharya S andStoffel M The state and fate of Himalayan Glaciers Science336 310ndash314 doi101126science1215828 2012

Bollasina M Ming Y and Ramaswamy V Anthropogenicaerosols and the weakening of the South Asian summer mon-soon Science 334 502ndash505 doi101126science12049942011

Bookhagen B and Burbank D Toward a complete Himalayan hy-drological budget spatiotemporal distribution of snowmelt andrainfall and their impact on river discharge J Geophys Res115 F03019 doi1010292009JF001426 2010

Bretherton C Widmann M Dymnikov V Wallace J and BladeacuteI The effective number of spatial degrees of freedom of a time-varying field J Climate 12 1990ndash2009 1999

Copland L Pope S Bishop M Shroder J Clendon PBush A Kamp U Seong Y and Owen L Glacier veloc-ities across the central Karakoram Ann Glaciol 50 41ndash49doi103189172756409789624229 2009

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

1284 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Copland L Sylvestre T Bishop M Schroder J Seong YOwen L Bush A and Kamp U Expanded and recently in-creased glacier surging in the Karakoram Arct Antarct AlpRes 43 503ndash516 doi1016571938-4246-434503 2011

Dobhal D Gergan J and Thayyen R Mass balance studies ofthe Dokirani Glacier from 1992 to 2000 Garhwal Himalaya In-dia Bull Glaciol Res 25 9ndash17 2008

Fowler H and Archer D Conflicting signals of climatic change inthe upper Indus basin J Climate 19 4276ndash4293 2006

Frey H Paul F and Strozzi T Compilation of a glacier in-ventory for the western Himalayas from satellite data methodschallenges and results Remote Sens Environ 124 832ndash843doi101016jrse201206020 2012

Fujita K Effect of precipitation seasonality on climatic sensitiv-ity of glacier mass balance Earth Planet Sc Lett 276 14ndash19doi101016jepsl200808028 2008

Fujita K and Nuimura T Spatially heterogeneous wastage of Hi-malayan glaciers P Natl Acad Sci USA 108 14011ndash14014doi101073pnas1106242108 2011

Fujita K Sakai A Nuimura T Yamaguchi S and SharmaR R Recent changes in Imja Glacial Lake and its dammingmoraine in the Nepal Himalaya revealed by in situ surveys andmulti-temporal ASTER imagery Environ Res Lett 4 1ndash7doi1010881748-932644045205 2009

Gardelle J Arnaud Y and Berthier E Contrasted evolution ofglacial lakes along the Hindu Kush Himalaya mountain rangebetween 1990 and 2009 Global Planet Change 75 47ndash55doi101016jgloplacha201010003 2011

Gardelle J Berthier E and Arnaud Y Slight mass gainof Karakoram glaciers in the early twenty-first century NatGeosci 5 322ndash325 doi101038NGEO1450 2012a

Gardelle J Berthier E and Arnaud Y Impact of resolu-tion and radar penetration on glacier elevation changes com-puted from DEM differencing J Glaciol 58 419ndash422doi1031892012JoG11J175 2012b

Gardner A Moholdt G Arendt A and Wouters B Acceleratedcontributions of Canadarsquos Baffin and Bylot Island glaciers to sealevel rise over the past half century The Cryosphere 6 1103ndash1125 doi105194tc-6-1103-2012 2012

Gardner A S Moholdt G Cogley J G Wouters B ArendtA A Wahr J Berthier E Hock R Pfeffer W T KaserG Ligtenberg S R M Bolch T Sharp M J Hagen J Ovan den Broeke M R and Paul F A Reconciled Estimate ofGlacier Contributions to Sea Level Rise 2003 to 2009 Science340 852ndash857 doi101126science1234532 2013

Heid T and Kaumlaumlb A Repeat optical satellite images revealwidespread and long term decrease in land-terminating glacierspeeds The Cryosphere 6 467ndash478 doi105194tc-6-467-2012 2012

Hewitt K The Karakoram anomaly Glacier expansion and theelevation effect Karakoram Himalaya Mt Res Dev 25 332ndash340 2005

Hewitt K Tributary glaciers surges an exceptional concentrationat Panmah Glacier Karakoram Himalaya J Glaciol 53 181ndash188 2007

Huss M Density assumptions for converting geodetic glaciervolume change to mass change The Cryosphere 7 877ndash887doi105194tc-7-877-2013 2013

Immerzeel W VanBeek L P H and Bierkens M F P ClimateChange Will Affect the Asian Water Towers Science 328 1382ndash1385 doi101126science1183188 2010

Immerzeel W Pellicciotti F and Shrestha A Glaciers as a proxyto quantify the spatial distribution of precipitation in the Hunzabasin Mt Res Dev 32 30ndash38 doi101659MRD-JOURNAL-D-11-000971 2012

Ives J Shrestha R and Mool P Formation of Glacial Lakes inthe Hindu Kush-Himalayas and GLOF Risk Assessment Inter-national Center for Integrated Mountain Development 2010

Jacob T Wahr J Pfeffer W and Swenson S Recent contri-butions of glaciers and ice caps to sea level rise Nature 482514ndash518 doi101038nature10847 2012

Kaumlaumlb A Combination of SRTM3 and repeat ASTERdata for deriving alpine glacier flow velocities in theBhutan Himalaya Remote Sens Environ 94 463ndash474doi101016jrse200411003 2005

Kaumlaumlb A Berthier E Nuth C Gardelle J and Arnaud Y Con-trasting patterns of early 21st century glacier mass change inthe Himalayas Nature 488 495ndash498 doi101038nature113242012

Kaser G Grosshauser M and Marzeion B Contribu-tion of glaciers to water availability in different climateregimes P Natl Acad Sci USA 107 20223ndash20227doi101073pnas1008162107 2010

Korona J Berthier E MBernard Reacutemy F and ThouvenotE SPIRIT SPOT 5 stereoscopic survey of Polar Ice Refer-ence Images and Topographies during the fourth InterationalPolar Year (2007ndash2009) ISPRS J Photogramm 64 204ndash212doi101016jisprsjprs200810005 2009

Kotlyakov V Osipova G and Tsvetkov D Monitoring surgingglaciers of the Pamirs central Asia from space Ann Glaciol48 125ndash134 doi103189172756408784700608 2008

Kulkarni A Rathore B Singh S and Bahuguna I Un-derstanding changes in the Himalayan cryosphere using re-mote sensing techniques Int J Remote Sens 32 601ndash615doi101080014311612010517802 2011

Lejeune Y Bertrand J-M Wagnon P and Morin S A phys-ically based model of the year-round surface energy and massbalance of debris-covered glaciers J Glaciol 59 327ndash344doi1031892013JoG12J149 2013

Marzeion B Jarosch A H and Hofer M Past and future sea-level change from the surface mass balance of glaciers TheCryosphere 6 1295ndash1322 doi105194tc-6-1295-2012 2012

Matsuo K and Heki K Time-variable ice loss in Asian highmountains from satellite gravimetry Earth Planet Sc Lett 29030ndash36 2010

Mattson L Gardner J and Young G Ablation on debris cov-ered glaciers an example from the Rakhiot Glacier Punjab Hi-malaya Proceedings of the Kathmandu Symposium November1992 IAHS Publ no 218 1993

Mihalcea C Mayer C Diolaiuti G Lambrecht A SmiragliaC and Tartari G Ice ablation and meteorological conditions onthe debris-covered area of Baltoro glacier Karakoram PakistanAnn Glaciol 43 292ndash300 doi1031891727564067818121042006

Mihalcea C Mayer C Diolaiuti G DrsquoAgata C Smi-raglia C Lambrecht A Vuillermoz E and Tartari GSpatial distribution of debris thickness and melting from

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya 1285

remote-sensing and meteorological data at debris-covered Bal-toro glacier Karakoram Pakistan Ann Glaciol 48 49ndash57doi103189172756408784700680 2008

Nuimura T Fujita K Yamaguchi S and Sharma R Eleva-tion changes of glaciers revealed by multitemporal digital el-evation models calibrated by GPS survey in the Khumbu re-gion Nepal Himalaya 1992ndash2008 J Glaciol 58 648ndash656doi1031892012JoG11J061 2012

Nuth C and Kaumlaumlb A Co-registration and bias corrections of satel-lite elevation data sets for quantifying glacier thickness changeThe Cryosphere 5 271ndash290 doi105194tc-5-271-2011 2011

Nuth C Schuler T Kohler J Altena B and Hagen J Es-timating the long-term calving flux of Kronebreen Svalbardfrom geodetic elevation changes and mass-balance modeling JGlaciol 58 119ndash135 doi1031892012JoG11J036 2012

Ohmura A Observed Mass Balance of Mountain Glaciers andGreenland Ice Sheet in the 20th Century and the Present TrendsSurv Geophys 32 537ndash554 doi101007s10712-011-9124-42011

Oerlemans J Glaciers and climate change A A Balkema Pub-lishers Rotterdam 2001

Papa F Bala S K Pandey R K Durand F Gopalakrishna VV Rahman A and Rossow W B Ganga-Brahmaputra riverdischarge from Jason-2 radar altimetry An update to the long-term satellite-derived estimates of continental freshwater forcingflux into the Bay of Bengal J Geophys Res 117 C11 C11021doi1010292012JC008158 2012

Paul F Calculation of glacier elevation changes with SRTM isthere an elevation-dependent bias J Glaciol 54 945ndash946doi103189002214308787779960 2008

Paul F Barrand N E Berthier E Bolch T Casey K FreyH Joshi S P Konovalov V Le Bris R Moelg N Nuth CPope A Racoviteanu A Rastner P Raup B Scharrer KSteffen S and Winswold S On the accuracy of glacier outlinesderived from remote sensing data Ann Glaciol 54 171ndash182doi1031892013AoG63A296 2013

Pieczonka T Bolch T Junfeng W and Shiyin L Heteroge-neous mass loss of glaciers in the Aksu-Tarim Catchment (Cen-tral Tien Shan) revealed by 1976 KH-9 Hexagon and 2009SPOT-5 stereo imagery Remote Sens Environ 130 233ndash244doi101016jrse201211020 2013

Quincey D Richardson S Luckman A Lucas RReynolds J Hambrey M and Glasser N Early recog-nition of glacial lake hazards in the Himalaya using re-mote sensing datasets Global Planet Change 56 137ndash152doi101016jgloplacha200607013 2007

Quincey D Copland L Mayer C Bishop M Luck-mann A and Belograve M Ice velocity and climate varia-tions for Baltoro Glacier Pakistan J Glaciol 55 1061ndash1071doi103189002214309790794913 2009

Quincey D Braun M Glasser N Bishop M Hewitt K andLuckman A Karakoram glacier surge dynamics Geophys ResLett 38 L18504 doi1010292011GL049004 2011

Rabatel A Dedieu J and Vincent C Using remote-sensing datato determine equilibrium-line altitude and mass-balance time se-ries validation on three French glaciers 1994-2002 J Glaciol51 539ndash546 doi103189172756505781829106 2005

Rabus B Eineder M Roth A and Bamler R The shuttle radartopography mission-a new class of digital elevation models ac-

quired by spaceborne radar ISPRS J Photogramm 57 241ndash262 doi101016S0924-2716(02)00124-7 2003

Racoviteanu A Paul F Raup B Khalsa S and Armstrong RChallenges and recommendations in mapping of glacier param-eters from space results of the 2008 Global Land Ice Measure-ments from Space (GLIMS) workshop Boulder Colorado USAAnn Glaciol 50 53ndash69 doi1031891727564107905958042009

Radic V and Hock R Regionally differentiated contribution ofmountain glaciers and ice caps to future sea-level rise NatGeosci 4 91ndash94 2011

Rasul G Climate Data Modelling and Analysis of the In-dus Ecoregion World Wide Fund for Nature ndash Pak-istan httpwwwindiaenvironmentportalorginfilesfileccap_climatechangepdf (last access 2 July 2013) 2012

Raup B Racoviteanu A Khalsa S Helm C Armstrong R andArnaud Y The GLIMS geospatial glacier database a new toolfor studying glacier change Global Planet Change 56 101ndash110doi101016jgloplacha200607018 2007

Rignot E Echelmeyer K and Krabill W Penetration depth ofinterferometric synthetic-aperture radar signals in snow and iceGeophys Res Lett 28 3501ndash3504 2001

Sakai A Takeuchi N Fujita K and Nakawo M Role ofsupraglacial ponds in the ablation process of a debris-coveredglacier in the Nepal Himalayas in Debris-covered glaciersedited by Nawako M Raymond C and Fountain A 119ndash130 2000

Sakai A Nakawo M and Fujita K Distribution charac-teristics and energy balance of ice cliffs on debris-coveredglaciers Nepal Himalaya Arct Antarct Alp Res 34 12ndash19doi1023071552503 2002

Sapiano J J Harrison W D and Echelmeyer K A Elevationvolume and terminus changes of nine glaciers in North AmericaJ Glaciol 44 119ndash135 1998

Scherler D Bookhagen B and Strecker M Spatially variableresponse of Himalayan glaciers to climate change affected bydebris cover Nat Geosci 4 156ndash159 doi101038ngeo10682011a

Scherler D Bookhagen B and Strecker M Hillslope-glacier coupling The interplay of topography and glacialdynamics in High Asia J Geophys Res 116 F02019doi1010292010JF001751 2011b

Shekhar M Chand H Kumar S Srinivasan K and Ganju AClimate-change studies in the western Himalaya Ann Glaciol51 105ndash112 doi103189172756410791386508 2010

Shi Y Liu C and Kang E The Glacier Inventory of China AnnGlaciol 50 1ndash4 doi103189172756410790595831 2009

Shrestha A Wake C Mayewski P and Dibb J Maximumtemperature trends in the Himalaya and its vicinity an anal-ysis based on temperature records from Nepal for the pe-riod 1971ndash94 J Climate 12 2775ndash2786 doi1011751520-0442(1999)012lt2775MTTITHgt20CO2 1999

Span N and Kuhn M Simulating annual glacier flow witha linear reservoir model J Geophys Res 108 4313doi1010292002JD002828 2003

Ulaby F T Moore R K and Fung A K Microwave remotesensing active and passive Vol 3 From theory to applicationsArtech House 1986

wwwthe-cryospherenet712632013 The Cryosphere 7 1263ndash1286 2013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013

1286 J Gardelle et al Region-wide glacier mass balances over the Pamir-Karakoram-Himalaya

Vincent C Influence of climate change over the 20th Century onfour French glacier mass balances J Geophys Res 107 4375doi1010292001JD000832 2002

Vincent C Ramanathan Al Wagnon P Dobhal D P Linda ABerthier E Sharma P Arnaud Y Azam M F Jose P G andGardelle J Balanced conditions or slight mass gain of glaciersin the Lahaul and Spiti region (northern India Himalaya) duringthe nineties preceded recent mass loss The Cryosphere 7 569ndash582 doi105194tc-7-569-2013 2013

Wagnon P Linda A Arnaud Y Kumar R Sharma P Vin-cent C Pottakkal J Berthier E Ramanathan A Has-nain S and Chevallier P Four years of mass balance onChhota Shigri Glacier Himachal Pradesh India a new bench-mark glacier in the western Himalaya J Glaciol 53 603ndash611doi103189002214307784409306 2007

Wagnon P Vincent C Arnaud Y Berthier E Vuillermoz EGruber S Meacuteneacutegoz M Gilbert A Dumont M Shea JM Stumm D and Pokhrel B K Seasonal and annual massbalances of Mera and Pokalde glaciers (Nepal Himalaya) since2007 The Cryosphere Discuss 7 3337ndash3378 doi105194tcd-7-3337-2013 2013

Wake C Glaciochemical investigations as a tool for determiningthe spatial and seasonal variation of snow accumulation in thecentral Karakoram northern Pakistan Ann Glaciol 13 279ndash284 1989

WGMS Fluctuations of Glaciers 2005ndash2010 (Vol X) editedby Zemp M Frey H Gaumlrtner-Roer I Nussbaumer S UHoelzle M Paul F and Haeberli W ICSU (WDS)IUGG(IACS)UNEP UNESCOWMO World Glacier MonitoringService Zurich Switzerland Based on database versiondoi105904wgms-fog-2012-11 2012

Yao T Thompson L Yang W Yu W Gao Y Guo X YangX Duan K Zhao H Xu B Pu J Lu A Xiang Y KattelD B and Joswiak D Different glacier status with atmosphericcirculations in Tibetan Plateau and surroundings Nature ClimChange 2 663ndash667 doi101038nclimate1580 2012

Yde J and Paasche Oslash Reconstructing climate change notall glaciers suitable EOS Transactions American GeophysicalUnion 91 189ndash190 doi1010292010EO210001 2010

Zhang G Yao T Xie H Kang S and Lei Y Increased massover the Tibetan Plateau From lakes or glaciers Geophys ResLett 40 2125ndash2130 doi101002grl50462 2013a

Zhang Y Hirabayashi Y Fujita K Liu S and Liu QSpatial debris-cover effect on the maritime glaciers of MountGongga south-eastern Tibetan Plateau The Cryosphere Dis-cuss 7 2413ndash2453 doi105194tcd-7-2413-2013 2013b

Zwally H Schutz B Abdalati W Abshire J Bentley C Bren-ner A Bufton J Dezio J Hancock D Harding D HerringT Minster B Quinn K Palm S Spinhirne J and ThomasR ICESatrsquos laser measurements of polar ice atmosphereocean and land J Geodyn 34 405ndash445 doi101016S0264-3707(02)00042-X 2002

The Cryosphere 7 1263ndash1286 2013 wwwthe-cryospherenet712632013