Embed Size (px)
Citation preview
Andean HighlandsImplications of climate change
Anji Seth, Jeanne Thibeault (UConn)Magali Garcia (UMSA, La Paz)
SANREM CRSPSustainable Agriculture and Natural Resource Management
Collaborative Research Support Program
• promotes stakeholder empowerment and improved livelihoods through knowledge - based sustainable agriculture and natural resource management systems.
• emphasis on developing economies
Adapting to Change in the Andean Highlands: Practices and Strategies to Address Climate and Market Risks in Vulnerable Agro-Eco Systems
Andean Highlands
Altiplano ClimateFraction of annual precipitation: DJF, JJA
Dec - Feb
84˚W 80˚W 76˚W 72˚W 68˚W 64˚W 60˚W 56˚W 52˚W 48˚W 44˚W 40˚W 36˚WLongitude
40˚S
36˚S
32˚S
28˚S
24˚S
20˚S
16˚S
12˚S
8˚S
4˚S
0˚4˚
N8˚
NLa
titud
e
Jun-Aug
84˚W 80˚W 76˚W 72˚W 68˚W 64˚W 60˚W 56˚W 52˚W 48˚W 44˚W 40˚W 36˚WLongitude
40˚S
36˚S
32˚S
28˚S
24˚S
20˚S
16˚S
12˚S
8˚S
4˚S
0˚4˚N
8˚N
Latitude
0 10 20 30 40 50 60 70 80 90 100precipitation
Overview
• SANREM CRSP Background
• Altiplano climate - mean
• Altiplano climate - variability
• Altiplano climate - projections
• Discussion
Altiplano ClimateFeatures of South American Warm Season Rains
!"#$"%&'()$*&+,-&+."/$*01&2(13((1&
! "
!
!
!
!
#$%&'(! )*! +,-(./0$,! $11&20'/0$34! 35! 0-(! 63'0-! /47! +3&0-!8.('$,/4!.3423342! 2920(.2! :1(50! /47! '$%-0! ;/4(12<!
'(2;(,0$=(19>*!+-/7$4%!$47$,/0(2!;'(,$;$0/0$34!/47!7/2-(7!1$4(2!$47$,/0(!,34=('%(4,(!?34(2*!+./11!/''3@2!2-3@!
13@A1(=(1!:BCC!-D/>!@$472<!/47!0-$,E!/''3@2!'(;'(2(40!13@A1(=(1!F(02*!84!GHG!2-3@2!/!2&I0'3;$,/1!2&'5/,(!-$%-!
,(40('<!/47!/4!G8G!$47$,/0(2!0-(!.342334!/40$,9,134(*!:87/;0(7!5'3.!J*!K3&2E9*>!
!
!
!
#$%&'(! "*! +,-(./0$,! =('0$,/1! 2(,0$34! 53'! 0-(! ,3''(2;347$4%! 2&..('! 2(/234! /0! /'3&47! "CL6! :1(50! ;/4(1>! /47!
23&0-@(20A43'0-(/20! :'$%-0! ;/4(1>*! M(%$342! 35! 7((;! ,34=(,0$34! /47! 13@A1(=(1! F(02! /'(! $47$,/0(7*! :D/4(1! 53'!
68N+!/7/;0(7!5'3.!O*!H$%%$42*>!
!
#$%&'(2!)!/47!"!2-3@!2,-(./0$,/119!0-(!./F3'!5(/0&'(2!35!0-(!@/'.!2(/234!,$',&1/0$34!3=('!63'0-!
/47!+3&0-!8.('$,/*!P-(!68N+!$2!,-/'/,0('$?(7!I9!/!'(%$34!35!$40(42(!;'(,$;$0/0$34!(./4/0$4%!5'3.!
0-(!(/20('4!D/,$5$,! $40('0'3;$,/1! ,34=('%(4,(!?34(! :QPRS><! (T0(47$4%!43'0-@/'7!3=('!N(T$,3! 03! 0-(!
23&0-@(20!U4$0(7!+0/0(2! :U*+*><!@$0-! 1/'%(20! =/1&(2! 3=('! 0-(!@(20('4! 213;(2! 35! 0-(!.3&40/$4! '/4%(*!
H$%-!=/1&(2!35!;'(,$;$0/0$34!/123!(T0(47!43'0-(/20@/'7!3=('!0-(!V&15!35!N(T$,3<!'(/,-$4%!&;!/134%!0-(!
(/20('4! 51/4E! 35! 63'0-! 8.('$,/<! /47!.('%$4%! $403! 0-(! 63'0-! 801/40$,! 203'.! 0'/,E*! P-('(! $2! /123! /!
'(1/0$=(!./T$.&.!03!0-(!23&0-@(20!35!0-(!V'(/0!W/E(2*!P-(!,340$4(40/1!(/20A@(20!,340'/20!I(0@((4!0-(!
/'$7!@(20!/47! 0-(!-&.$7!(/20! $2!/!E(9!,-/'/,0('$20$,!35!/!.342334/1!,$',&1/0$34<!@-$,-!@(!(1/I3'/0(!
&;34!$4!2(,0$34!X*!P-(!&;;('A1(=(1!.342334!/40$,9,134(!/223,$/0(7!@$0-!0-(!68N+!2-$502!43'0-@/'7!
@$0-! 2(/234! 5'3.! 23&0-@(20('4! N(T$,3! 03! 43'0-@(20('4! N(T$,3! /47! 23&0-@(20('4! U4$0(7! +0/0(2*!
N3$20&'(!0'/42;3'0!3403!0-(!63'0-!8.('$,/4!,340$4(40!$2!/223,$/0(7!@$0-!I'3/7A2,/1(!/7=(,0$34!5'3.!
0-(!V&15!35!N(T$,3<!/47!@$0-!$.;3'0/40!13@A1(=(1!F(02!:WWY2>!3=('!0-(!V&15!35!R/1$53'4$/!/47!(/20!35!
0-(!M3,E$(2*!P-(!1/00('!WWY!$2!/!@/'.!2(/234<!;'$./'$19!43,0&'4/1<!5(/0&'(Z!1(22!7(0/$1!$2!E43@4!/I3&0!
0-(!30-('!34(*!
!"#$"%&'()$*&+,-&+."/$*01&2(13((1&
! "#
$%&'!()(*+!*,)-&.*'!-)/0+!1*)23+*'!01!2%*!45647!4*82&0.!"!(+*'*.2'!89&-)2090:&8)9!)'(*82';!'38%!
)'! *<0932&0.;! %*)2&.:! =&'2+&>32&0.! ).=! 2%*! )''08&)2*=! 8&+839)2&0.7! ?)+&)>&9&2@! &.! )99! 2&-*A'8)9*'! &'!
0329&.*=!&.!'*82&0.!B7!4*82&0.'!#!).=!C!+*<&*D!-0=*9&.:!'23=&*'!).=!=&'83''!(+*=&82)>&9&2@!)'(*82'7!E.!
80-().&0.! ()(*+'! 6*8%0'0! !"# $%&! F"GG#H! 80-()+*! 2%*! I0+2%! ).=! 4032%! 5-*+&8).! -0.'00.'! ).=!
J0(*9*D'K&!!"#$%&!F"GG#H!=&'83''!2%*!I0+2%!5-*+&8).!-0.'00.7!L2%*+!+*9*<).2!'03+8*'!01!&.10+-)2&0.!
0.!2%*!4564!)+*!2%*!()(*+'!>@!M)*:9*!!"#$%&!F"GG"H!).=!?*+)!!"#$%&!F"GG#H7!!
!"#$%&'()*%*+&,(%#-./0,).#
!"#$%&%'!(#)%!*+'%,#
E.!)3'2+)9!'3--*+;!)'!2%*!-)/0+!%*)2&.:!N0.*!-&:+)2*'!20!2%*!'3>2+0(&8';!)! 2%*+-)9!90DA(+*''3+*!
'@'2*-! =*<*90('! 0<*+! 2%*! O%)80! +*:&0.;! &.! 8*.2+)9! 4032%! 5-*+&8)7! $%*! 90D! (+*''3+*! '@'2*-! 0<*+!
.0+2%*+.!5+:*.2&.)!).=!D*'2*+.!M)+):3)@!&'!)!89&-)2090:&8)9!1*)23+*!(+*'*.2!2%+03:%032!2%*!@*)+;!>32!
'2+0.:*'2!=3+&.:!2%*!'3--*+7!
!
!
!
P&:7!"7!6)&.!1*)23+*'!01!2%*!45647!Q*8*->*+AP*>+3)+@!-*).!FRSTSARSSCH!S"C!%M)!<*820+!D&.=!).=!"GG!%M)!
'2+*)-9&.*'!1+0-!2%*!IOUMVIO5J!+*).)9@'&'!)+8%&<*;!).=!-*+:*=!')2*99&2*!*'2&-)2*'!).=!'2)2&0.!0>'*+<)2&0.'!01!
(+*8&(&2)2&0.! F--;! '%)=&.:H7! $%*! (0'&2&0.! 01! 2%*! W09&<&).!X&:%! F5H! ).=! 2%*! '3>2+0(&8)9!529).2&8! ).=! M)8&1&8!
'3+1)8*!%&:%!(+*''3+*!8*.2*+'!FXH!)+*!&.=&8)2*=7!$%*!)((+0,&-)2*!),&'!01!2%*!4032%!529).2&8!O0.<*+:*.8*!Y0.*!&'!
&.=&8)2*=!>@!2%*!%*)<@!=)'%*=!9&.*!F)=)(2*=!1+0-!?7!Z03'K@!).=!67!X)9(*+2H7!
!
$%*! '032%D*'2A.0+2%*)'2! &.2*+A%*-&'(%*+&8! (+*''3+*! :+)=&*.2! >*2D**.! 2%*! 4032%! 5-*+&8).! 90D!
).=! 2%*! .0+2%D*'2*+.! 4)%)+)! '2+*.:2%*.';! *.%).8&.:! 2%*! 2+0(&8)9! .0+2%*)'2*+9@! 2+)=*!D&.='! FP&:7! "H7!
5.0-)903'! F8+0''! *[3)20+&)9! 190D! (*.*2+)2*'! 2%*! 80.2&.*.2;! 8)++@&.:! -0&'23+*7! $%*! 190D! >*80-*'!
.0+2%D*'2*+9@;! &'! 8%)..*9*=! '032%D)+=! >@! 2%*! 5.=*'! -03.2)&.';! ).=! 23+.'! 8908KD&'*! )+03.=! 2%*!
Altiplano ClimateFeatures of South American Warm Season Rains
!"#$"%&'()$*&+,-&+."/$*01&2(13((1&
! "
!
!
!
!
#$%&'(! )*! +,-(./0$,! $11&20'/0$34! 35! 0-(! 63'0-! /47! +3&0-!8.('$,/4!.3423342! 2920(.2! :1(50! /47! '$%-0! ;/4(12<!
'(2;(,0$=(19>*!+-/7$4%!$47$,/0(2!;'(,$;$0/0$34!/47!7/2-(7!1$4(2!$47$,/0(!,34=('%(4,(!?34(2*!+./11!/''3@2!2-3@!
13@A1(=(1!:BCC!-D/>!@$472<!/47!0-$,E!/''3@2!'(;'(2(40!13@A1(=(1!F(02*!84!GHG!2-3@2!/!2&I0'3;$,/1!2&'5/,(!-$%-!
,(40('<!/47!/4!G8G!$47$,/0(2!0-(!.342334!/40$,9,134(*!:87/;0(7!5'3.!J*!K3&2E9*>!
!
!
!
#$%&'(! "*! +,-(./0$,! =('0$,/1! 2(,0$34! 53'! 0-(! ,3''(2;347$4%! 2&..('! 2(/234! /0! /'3&47! "CL6! :1(50! ;/4(1>! /47!
23&0-@(20A43'0-(/20! :'$%-0! ;/4(1>*! M(%$342! 35! 7((;! ,34=(,0$34! /47! 13@A1(=(1! F(02! /'(! $47$,/0(7*! :D/4(1! 53'!
68N+!/7/;0(7!5'3.!O*!H$%%$42*>!
!
#$%&'(2!)!/47!"!2-3@!2,-(./0$,/119!0-(!./F3'!5(/0&'(2!35!0-(!@/'.!2(/234!,$',&1/0$34!3=('!63'0-!
/47!+3&0-!8.('$,/*!P-(!68N+!$2!,-/'/,0('$?(7!I9!/!'(%$34!35!$40(42(!;'(,$;$0/0$34!(./4/0$4%!5'3.!
0-(!(/20('4!D/,$5$,! $40('0'3;$,/1! ,34=('%(4,(!?34(! :QPRS><! (T0(47$4%!43'0-@/'7!3=('!N(T$,3! 03! 0-(!
23&0-@(20!U4$0(7!+0/0(2! :U*+*><!@$0-! 1/'%(20! =/1&(2! 3=('! 0-(!@(20('4! 213;(2! 35! 0-(!.3&40/$4! '/4%(*!
H$%-!=/1&(2!35!;'(,$;$0/0$34!/123!(T0(47!43'0-(/20@/'7!3=('!0-(!V&15!35!N(T$,3<!'(/,-$4%!&;!/134%!0-(!
(/20('4! 51/4E! 35! 63'0-! 8.('$,/<! /47!.('%$4%! $403! 0-(! 63'0-! 801/40$,! 203'.! 0'/,E*! P-('(! $2! /123! /!
'(1/0$=(!./T$.&.!03!0-(!23&0-@(20!35!0-(!V'(/0!W/E(2*!P-(!,340$4(40/1!(/20A@(20!,340'/20!I(0@((4!0-(!
/'$7!@(20!/47! 0-(!-&.$7!(/20! $2!/!E(9!,-/'/,0('$20$,!35!/!.342334/1!,$',&1/0$34<!@-$,-!@(!(1/I3'/0(!
&;34!$4!2(,0$34!X*!P-(!&;;('A1(=(1!.342334!/40$,9,134(!/223,$/0(7!@$0-!0-(!68N+!2-$502!43'0-@/'7!
@$0-! 2(/234! 5'3.! 23&0-@(20('4! N(T$,3! 03! 43'0-@(20('4! N(T$,3! /47! 23&0-@(20('4! U4$0(7! +0/0(2*!
N3$20&'(!0'/42;3'0!3403!0-(!63'0-!8.('$,/4!,340$4(40!$2!/223,$/0(7!@$0-!I'3/7A2,/1(!/7=(,0$34!5'3.!
0-(!V&15!35!N(T$,3<!/47!@$0-!$.;3'0/40!13@A1(=(1!F(02!:WWY2>!3=('!0-(!V&15!35!R/1$53'4$/!/47!(/20!35!
0-(!M3,E$(2*!P-(!1/00('!WWY!$2!/!@/'.!2(/234<!;'$./'$19!43,0&'4/1<!5(/0&'(Z!1(22!7(0/$1!$2!E43@4!/I3&0!
0-(!30-('!34(*!
Altiplano Variability15 JUNE 2001 2781G A R R E A U D A N D A C E I T U N O
FIG. 2. (a) Summer rainfall anomalies on a 5! " 5! grid box centered at 17.5!S 67.5!W.The year mark the Jan–Feb part of the respective austral summer. Monthly values obtainedfrom gridded rainfall data available from NASA Goddard Institute for Spatial Studies (Daiet al. 1997). Gridded data was synthesized from original station records on this region compiledat the Climate Research Unit, University of East Anglia, UK (Hulme 1992). (b) Variation ofthe level of Lake Titicaca (northern Altiplano) measured at Puno (15.9!S, 70.0!W, 3800 m)between 1 Dec and 28 Feb (summertime increment). In both panels cross (filled circle) indicatesEl Nino (La Nina) conditions during the corresponding austral summer.
circulation, and convection anomalies related to sum-mertime rainfall variability and their impact upon in-traseasonal episodes. A physical link between global-scale phenomena (including ENSO) and interannualrainfall variability over the Altiplano emerges from thisanalysis. A summary of our results is presented in sec-tion 5.
2. Datasets and methodology
Daily and monthly outgoing longwave radiation(OLR) on a 2.5! " 2.5! grid box centered at 17.5!S,70!W was used as an index of the convective precipi-tation over the Altiplano (CI). Original OLR fields ona 2.5! lat–long grid are available from December 1974[see Liebmann and Smith (1996) for further details onthis dataset] and they have been widely used as a proxyof rainfall over tropical and subtropical regions (e.g.,Meisner and Arkin 1987). Despite its coarse spatial res-olution and lack of direct relationship with the amountof rain, interannual fluctuations of CI agree well with
other rainfall estimates over the Altiplano (Fig. 3). Ininterpreting subsequent results it is worth keeping inmind that the linear dependence between summer meanCI and summer rainfall (station data) is, roughly,#(Rain)/#(-CI) $ 100 mm (10 W m%2)%1 (Fig. 3a).The large-scale tropospheric circulation was charac-
terized using the National Centers for EnvironmentalPrediction–National Center for Atmospheric Research(NCEP–NCAR) reanalysis fields (Kalnay et al. 1996)on global 2.5! lat–long grids available since 1958. Be-cause the enhanced amount of assimilated data (includ-ing satellite wind data since 1979) the reanalysis arethought to realistically portray the atmospheric circu-lation on synoptic, seasonal, and interannual scales,even in regions with sparse conventional observations.Over the Altiplano, reanalysis data underestimate theamplitude of the mean diurnal cycle of several variables(Aceituno and Montecinos 2000) but they do capturemost of the intraseasonal variability in moisture andmidlevel winds (Garreaud 2000).To display the large-scale patterns of convection and
Text
Garreaud & Aceituno, 2001
Summer Rainfall AnomaliesGridded precipitation from CRU/Dai
El Nino - Dry?La Nina - Wet?
Lake Titicaca summertime increment
Altiplano Variability2786 VOLUME 14J O U R N A L O F C L I M A T E
FIG. 7. Regression maps in the interannual range (see section 2 for details on calculation and statistical significance). (a) OLR (shaded,scale in units of W m!2 per std dev) and 200-hPa wind regressed upon CI (OLR over the Altiplano). (b) 200-hPa height and winds regressedupon CI. Contour interval is 30 m per std dev. Negative values in dashed line. The zero contour is omitted. Only values and wind vectorsstatistically significant at the 95% confidence level are shown. Reference wind vector (in m s!1) at the bottom of the figure.
subtropical South America. For instance, while the en-hanced subtropical jet stream (westerly flow) during ElNino years tends to suppress convection over the Al-tiplano it also favors the maintenance of persistent fron-tal activity in southern Brazil (e.g., Kousky et al. 1984;Lenters and Cook 1999). In between these coherent butopposite signals, rainfall anomalies over the Bolivianlowlands to the east of the Andes are largely indepen-dent of ENSO (Ronchail 1998). ENSO-related warmingor cooling of the tropical troposphere also explains astrong, direct relationship between ENSO phases andthe low-level temperature anomalies over the Altiplano,as documented in Vuille (1999) and Vuille et al. (2000).Given the marked similarity between circulation
anomalies associated with Altiplano rainfall variabilityand the canonical ENSO signal, one may wonder whythe ENSO–Altiplano rainfall relationship is rather weak.To understand this paradox we examine 200-hPa windand geopotential anomalies (full field minus climatol-ogy) for 2 months when the relationship does not hold(Fig. 8). Our first example is January 1973, during themature stage of a moderate El Nino (SOI " !0.5) andAltiplano rainfall slightly above normal. Troposphericwarming over the tropical eastern Pacific and SouthAmerica is evident during this month (positive 200-hPaheight anomalies, Fig. 8a). However, the anomalouswarming extended well into the subtropics so that theincrease in upper-level zonal wind take places between
Garreaud & Aceituno, 2001
Summer rainfall anomaliesupper tropospheric windsand tropospheric temperatures
Dry - upper level westerliesand warm troposphere
Wet - easterlies and cooler troposphere
Sensitive to latitude of enhanced westerlies
Altiplano Variability
110˚W 100˚W 90˚W 80˚W 70˚W 60˚W 50˚W 40˚W 30˚W 20˚W 10˚WLongitude
50˚S
40˚S
30˚S
20˚S
10˚S
0˚10˚N
20˚N
Latitude
110˚W 100˚W 90˚W 80˚W 70˚W 60˚W 50˚W 40˚W 30˚W 20˚W 10˚WLongitude
50˚S
40˚S
30˚S
20˚S
10˚S
0˚10˚N
20˚N
Latitude
110˚W 100˚W 90˚W 80˚W 70˚W 60˚W 50˚W 40˚W 30˚W 20˚W 10˚WLongitude
50˚S
40˚S
30˚S
20˚S
10˚S
0˚10˚N
20˚N
Latitude
110˚W 100˚W 90˚W 80˚W 70˚W 60˚W 50˚W 40˚W 30˚W 20˚W 10˚WLongitude
50˚S
40˚S
30˚S
20˚S
10˚S
0˚10˚N
20˚N
Latitude
110˚W 100˚W 90˚W 80˚W 70˚W 60˚W 50˚W 40˚W 30˚W 20˚W 10˚WLongitude
50˚S
40˚S
30˚S
20˚S
10˚S
0˚10˚N
20˚N
Latitude
110˚W 100˚W 90˚W 80˚W 70˚W 60˚W 50˚W 40˚W 30˚W 20˚W 10˚WLongitude
50˚S
40˚S
30˚S
20˚S
10˚S
0˚10˚N
20˚N
Latitude
El NinoLa Nina
SON
DJF
MAM
Altiplano VariabilityGarreaud & Aceituno, 2001
Summer rainfall anomaliesupper tropospheric winds
Dry - upper level westerlies
Wet - easterlies Fig. 5. (a) Correlation ¢eld of PC1 with DJF SSTA (GISST 2.3a). Contour interval is 0.2, negative contours are dashed and thezero line is omitted. (b) As in panel a, but for DJF 200 hPa geopotential height anomalies. Contour interval is 0.1, negative con-tours are dashed. (c) Regression ¢eld of PC1 with DJF 200 hPa wind anomalies between 1957 and 1996. Values signi¢cant at the95% con¢dence level based on two-tailed Student’s t-test are shaded in gray in panels a and b. Wind vectors in panel c areshown only if either zonal or meridional component is signi¢cant at 95%-con¢dence level. Scale for wind vectors in panel c isgiven below. Reference period for the anomaly ¢elds is 1961^1990. Wind and geopotential height ¢elds from NCEP-NCAR rean-alysis.
PALAEO 3054 16-4-03
R. Garreaud et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 194 (2003) 5^22 15
Altiplano Variability
Altiplano moisture are controlled by changes inthe moisture transport, rather than by moisturechanges in its source region.
3.3. Interannual variability
On the interannual timescale, the Altiplano ex-periences strong precipitation £uctuations, rang-ing from extremely dry to very wet australsummer conditions. Between 1957 and 1996,DJF precipitation recorded at Copacabana(16.2‡S, 69.1‡W; 3815 m), located on the shoreof Lake Titicaca, for example, ranged from 203mm in 1990/91 to 850 mm in 1983/84. In the morearid southwestern part of the Altiplano the £uc-tuations are even more impressive, as indicated by
the station Coyacagua in northern Chile (20‡S,68.8‡W; 3990 m) where the minimum DJF pre-cipitation sum recorded was 11 mm in 1982/83,while in the next rainy season, 1983/84, the max-imum record was set with 277 mm. This stronginterannual variability of summer precipitationhas been described in a number of studies andthere is a general agreement that a signi¢cantfraction of this variability is related to the ElNin‹o Southern Oscillation (ENSO) phenomenon(e.g. Thompson et al., 1984; Aceituno, 1988;Ronchail, 1995; Lenters and Cook, 1999; Vuille,1999; Vuille et al., 2000a; Arnaud et al., 2001;Garreaud and Aceituno, 2001). All of these stud-ies concluded that El Nin‹o years (warm phase ofENSO) tend to be dry, while La Nin‹a years
Fig. 3. Schematic representation of the circulation patterns and di¡erent air masses over and adjacent to the central Andes, in avertical-longitude section at the latitudes of the Altiplano, for (a) rainy episodes and (b) dry episodes. Large, open arrows indi-cate the sense of the upper-level, large-scale £ow. Solid (dashed) curves represent the transport of moist (dry) air by the regionalcirculation over the Andean slopes. Thin vertical arrows represent the large-scale subsidence over the subtropical SE Paci¢c thatmaintains the strong trade inversion (solid line).
PALAEO 3054 16-4-03
R. Garreaud et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 194 (2003) 5^2212
Garreaud et al., 2003
Summer rainfall related to direction of upper tropospheric winds at• intraseasonal• annual cycle• interannual• interglacial?timescales.
Altiplano Trends82 MATHIAS VUILLE ET AL.
Figure 3. Trends in station precipitation (mm yr−1) between 1950 and 1994 for (a) annual sum,(b) DJF, (c) JJA. Upward (downward) pointing triangles indicate an increase (decrease) in precipita-tion. Note different scaling in (c). Filled (open) triangles indicate that the trend is (not) significant atthe 95%-confidence level. (d) As in (a) but trend in annual precipitation (in % yr−1) versus elevation.(e) As in (d) but for DJF, (f) as in (d) but for JJA. Stations below ∼2500 m are represented by squares(western slope) and diamonds (eastern slope); stations above 2500 are shown as circles. Trend showschange (in % yr−1) compared to long-term mean (1950–1994) precipitation. Filled (open) symbolsindicate that the trend is (not) significant at the 95%-confidence level.
Peru/Bolivia border most stations indicate a precipitation decrease for the annualtotal and during the main rainy season DJF (Figure 3b). In JJA most stations indi-cate an increase in precipitation, in particular the lowland stations to the east of theAndes and the Altiplano region of northern Bolivia and southern Peru (Figure 3c).Even in the case of a coherent regional signal however, individual station trendsare mostly insignificant. Of the 42 stations analyzed only 5 (2) show a significantincrease (decrease) in the annual precipitation amount. Furthermore one needs totake into account that several stations are located very close to one another and arethus not truly independent records, because they capture the same local climaticsignal.
Vuille et al., 2003
Precipitation trends are minor and cannot explainobserved retreat of glaciers.
Altiplano Trends84 MATHIAS VUILLE ET AL.
Figure 4. Temperature trend and annual departure from 1961–1990 average (± one standard devia-tion) in the tropical Andes (∼1◦ N–23◦ S) between 1950 and 1994 based on (a) 277 station records(modified from Vuille and Bradley, 2000), (b) CRU05 data and (c) ECHAM-4 T30 simulation. Trendsin (b) and (c) are based on sub-sampling of grid cells to match location of station records used in (a).
that is slightly decreasing with elevation above 3500 m in both data sets. Thisvertical structure of the temperature trend is different from what is observed inTibet or the European Alps, where the warming is more pronounced at higherelevations. However, the high altitude warming in those regions is probably relatedto a decrease in spring snow cover, lower albedo values and a positive feedbackon temperature (i.e., Liu and Chen, 2000). This mechanism is not as important ina tropical environment, where only a few mountain peaks reach above the 0 ◦Cisotherm at ∼5000 m, and, due to the lack of thermal seasons, there is no ‘winter’and ‘spring’ snow cover, which could influence the thermal regime in a way similarto that in the mid-latitudes (Kaser and Georges, 1999).
3.4. HUMIDITY
A significant increase in near-surface and tropospheric humidity over the lastdecades has been reported from both the eastern and western tropical Pacific
Vuille et al., 2003
Temperature trends are substantial, and are a likely cause of glacial retreat.
Trend is larger on the western slope of the Andes.
Altiplano Projections
ties to emphasize those measurements of ourlocal astrophysical laboratory that will best illu-minate how these fundamental cosmic entitiesoperate. Moreover, observations to date alreadyindicate that the Saturn system is literally chang-ing before our eyes. We anticipate that even moredramatic transformations in our neighborhood’sastrophysical laboratory will be monitored byCassini’s instruments over the next several years.
References and Notes1. Special Issue on Cassini at Saturn, Science 307 (25
February 2005).2. J. J. Lissauer, J. N. Cuzzi, in Protostars and Planets II, D.
C. Black, M. S. Matthews, Eds. (Univ. of Arizona Press,Tucson, AZ, 1985), pp. 920–958.
3. Special Issue on Disks in Space, Science 307 (7 January2005).
4. M. S. Tiscareno et al., paper presented at the AmericanGeophysical Union fall meeting, 5 to 9 December 2005,San Francisco (abstract FM-P33B0245T2005).
5. J. E. Colwell, personal communication.6. J. E. Colwell, L. W. Esposito, M. Sremcevic, Geophys. Res.
Lett. 33, L07201 (2006).7. M. M. Hedman, P. D. Nicholson, B. D. Wallis, paper pre-
sented at the American Geophysical Union fall meeting,
5 to 9 December 2005, San Francisco (abstract FM-P31D02H2005).
8. H. Salo, Icarus 117, 287 (1995).9. A. W. Harris, in Planetary Rings, R. J. Greenberg, A.
Brahic, Eds. (Univ. of Arizona Press, Tucson, AZ, 1984),pp. 641–658.
10. M. S. Tiscareno et al., Nature 440, 648 (2006).11. P. Kalas, J. R. Graham, M. Clampin, Nature 435, 1067
(2005).12. Support from the Cassini project is gratefully acknowl-
edged. We thank M. Tiscareno, M. Hedman, and P.Nicholson for comments.
10.1126/science.1114856
1755
According to general circula-tion models of future cli-mate in a world with double
the preindustrial carbon dioxide(CO
2) concentrations, the rate of
warming in the lower tropospherewill increase with altitude. Thus,temperatures will rise more in thehigh mountains than at lower eleva-tions (see the figure) (1). Maximumtemperature increases are predictedto occur in the high mountains ofEcuador, Peru, Bolivia, and north-ern Chile. If the models are correct,the changes will have importantconsequences for mountain glaciersand for communities that rely onglacier-fed water supplies.
Is there evidence that tempera-tures are changing more at higherthan at lower elevations? Althoughsurface temperatures may not be thesame as in the free air, in high moun-tain regions the differences are small(2), and changes in temperature should thus besimilar at the surface and in the adjacent free air.Unfortunately, few instrumental observations areavailable above ~4000 m. The magnitude ofrecent temperature change in the highest moun-tains is therefore poorly documented. An analysisof 268 mountain station records between 1°N and
23°S along the tropical Andes indicates a temper-ature increase of 0.11°C/decade (compared withthe global average of 0.06°C/decade) between1939 and 1998; 8 of the 12 warmest years wererecorded in the last 16 years of this period (3).Further insight can be obtained from glaciers andice caps in the very highest mountain regions,which are strongly affected by rising tempera-tures. In these high-altitude areas, ice masses aredeclining rapidly (4–6). Indeed, glacier retreat isunder way in all Andean countries, fromColumbia and Venezuela to Chile (7).
A convergence of factors contribute to thesechanges. Rising freezing levels (the level wheretemperatures fall to 0°C in the atmosphere) (8, 9)
lead to increased melting and to increased expo-sure of the glacier margins to rain rather thansnow (10). Higher near-surface humidity leadsto more of the available energy going into melt-ing snow and ice, rather than sublimation, whichrequires more energy to remove the same massof ice. Therefore, during humid, cloudy condi-tions, there is often more ablation than duringdrier, cloud-free periods (6). In some areas,changes in the amount of cloud cover and thetiming of precipitation may have contributed toglacier mass loss through their impact on albedo(surface reflectivity) and the net radiation bal-ance (11). As these processes continue and snowis removed, more of the less reflective ice isexposed and absorption of the intense high-ele-vation radiation increases, thus accelerating thechanges under way through positive feedbacks.
The processes involved in mass-balancechanges at any one location are complex, buttemperature is a good proxy (12) for all theseprocesses, and most of the observed changes arelinked to the rise in temperature over recentdecades (5). Further warming of the magnitudeshown in the figure will thus have a strong nega-tive impact on glaciers throughout the Cordilleraof North and South America. Many glaciers maycompletely disappear in the next few decades,with important consequences for people livingin the region (7).
Although an increase in glacier melting ini-tially increases runoff, the disappearance of gla-ciers will cause very abrupt changes in stream-flow, because of the lack of a glacial buffer dur-ing the dry season. This will affect the availabil-ity of drinking water, and of water for agricultureand hydropower production.
In the High Andes, the potential impact ofsuch changes on water supplies for human con-
Climate models predict that greenhouse
warming will cause temperatures to rise faster at
higher than at lower altitudes. In the tropical
Andes, glaciers may soon disappear, with poten-
tially grave consequences for water supplies.
Threats to Water Supplies in theTropical AndesRaymond S. Bradley, Mathias Vuille, Henry F. Diaz, Walter Vergara
CLIMATE CHANGE
R. S. Bradley and M. Vuille are at the Climate SystemResearch Center, Department of Geosciences, University ofMassachusetts, Amherst, MA 01003, USA. H. F. Diaz is atthe Earth System Research Laboratory, National Oceanicand Atmospheric Administration, Boulder, CO 80303, USA.W. Vergara is in the Latin America EnvironmentDepartment, World Bank, 1850 I Street, NW, Washington,DC 20433, USA. E-mail: [email protected] (R.S.B.)
40°S 0° 20°N 40°N 60°N20°S
Latitude
1000
2000
3000
4000
5000
6000
7000
8000
9000El
evat
ion
(m)
2
2.5
3
3.5
4
4.5
5
5.5
6
6.5
7
Air
tem
pera
ture
cha
nge
(°C)
Global warming in the American Cordillera. Projected changesin mean annual free-air temperatures between (1990 to 1999) and(2090 to 2099) along a transect from Alaska (68°N) to southernChile (50°S), following the axis of the American Cordillera moun-tain chain. Results are the mean of eight different general circula-tion models used in the 4th assessment of the IntergovernmentalPanel on Climate Change (IPCC) (15), using CO2 levels from sce-nario A2 in (16). Black triangles denote the highest mountains ateach latitude; areas blocked in white have no data (surface orbelow in the models). Data from (15).
www.sciencemag.org SCIENCE VOL 312 23 JUNE 2006
PERSPECTIVES
Published by AAAS
ties to emphasize those measurements of ourlocal astrophysical laboratory that will best illu-minate how these fundamental cosmic entitiesoperate. Moreover, observations to date alreadyindicate that the Saturn system is literally chang-ing before our eyes. We anticipate that even moredramatic transformations in our neighborhood’sastrophysical laboratory will be monitored byCassini’s instruments over the next several years.
References and Notes1. Special Issue on Cassini at Saturn, Science 307 (25
February 2005).2. J. J. Lissauer, J. N. Cuzzi, in Protostars and Planets II, D.
C. Black, M. S. Matthews, Eds. (Univ. of Arizona Press,Tucson, AZ, 1985), pp. 920–958.
3. Special Issue on Disks in Space, Science 307 (7 January2005).
4. M. S. Tiscareno et al., paper presented at the AmericanGeophysical Union fall meeting, 5 to 9 December 2005,San Francisco (abstract FM-P33B0245T2005).
5. J. E. Colwell, personal communication.6. J. E. Colwell, L. W. Esposito, M. Sremcevic, Geophys. Res.
Lett. 33, L07201 (2006).7. M. M. Hedman, P. D. Nicholson, B. D. Wallis, paper pre-
sented at the American Geophysical Union fall meeting,
5 to 9 December 2005, San Francisco (abstract FM-P31D02H2005).
8. H. Salo, Icarus 117, 287 (1995).9. A. W. Harris, in Planetary Rings, R. J. Greenberg, A.
Brahic, Eds. (Univ. of Arizona Press, Tucson, AZ, 1984),pp. 641–658.
10. M. S. Tiscareno et al., Nature 440, 648 (2006).11. P. Kalas, J. R. Graham, M. Clampin, Nature 435, 1067
(2005).12. Support from the Cassini project is gratefully acknowl-
edged. We thank M. Tiscareno, M. Hedman, and P.Nicholson for comments.
10.1126/science.1114856
1755
According to general circula-tion models of future cli-mate in a world with double
the preindustrial carbon dioxide(CO
2) concentrations, the rate of
warming in the lower tropospherewill increase with altitude. Thus,temperatures will rise more in thehigh mountains than at lower eleva-tions (see the figure) (1). Maximumtemperature increases are predictedto occur in the high mountains ofEcuador, Peru, Bolivia, and north-ern Chile. If the models are correct,the changes will have importantconsequences for mountain glaciersand for communities that rely onglacier-fed water supplies.
Is there evidence that tempera-tures are changing more at higherthan at lower elevations? Althoughsurface temperatures may not be thesame as in the free air, in high moun-tain regions the differences are small(2), and changes in temperature should thus besimilar at the surface and in the adjacent free air.Unfortunately, few instrumental observations areavailable above ~4000 m. The magnitude ofrecent temperature change in the highest moun-tains is therefore poorly documented. An analysisof 268 mountain station records between 1°N and
23°S along the tropical Andes indicates a temper-ature increase of 0.11°C/decade (compared withthe global average of 0.06°C/decade) between1939 and 1998; 8 of the 12 warmest years wererecorded in the last 16 years of this period (3).Further insight can be obtained from glaciers andice caps in the very highest mountain regions,which are strongly affected by rising tempera-tures. In these high-altitude areas, ice masses aredeclining rapidly (4–6). Indeed, glacier retreat isunder way in all Andean countries, fromColumbia and Venezuela to Chile (7).
A convergence of factors contribute to thesechanges. Rising freezing levels (the level wheretemperatures fall to 0°C in the atmosphere) (8, 9)
lead to increased melting and to increased expo-sure of the glacier margins to rain rather thansnow (10). Higher near-surface humidity leadsto more of the available energy going into melt-ing snow and ice, rather than sublimation, whichrequires more energy to remove the same massof ice. Therefore, during humid, cloudy condi-tions, there is often more ablation than duringdrier, cloud-free periods (6). In some areas,changes in the amount of cloud cover and thetiming of precipitation may have contributed toglacier mass loss through their impact on albedo(surface reflectivity) and the net radiation bal-ance (11). As these processes continue and snowis removed, more of the less reflective ice isexposed and absorption of the intense high-ele-vation radiation increases, thus accelerating thechanges under way through positive feedbacks.
The processes involved in mass-balancechanges at any one location are complex, buttemperature is a good proxy (12) for all theseprocesses, and most of the observed changes arelinked to the rise in temperature over recentdecades (5). Further warming of the magnitudeshown in the figure will thus have a strong nega-tive impact on glaciers throughout the Cordilleraof North and South America. Many glaciers maycompletely disappear in the next few decades,with important consequences for people livingin the region (7).
Although an increase in glacier melting ini-tially increases runoff, the disappearance of gla-ciers will cause very abrupt changes in stream-flow, because of the lack of a glacial buffer dur-ing the dry season. This will affect the availabil-ity of drinking water, and of water for agricultureand hydropower production.
In the High Andes, the potential impact ofsuch changes on water supplies for human con-
Climate models predict that greenhouse
warming will cause temperatures to rise faster at
higher than at lower altitudes. In the tropical
Andes, glaciers may soon disappear, with poten-
tially grave consequences for water supplies.
Threats to Water Supplies in theTropical AndesRaymond S. Bradley, Mathias Vuille, Henry F. Diaz, Walter Vergara
CLIMATE CHANGE
R. S. Bradley and M. Vuille are at the Climate SystemResearch Center, Department of Geosciences, University ofMassachusetts, Amherst, MA 01003, USA. H. F. Diaz is atthe Earth System Research Laboratory, National Oceanicand Atmospheric Administration, Boulder, CO 80303, USA.W. Vergara is in the Latin America EnvironmentDepartment, World Bank, 1850 I Street, NW, Washington,DC 20433, USA. E-mail: [email protected] (R.S.B.)
40°S 0° 20°N 40°N 60°N20°S
Latitude
1000
2000
3000
4000
5000
6000
7000
8000
9000
Elev
atio
n (m
)
2
2.5
3
3.5
4
4.5
5
5.5
6
6.5
7
Air
tem
pera
ture
cha
nge
(°C)
Global warming in the American Cordillera. Projected changesin mean annual free-air temperatures between (1990 to 1999) and(2090 to 2099) along a transect from Alaska (68°N) to southernChile (50°S), following the axis of the American Cordillera moun-tain chain. Results are the mean of eight different general circula-tion models used in the 4th assessment of the IntergovernmentalPanel on Climate Change (IPCC) (15), using CO2 levels from sce-nario A2 in (16). Black triangles denote the highest mountains ateach latitude; areas blocked in white have no data (surface orbelow in the models). Data from (15).
www.sciencemag.org SCIENCE VOL 312 23 JUNE 2006
PERSPECTIVES
Published by AAAS
HypothesesAnticipated Changes
• Climate Variability - • Climate Change - precipitation, frost days, heat waves
Altiplano ClimatePrecipitation - Obs, AR4 Models
Altiplano ClimatePrecipitation - Obs, AR4 Models
Altiplano ClimatePrecipitation & 200 hPa Wind - AR4 Models
2780 VOLUME 14J O U R N A L O F C L I M A T E
FIG. 1. Climatological summertime mean winds at the 200-hPa level and outgoing longwave radiation (OLR). Wind vector reference atthe bottom is 30 m s!1. Light and medium shading indicates OLR less than 235 and 215 W m!2, respectively. Thin solid contours indicateterrain elevation in excess of 2000 and 4000 m. Center of the Bolivian High is indicated by a letter H.
Consistent with its transitional character in time andspace, rainfall over the Altiplano is highly sensitive tolarge-scale circulation anomalies. On the interannualtimescale, surface-based observations indicate alternat-ing of drought and very wet summer condition (e.g.,Ronchail 1998). For instance, the area-average sum-mertime rainfall1 shown in Fig. 2 ranges from 50 to"500 mm season!1 and pronounced fluctuations arealso observed in the time series of the summertime in-crement of the level of Lake Titicaca (an integratedindex of summertime rainfall over the northern Alti-plano).Several studies have shown that a significant fraction
of the interannual variability of precipitation on the Al-tiplano is related to the El Nino–Southern Oscillation(ENSO) phenomenon (Thompson et al. 1984; Francouand Pizarro 1985; Aceituno 1988; Aceituno and Gar-reaud 1995; Ronchail 1995, 1998; Lenters and Cook1999; Vuille 1999). Based on different estimates of sum-mertime precipitation (rainfall station data, Titicaca lakelevel, snow accumulation) and statistical techniquesthese studies document a tendency for wet conditionsduring the cold ENSO phase (La Nina years) and dryconditions during the warm ENSO phase (El Ninoyears). As evident in Fig. 2, however, neither all rainiest/driest summers occur during extreme ENSO years, nor
1 The rainfall data in Fig. 2 must be taken with caution, becauseit is based on a small and variable number of stations (2–4) withgaps in its original records. Furthermore, the 5# $ 5# lat–long rainfallaverage is a large simplification of the real conditions over the Al-tiplano, possibly biased by the more humid conditions in the northeastsector of the basin. The convective and episodic nature of the rainfallover the Altiplano also substract representativeness to area averagerainfall estimated from a reduced number of stations.
all El Nino/La Nina years are associated with extremerainfall anomalies (see also Aceituno and Garreaud1995; Vuille 1999). The weakness of the relationshipbetween ENSO and summertime rainfall could be atleast partially caused by the large spatial and intrasea-sonal variability of the precipitation over the Altiplano.For instance, recent work by Vuille et al. (2000) showsthat ENSO-related rainfall anomalies are much morepronounced and statistically significant over the westernsector compared with those over the northeastern sectorof the Altiplano. In any case, the physical link betweenglobal-scale circulation anomalies (e.g., ENSO) and Al-tiplano rainfall anomalies is unclear.In the present work we investigate the interannual
rainfall variability over the Altiplano in a more generalcontext, without focusing a priori on the ENSO signal.Specifically, we aim to characterize the year-to-yearfluctuations of the summertime rainfall over the centralAndes and their relationship with rainfall changes else-where (e.g., over the rest of South America) and large-scale circulation anomalies. Only after this step we in-vestigate whether or not the ENSO signal projects uponthe anomaly fields associated with Altiplano rainfall var-iability. Satellite measurements of outgoing longwaveradiation (OLR) are used as a proxy of convective rain-fall over the Altiplano. Comparison of OLR with otherestimates of precipitation and a description of the da-tasets used in this study are presented in section 2. Con-sidering that year-to-year variability is ultimately pro-duced by changes in the number or intensity of rainfallepisodes within the summer season, we first review thelarge-scale circulation patterns associated with intrasea-sonal rainfall variability on the Altiplano (section 3). Insection 4 we document the seasonal-mean, large-scale
Altiplano VariabilityPrecipitation & 200 hPa Wind - AR4 Models
Fig. 5. (a) Correlation ¢eld of PC1 with DJF SSTA (GISST 2.3a). Contour interval is 0.2, negative contours are dashed and thezero line is omitted. (b) As in panel a, but for DJF 200 hPa geopotential height anomalies. Contour interval is 0.1, negative con-tours are dashed. (c) Regression ¢eld of PC1 with DJF 200 hPa wind anomalies between 1957 and 1996. Values signi¢cant at the95% con¢dence level based on two-tailed Student’s t-test are shaded in gray in panels a and b. Wind vectors in panel c areshown only if either zonal or meridional component is signi¢cant at 95%-con¢dence level. Scale for wind vectors in panel c isgiven below. Reference period for the anomaly ¢elds is 1961^1990. Wind and geopotential height ¢elds from NCEP-NCAR rean-alysis.
PALAEO 3054 16-4-03
R. Garreaud et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 194 (2003) 5^22 15
Altiplano Projections
Altiplano ProjectionsPrecipitation & 200 hPa Wind - AR4 Models
Altiplano VariabilityPrecipitation & 200 hPa Wind - AR4 Models
Altiplano Variability
Discussion
• The upper tropospheric zonal wind hypothesis...
• The models appear to simulate this mechanism in the present
• The models respond quite differently in 21st Century climate - is there a change in the mechanism relating precipitation to upper level zonal wind?
Acknowledgments
