Atmospheric salt deposition in a tropical mountain ... · Sea salt (NaCl) has recently been proven to be of the utmost importance for ecosystem functioning in Amazon lowland forests

Atmos Chem Phys 16 10241ndash10261 2016wwwatmos-chem-physnet16102412016doi105194acp-16-10241-2016copy Author(s) 2016 CC Attribution 30 License

Atmospheric salt deposition in a tropical mountain rainforest at theeastern Andean slopes of south Ecuador ndash Pacific or Atlantic originSandro Makowski Giannoni1 Katja Trachte1 Ruetger Rollenbeck1 Lukas Lehnert1 Julia Fuchs2 and Joerg Bendix1

1Laboratory for Climatology and Remote Sensing (LCRS) Institute of Physical Geography Department of GeographyPhilipps-Universitaet Marburg Germany2Laboratory of Climatology Institute of Physical Geography Department of GeographyRuhr-Universitaet Bochum Germany

Correspondence to Sandro Makowski Giannoni (makowsksstudentsuni-marburgde)

Received 17 June 2015 ndash Published in Atmos Chem Phys Discuss 8 October 2015Revised 10 July 2016 ndash Accepted 13 July 2016 ndash Published 12 August 2016

Abstract Sea salt (NaCl) has recently been proven to be ofthe utmost importance for ecosystem functioning in Amazonlowland forests because of its impact on herbivory litter de-composition and thus carbon cycling Sea salt depositionshould generally decline as distance from its marine sourceincreases For the Amazon a negative eastndashwest gradient ofsea salt availability is assumed as a consequence of the bar-rier effect of the Andes Mountains for Pacific air massesHowever this generalized pattern may not hold for the tropi-cal mountain rainforest in the Andes of southern Ecuador Toanalyse sea salt availability we investigated the depositionof sodium (Na+) and chloride (Clminus) which are good proxiesof sea spray aerosol Because of the complexity of the ter-rain and related cloud and rain formation processes sea saltdeposition was analysed from both rain and occult precip-itation (OP) along an altitudinal gradient over a period be-tween 2004 and 2009 To assess the influence of easterly andwesterly air masses on the deposition of sodium and chlo-ride over southern Ecuador sea salt aerosol concentrationdata from the Monitoring Atmospheric Composition and Cli-mate (MACC) reanalysis data set and back-trajectory statis-tical methods were combined Our results based on deposi-tion time series show a clear difference in the temporal vari-ation of sodium and chloride concentration and Na+ Clminus

ratio in relation to height and exposure to winds At higherelevations sodium and chloride present a higher seasonalityand the Na+ Clminus ratio is closer to that of sea salt Medium-to long-range sea salt transport exhibited a similar seasonal-ity which shows the link between our measurements at highelevations and the sea salt synoptic transport Although the

influence of the easterlies was predominant regarding the at-mospheric circulation the statistical analysis of trajectoriesand hybrid receptor models revealed a stronger impact ofthe north equatorial Atlantic Caribbean and Pacific sea saltsources on the atmospheric sea salt concentration in southernEcuador The highest concentration in rain and cloud waterwas found between September and February when air massesoriginated from the north equatorial Atlantic the CaribbeanSea and the equatorial Pacific Together these sources ac-counted for around 824 of the sea salt budget over south-ern Ecuador

1 Introduction

Poor substrate and intense leaching by precipitation maketropical forests particularly prone to nutrient deficiencyWhile phosphorus is mainly considered a limitation to netprimary productivity (NPP) in lowland Amazonian tropicalforests phosphorus and nitrogen co-limit growth in the trop-ical mountain rainforests as in southern Ecuador (Homeieret al 2012 Koehler et al 2009 Tanner et al 1998 Vi-tousek 1984 Wolf et al 2011 Wullaert et al 2010) Be-cause of a worldwide increase in nitrogen and phosphorusemissions and a pronounced increase in emissions from de-veloping countries where the majority of tropical forestsare located atmospheric deposition in these countries hasgained attention (Dentener et al 2006 Galloway et al 2008Phoenix et al 2006)

Published by Copernicus Publications on behalf of the European Geosciences Union

10242 S Makowski Giannoni et al Atmospheric deposition of sea salt in southeastern Ecuador

In several tropical and temperate forests human interven-tion in the nitrogen and phosphorus cycles have has beendocumented (Mahowald et al 2005 Matson et al 2002Phoenix et al 2006 Tipping et al 2014 Yu et al 2015)Because nutrient availability regulates ecosystem processesand functions the changes currently affecting the nitrogenand phosphorus budgets are expected to have wide-reachingimpacts in forest ecosystem structure and diversity (Bobbinket al 2010 Homeier et al 2012 Matson et al 2014 Pentildeue-las et al 2013 Pett-Ridge 2009 Wang et al 2014 Wilckeet al 2013) The role of sea salt availability has very re-cently gained attentions as it has been found to conditionthe behaviour of herbivores in addition to affecting carboncycling and organic matter decomposition in tropical ecosys-tems (Dudley et al 2012 Kaspari et al 2008 2009 Pow-ell et al 2009 Voigt et al 2008) At the western rim ofthe Amazon forest in Peru Ecuador and Colombia there isevidence that herbivorous and frugivorous birds and mam-mals visit mineral licks to compensate for low sodium con-centration in plant and fruit tissues (Lee et al 2009 Liz-cano and Cavelier 2004 Powell et al 2009 Voigt et al2008) Furthermore some taxa of arthropod have reportedlybegun practicing geophagy to deal with salt scarcity in plants(Kaspari et al 2008) Yet despite its pantropical importancesalt availability has hitherto been overlooked in most biogeo-graphic and biogeochemical studies (Dudley et al 2012)

By far the most important source of continental sea saltdepositions are the oceans Sea salt scarcity in Amazonianrainwater increases along a gradient from the Atlantic coasttowards the Andean range which acts as a natural orographicbarrier to the west The concentration of both sodium andchloride in rainwater diminishes significantly with increas-ing distance from the Atlantic Ocean (Talbot et al 1990)Additionally the ratio between both concentrations invertsfrom Clminus gt Na+ close to the ocean to Clminus lt Na+ far fromthe ocean (Tardy et al 2005) Consequently tropical moun-tain forests on the eastern slopes of the Andes and tropicallowland forests at the western edge of the Amazon are ex-pected to suffer from sea salt deprivation whereas forestscloser to the Atlantic coast are subject to large sea salt depo-sition (Dudley et al 2012)

The tropical mountain forests at the eastern Andean slopesin southern Ecuador may likely represent an exception of thisgeneralized pattern because of their location in the Huan-cabamba depression an area where the Andes rarely exceed3600 m in altitude This allows the transport of Pacific airmasses rich in sea salt As a result the mountain forest mightbenefit not only from sea salt transported by easterly airmasses from the Atlantic but also by sea salt originating fromPacific air masses Depending on the strength of the contribu-tion to sodium and chloride deposition originating from thePacific the combined input from Atlantic and Pacific sourcescould result in a greater sodium and chloride availability thanthat found for the lowland forests on the western rim of theAmazon (Dudley et al 2012)

However little research has investigated the deposition ofatmospheric sodium and chloride in the tropical forests ofthe southern Ecuadorian Andes Furthermore any such re-search has yet to identify their sources In this context an in-vestigation of the deposition by occult precipitation (OP) isparticularly important because OP comprises an extremelyhigh proportion of total precipitation in tropical mountainforests OP is the water supplied to soil or vegetation by lightdrizzle wind-driven rain and fog andor clouds that conven-tional rain gauges cannot measure An exception is the workof Fabian et al (2009) who estimated the origin of the localsea salt deposition by visual interpretation of single back tra-jectories To our knowledge neither a comprehensive quan-titative investigation on sea salt sources nor any estimates oftheir contribution to the atmospheric deposition have beenconducted yet

As a consequence of the knowledge gaps regarding thesea salt sources of deposition in the Andes of southeast-ern Ecuador the aims of this study are as follows (1) tocharacterize sodium and chloride atmospheric deposition byrain and OP along an altitudinal gradient and at differenttopographic locations in a tropical mountain rainforest site(2) to identify potential Pacific Atlantic and continental ge-ographic sources of sea salt concentration over the Andesof southern Ecuador by applying back-trajectory statisticaltechniques and reanalysis data of atmospheric compositionand (3) to estimate the contribution of each source area to theatmospheric sea salt concentration in the Andes of southernEcuador

2 Study area

The study area is located at the northwestern edge of theAmazon basin (400prime S 7905primeW) at the southeastern An-des of Ecuador approximately a 100 km straight line distanceaway from the Pacific coast and around 2000 km from theclosest part of the Atlantic coast (Fig 1) The study area con-tains the San Francisco valley deeply incised into the east-ern slope of the main Andes range Since 2002 two succes-sive multidisciplinary research programs have investigatedthe Reserva Bioloacutegica San Francisco (RBSF) located on thenorthern slopes of the valley and some areas outside of thereserve (Beck et al 2008 Bendix et al 2013)

The Andes in this area are characterized by lower ele-vations and higher geomorphological complexity comparedto other parts of the mountain chain in northern EcuadorPeru and Colombia Since studies have shown that exposureand altitude affect deposition patterns (Griffith et al 2015Kirchner et al 2014 Lovett and Kinsman 1990 MakowskiGiannoni et al 2013) a precondition for the study of sea saltdeposition is to collect measurements along a large altitudi-nal gradient and at different slope exposures

The climate of the catchment is mainly determined bythe constant tropical easterlies However the trade winds

Atmos Chem Phys 16 10241ndash10261 2016 wwwatmos-chem-physnet16102412016

S Makowski Giannoni et al Atmospheric deposition of sea salt in southeastern Ecuador 10243

Figure 1 Map of the study area (a) Location of the study area in the Huancabamba depression of the Andes in South America (b) Detailedmap of the rain and OP sampling sites installed in the study area

are weakened each year between November and March andwesterly wind bursts occur locally because of the low altitudeof the mountains Those westerly winds are transporting Pa-cific air masses into the study area (see Fig 1 Bendix et al2008a Emck 2007) Precipitation responds to the displace-ment of the intertropical convergence zone (ITCZ) and theintensity of the tropical easterlies The highest rainfall oc-curs between June and August when the easterly winds pre-dominate carrying humid air masses from the lowlands ofthe Amazon The topography forces the humid air upwardsleading to high rainfall totals especially in the higher partsof the mountains and the peaksrsquo immersion into the cloudsresulting in OP (Bendix et al 2006 2008b Richter et al2013 Rollenbeck et al 2011) In the period from 2004 to2009 average rainfall varied from 1500 to 6500 mm per yearbetween 1960 and 3180 m In the highest regions OP con-tributes up to 35 of total precipitation (Rollenbeck et al2011) A short dry season occurs between November andMarch when Pacific air masses are transported to the area byoccasional westerlies Such air masses occur less than 20 of the time (Richter et al 2013) and are accompanied byconvective activity

3 Data and methods

The methodology is comprised of two components First weanalyse local salt (sodium and chloride) concentrations byassessing the sodium and chloride concentrations in samplesof rain and OP along an altitudinal gradient To do so weuse a statistical approach due to the complexity of the ter-rain Second we attempt to derive the source of the salt Thesecond part focuses on describing the transport pathways as-

sociated with the general atmospheric circulation patterns todetect potential source areas of sea salt Our goal is to drawconnections between the contributions each respective atmo-spheric sea salt source has on our study area Back-trajectoryanalysis was used to achieve this goal

31 Sample collection materials and data

Three meteorological stations (MSs) were installed on thenorth-facing slopes of the San Francisco valley along an alti-tudinal transect ranging from 1960 to 3180 m in elevation Afourth station (El Tiro 2725 m) was installed at a mountainpass about 4 km up the valley on the Cordillera Real (Fig 1)

Rain and OP samples were collected at each station be-tween 2004 and 2009 While rain water was collected intotaling gauges (UMS 200 made of polyethylene to war-rant chemical inertia) OP was collected in 1 m2 mesh gridfog collectors designed according to Schemenauer and Cere-cedarsquos proposal (Schemenauer and Cereceda 1994) Detailsabout rain and fog measurement techniques calibration anddata handling are described in Rollenbeck et al (2007)Fabian et al (2009) and Rollenbeck et al (2011)

Rain and OP samples were collected at almost regularweekly intervals The samples were filtered and immediatelystored in frozen state before being sent to the laboratory forion analyses All samples were analysed at the Universityof Munichrsquos Weihenstephan Center (TUM-WZW) for majorions (K+ Na+ NH4

+ Ca2+ Mg2+ Clminus SO42minus NO3

minus)Cations were analysed according to the inductivity-coupledplasma method (Perkin Elmer Optima 3000) while ion chro-matography (Dionex DX-210) was used for anions The de-tection limits are 01 and 02 mg Lminus1 for sodium and chlo-ride respectively

wwwatmos-chem-physnet16102412016 Atmos Chem Phys 16 10241ndash10261 2016

The sea salt mixing ratio of the Monitoring AtmosphericComposition and Climate (MACC) reanalysis data set wasused as a proxy for the sea salt concentration in the atmo-sphere with a horizontal resolution of 075 by 075 (In-ness et al 2013) In this data set the concentration of seasalt generated by wind stress on the ocean surface was deter-mined based on a source function developed by Guelle et al(2001) and Schulz et al (2004) accounting for sedimenta-tion as well as wet and dry deposition processes The seasalt concentration was integrated for three size bins (003ndash05 05ndash50 and 50ndash200 microm) and calculated for 60 verticalmodel levels with the upper model limit at 01 hPa (Benedettiet al 2009 Morcrette et al 2009) To our knowledge this isthe only available global sea salt concentration data that spanthe period covered in this study Furthermore this reanaly-sis data set has performed well when compared to measuredsatellite and ground-based data (Benedetti et al 2009)

32 Statistical evaluation of sodium and chloride ionconcentration in rain and OP

Since sea spray aerosol consists mainly of chloride andsodium (Millero 2014) we used the ion concentration ofboth elements in rain and OP as proxies of sea salt atmo-spheric inputs into the ecosystem Because sodium is a con-servative ion in sea salt aerosol it is often used as a referencefor sea salt concentration in precipitation chemistry and at-mospheric chemistry modelling studies (Jaegleacute et al 2011Keene et al 1986 Pozzer et al 2012 Tardy et al 2005 Vetet al 2014) Chloride is more unstable as it photochemicallyreacts with other atmospheric ions (eg sulfur and nitrogenspecies) and it is depleted as a function of time spent in theatmosphere (Keene et al 1986)

Weekly sodium and chloride concentrations in watersamples from rain and OP were weighted with the totalweekly precipitation volume to calculate volume-weightedmonthly mean concentrations (VWMMs) With the calcu-lated VWMMs we compiled monthly time series of sea saltconcentration for a 6-year time series from 2004 to 2009which represented the temporal variation in the concentra-tion at each altitudinal level of the study area To identify dif-ferences in the distribution of sodium and chloride concen-trations between the sites we created boxplots of total con-centration over the 6-year evaluated period at each altitudi-nal level Additionally we computed total volume-weightedmeans (VWMs) to compare our observations with those fromother studies

We analysed the relationship to other ions (K+ NH4+

Ca2+ Mg2+ SO42minus NO3

minus) using a principal factor analy-sis (PFA) to locate common transport histories and the likelyorigin of sodium and chloride Before conducting the PFAthe data were normalized and scaled to achieve comparabledistributions

In coastal continental areas the Na+ Clminus molar ratio istypically that of sea salt (Keene et al 1986) This ratio was

calculated using the measurements from each altitude andserves as an indicator of the origin of both sodium and chlo-ride concentration in precipitation The ratio changes as afunction of distance from the ocean as chloride is photo-chemically depleted in the atmosphere

Finally to asses a likely impact of anabatic flows on thesodium and chloride budget we calculated wind directionrelative frequency plots

33 Back-trajectory and sourcendashreceptor analysis

The HYSPLIT model was used to generate back trajecto-ries of air masses encompassing 10 days with a resolutionof 1 day (Draxler and Hess 1998) Modeling was done us-ing the openair package (Carslaw and Ropkins 2012) for Rstatistical language The wind fields were derived from theERA-Interim reanalysis (Dee et al 2011) with a grid res-olution of 075 by 075 All trajectories had their originsat the San Francisco River catchment in southern EcuadorThe MACC reanalysis sea salt concentration data were setas proxy of sea salt concentration in the atmosphere for air-mass transport analysis by back-trajectory techniques To testthe link between the MACC sea salt concentration and thesodium and chloride concentrations actually measured in rainand OP both were linearly correlated Pearsonrsquos productndashmoment correlation coefficients were calculated between theconcentration at the two uppermost MSs (El Tiro and Cerrodel Consuelo) and the MACC sea salt concentration (see Ta-ble 1) Based on the correlation coefficients the MACC dataset at 700 hPa and the medium particle size (05ndash50 microm) waschosen as the input parameter for further examination byback-trajectory analysis because it yielded the highest cor-relation coefficient and significance level

Given the spatial resolution of the data set (075 by075) the outcome of this analysis can only provide evi-dence of synoptic transport pathways and sourcendashreceptor re-lationships for medium- to long-distance sources for an areaof approximately 80 by 80 km2 in the southern EcuadorianAndes Local-scale transport is not represented by the usedtrajectory models

We first aimed to identify the potential geographic originof the sea salt concentration over this wider area coveringour receptor site in southern Ecuador For this particular pur-pose we used sourcendashreceptor modelling as it has been suc-cessfully applied to determine likely geographic origins ofpollutants and aerosols (eg Fleming et al 2012 Hsu et al2003 Powell et al 2009 Riuttanen et al 2013 Robinsonet al 2011) Here two different hybrid receptor models wereused for comparison the potential source contribution func-tion (PSCF) and the adjusted concentration-weighted trajec-tory (CWT) running on a grid that covers the domain of the2192 generated trajectories between 2004 and 2009 Giventhe high seasonality of synoptic air mass transport we cal-culated the models on a seasonal basis (Bendix et al 2008aEmck 2007)

Table 1 Results from the correlation analysis between sea salt monthly mean concentration from the MACC reanalysis data and Na+ andClminus monthly mean concentration samples from El Tiro and Cerro del Consuelo meteorological stations Correlations were tested for thevarious elevations within the MACC data set

MACC1 (003ndash05 microm) MACC2 (05ndash5 microm) MACC3 (5ndash20 microm)

Clminus Na+ Mean Clminus Na+ Mean Clminus Na+ Mean

Cerro del Consuelo

700 hPa 0-36 035 018 052 052 040 048 047 052600 hPa 031 026 01 050 047 039 036 030 040500 hPa 027 019 003 047 036 030 022 013 028400 hPa 024 019 003 037 027 023 008 002 017300 hPa 011 002 minus005 025 016 023 001 minus005 014200 hPa 022 009 003 030 018 025 minus002 minus004 008

El Tiro

700 hPa 034 018 018 041 017 02 032 005 016600 hPa 037 022 02 040 014 018 019 minus008 007500 hPa 033 018 016 031 005 009 002 minus015 minus004400 hPa 024 014 012 014 minus002 0 minus014 minus016 minus012300 hPa 014 015 01 minus001 minus008 minus005 minus021 minus015 minus013200 hPa 015 015 019 minus001 minus012 001 minus017 001 minus004

Note p lt 005 p lt 001 p lt 0001

The PSCF (Malm et al 1986 Pekney et al 2006 Zengand Hopke 1989) calculates the probability that a source ofaerosol or pollutant observed at the ground measurement siteis located at a specific cell in the geographic space and isdefined by

PSCFij =mij

where nij is the number of trajectory points that passedthrough cell (ij ) and mij is the number of times that tra-jectory points passing through the cell (ij ) and correspondto a high sea salt concentration (above an arbitrary thresh-old) at the time of the trajectoryrsquos arrival at the receptor siteThe function is based on the premise that if a source is lo-cated at that specific location the air masses represented bythe trajectory passing through the collocated cell are likelyto collect and transport the material along the trajectory untilthe receptor site In this study we defined two concentrationthresholds the 75th percentile for moderate-to-high concen-tration and the 90th percentile for high concentration

The adjusted CWT function uses a grid domain to calcu-late a grid-wise logarithmic mean concentration of an aerosolor pollutant (Seibert et al 1994) and is defined by

ln(Cij

k=1tij

Nsumk=1

)tijk (2)

where i and j are the grid indices k the trajectory index N

the total number of trajectories Ck the pollutant concentra-

tion measured upon arrival of trajectory k and tijk the resi-dence time of trajectory k in grid cell (i j ) In this methoda weighted concentration is assigned to each pixel in the do-main This concentration is the average of the sample con-centrations at a given receptor that have associated trajecto-ries crossing the respective cell

In a second step we assessed the contribution of the maintransport pathways of sea salt to the observed concentrationover southern Ecuador For this purpose we integrated theMACC sea salt data to the back-trajectory cluster analysisAs for the sourcendashreceptor modelling approach cluster anal-ysis was applied on a seasonal basis to group similar air masshistories This revealed general circulation patterns and sub-sequently enabled to post-process concentration data in rela-tion to cluster origin and pathways A partitioning algorithmbased on spherical k means was used to define the appro-priate number of trajectory clusters as well as prior knowl-edge of the main wind systems affecting the receptor site Wetested different k values and chose the maximum number ofcluster objects (ie back trajectories) that most closely repro-duced the known conditions The cosine distance was used asmeasure of similaritydissimilarity between different trajec-tories Afterwards the frequency of trajectories representedby each cluster was determined

To estimate the contribution of the different seasonal trans-port pathways to the observed sea salt concentration the sin-gle trajectories belonging to each cluster object (cluster meantrajectory) were related to the sea salt concentration in thenearest neighbouring pixel within the study area In this waythe contribution of each cluster object to the sea salt con-

centration above the study site could be statistically evalu-ated Likewise to analyse sources and sinks of sea salt alongthe cluster mean trajectories we extracted the mean seasonalconcentration values from the MACC data pixels (from 2004to 2009) that matched the cluster mean trajectories in loca-tion time and height Seasonal sea salt concentrations mapswere calculated to further interpret the concentration alongthe cluster mean trajectories For this the MACC sea saltdata were vertically integrated between 875 hPa (the mini-mum height of the trajectory clusters) and 500 hPa (maxi-mum height) Additionally based on findings by Akagi et al(2011) showing that burning biomass is a contributor tochloride emissions the Copernicus atmosphere monitoringsystemrsquos (CAMS) global fire assimilation system (GFAS)(Kaiser et al 2012) was used to create seasonal maps of NOx

biomass burning fluxes over South America To estimate theNa+ and Clminus contribution from the biomass-burning aerosolmass we calculated the Na+ and Clminus mass ratio to NOx

based on emission factors from Ferek et al (1998) In thisway seasonal fractions of biomass-burning Na and Cl rela-tive to sea salt were calculated

4 Results

41 Sodium and chloride concentration

Our study area is characterized by the complex topographyof the Andes (see Fig 1) Hence temporal variation and dis-tribution is of interest in our study of sodium and chlorideconcentrations in rain and OP at different altitudes and topo-graphical locations

Figure 2 (left column) depicts the time series of sodiumand chloride concentrations at different altitudes Cerro delConsuelo MS situated at 3180 m demonstrated the clearesttemporal pattern in concentration where the highest peaksoccurred almost regularly between September and Febru-ary (Fig 2a) The highest concentrations of sodium wererecorded at Cerro del Consuelo and El Tiro (2825 m) (Fig 2aand b) Contrarily chloride concentration peaks in OP werehighest at the lowest MS Estacioacuten Cientiacutefica San Francisco(ECSF) (Fig 2d)

To compare the respective distributions the boxplots inFig 2 (right column) show the concentration of sodium andchloride for both rain and OP at every MS Overall no essen-tial variations between the concentration at each MS couldbe observed except for chloride in OP and rainwater at ECSF(Fig 2d) where reported values were much higher than thosemeasured at other elevations Regarding ion concentration insodium and chloride species a clear difference could be ob-served with chloride concentration in the interquartile rangeextending between 022 and 051 mg Lminus1 and sodium con-centration extending between 006 and 020 mg Lminus1

In rainwater samples the concentrations of chloride wereconsiderably higher than those of sodium at each MS A

larger range and higher extreme values were observed inthe chloride concentration Differences in sodium concen-trations between MSs at different altitudes were negligi-ble The concentrations showed a slight increase at tran-sect 1 (TS1) (median of 014 mg Lminus1) and El Tiro (medianof 013 mg Lminus1) and decreased again at the highest stationCerro del Consuelo (median of 007 mg Lminus1) Compared tothe rain samples OP contained a higher mean concentrationof sodium and chloride but also a greater range in its distri-bution (Fig 2 left column) The concentration of chloridewas also considerably higher than that of sodium with thehighest mean concentration (median of 062 mg Lminus1) at thelowest MS ECSF Sodium concentration peaked at El Tiro(017 mg Lminus1) At TS1 the mean concentration was lowest(median Na+ 009 mg Lminus1 and Clminus 03 mg Lminus1) increasedonce again at the highest elevations El Tiro and Cerro delConsuelo (median Na+ between 011 and 017 mg Lminus1 andClminus between 033 and 035 mg Lminus1)

A PFA of every major ion concentration was conducted togain insights into the origin of sea salt inputs for each MS(Table 2) This analysis indicated four components that ex-plain the majority of the variability in the data set The loadof sodium and chloride had a considerable bearing on eitherfactor 1 or 2 depending on the altitude and location of theMS and the precipitation type These two factors explainedat least 29 of the variability in the system (Table 2)

Sodium and chloride explained the greatest variability inthe systemrsquos rain samples except for TS1 (2660 m)

At Cerro del Consuelo sodium chloride and potassiumdominated the variability in rain given that they loaded tofactor 1 In OP samples biomass-burning compounds suchas nitrate sulfate and ammonium had a stronger signal load-ing to factor 1 Factor 2 was loaded by sodium and chlorideonly No other compounds loaded to this factor meaning thatsodium and chloride most likely originated in sea salt fromAtlantic and Pacific air masses

At El Tiro sea salt sources were exclusively present infactor 1 for rain and factor 2 for OP similar to Cerro delConsuelo As for Cerro del Consuelo sea salt explainedmost of the variance in rain followed by biomass-burningcompounds The opposite was true for OP where biomass-burning compounds dominated followed by sea salt

The situation at TS1 was more complex than at Cerro delConsuelo and El Tiro given the combined influence of the lo-cal mountain-valley breeze system and the synoptic systemIn rain biomass-burning compounds dominated the variabil-ity with significant loadings on factor 1 Factor 2 was onlyloaded with sodium and chloride In OP factor 1 representedsea salt and crustal material as sodium chloride and calciumloaded to this factor

In the rain samples at ECSF sodium and chloride loaded tofactor 1 together with K+ In the OP samples nitrate SO4

minusand K+ dominated the variability loading to factor 1 whilesodium and chloride loaded to factor 2

Na rain+

Cl rain-

Na OP+

Cl OP-

Jan 2007Jan 2006 Jan 2008Jan 2005Jan 2004 Jan 2009

Figure 2 Time series of Na+ and Clminus VWMM concentration in rain and OP These samples come from MSs at different altitudes andtopographical locations (a) Cerro del Consuelo (3180 m) (b) El Tiro (2825 m) (c) TS1 (2660 m) and (d) ECSF (1960 m) The shadedareas cover 6-month periods from September to February The boxplots in the right column show the distribution of each time series boxessymbolize the lower and upper quartile of the data vertical lines show ranges of observed concentration and points are outliers

Figure 3a shows the Na+ Clminus molar ratio calculatedin data from OP and rain water samples collected at eachMS along the altitudinal gradient studied The typical mo-lar Na+ Clminus ratio in precipitation for areas close to the seais 086 (dotted line in the figure) according to Keene et al(1986) The International Co-operative Programme on As-sessment and Monitoring of Air Pollution Effects on Forests(ICP) recommends an acceptable range of values between05 and 15 molar units (Clarke et al 2010) (dashed line inFig 3) Outliers were likely due to samples with concentra-tions too close to the detection limits When approaching thelower limits of the concentration the ratio becomes more un-stable and tends towards more extreme values

The highest stations show a stronger marine influence par-ticularly in OP The ratio fluctuates within the acceptablerange and is mostly close to the standard value of sea saltin precipitation from coastal areas This influence diminishes

as the altitude decreases especially for OP Median ratios of07 08 05 and 03 for Cerro del Consuelo El Tiro TS1and ECSF respectively also reflect a greater influence of seasalt at higher altitudes

A somewhat stronger seasonal behaviour was identifiedat the two highest stations (grey columns show the periodfrom September to February with the highest frequency ofintrusion by the easterlies) Figure 3b depicts the frequencyof trajectories on a yearly basis In the first 3 years (2004ndash2006) the occurrence of westerlies was more frequent and atthe same time the Na+ Clminus ratio was close to that of freshsea water (local Pacific influence) During 2007ndash2009 wheneasterlies were more frequent the Na+ Clminus ratio increaseddue to the increasing influence of distant Atlantic sources(chloride is depleted during transport) and the likely con-tribution of forest and agricultural fires (Reid et al 2004Akagi et al 2011)

Table 2 Loadings from PFA with varimax rotation of major ions in rain and OP samples from Cerro del Consuelo El Tiro TS1 and ECSFmeteorological stations

Rain OP

Factor 1 Factor 2 Factor 3 Factor 4 Factor 1 Factor 2 Factor 3 Factor 4

Cerro del Consuelo

NH4+ 005 014 081 000 088 009 015 008

Ca2+ 005 008 minus004 073 017 030 060 060Clminus 088 044 006 minus015 023 081 013 035Mg2+ 004 minus021 070 minus006 017 020 086 014NO3

minus 027 091 000 019 087 027 minus002 042K+ 085 006 minus002 028 084 024 025 009Na+ 088 037 019 minus000 019 083 024 005SO4

2minus 049 064 minus015 minus003 054 030 034 064

El Tiro

minus007 minus006 064 003 082 034 028 022Ca2+

minus002 011 003 040 034 029 068 009Clminus 098 009 minus011 minus011 031 090 025 001Mg2+ 005 minus004 087 minus003 048 038 052 038NO3

minusminus006 080 minus018 020 081 038 038 008

K+ 015 042 035 minus038 085 030 039 002Na+ 097 004 008 001 033 087 031 016SO4

2minus 013 084 minus000 008 061 032 063 minus002

minus004 005 059 027 003 012 009 068Ca2+

minus006 minus016 minus002 004 067 023 037 021Clminus 036 089 minus015 004 089 037 016 minus000Mg2+

minus022 minus009 078 004 045 minus015 055 046NO3

minus 091 014 minus020 027 022 089 000 018K+ 015 008 034 067 057 042 058 018Na+ -004 089 minus000 033 091 035 019 007SO4

NH4+ 001 078 011 minus010 034 042 039 043

Ca2+ 023 001 002 003 minus004 003 007 050Clminus 049 minus023 004 030 031 077 034 032Mg2+ -010 054 minus059 009 005 033 073 023NO3

minusminus000 022 075 031 080 007 001 001

K+ 065 minus002 minus010 003 052 037 064 004Na+ 080 002 015 022 013 079 025 minus002SO4

2minus 024 minus007 017 073 077 029 027 minus003

Locally driven winds such as thermally induced anabaticwinds can contribute to the transport of local sodium andchloride from the valley to the upper parts of the catchmentIn a previous study however Makowski Giannoni et al(2013) showed that anabatic winds do not impact MSs lo-cated on mountain tops where synoptic winds predominateFigure 4a and b show relative frequencies of the wind di-rection at ECSF and Cerro del Consuelo respectively Atthe lower altitudes (ECSF) a typical mountain-valley breeze

circulation system exists while at the crest (Cerro del Con-suelo) north-easterly wind directions predominated

42 Spatial allocation of sources and general transportpathways

In the previous section we analysed the temporal and alti-tudinal variation of sodium and chloride concentrations indeposition driven by rain and OP This section addresses the

051015

Cerro del Consuelo

051015

El Tiro

051015

04-01 04-07 05-01 05-07 06-01 06-07 07-01 07-07 08-01 08-07 09-01 09-07

051015

Rain NaCl OP NaCl(a)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

070(b)

Figure 3 (a) Time series of monthly Na+ Clminus molar ratios for thefour MSs along the altitudinal gradient Smooth lines are fitted assolid lines and the 95 confidence interval is shown by the shadedarea (b) Yearly CWT sea salt source maps for southern Ecuador

remaining question of where the sodium and chloride sourceareas are located geographically

The synoptic wind system over South America is drivenby strong seasonal circulation patterns Because the air masstransport to the receptor site is directly linked to the seasonalcycle of the large-scale circulation system (Bendix et al2008a Emck 2007) and thus sources of sea salt concentra-tion and their intensity may vary with the seasons we exam-ined seasonal patterns present in the sources and dominantair mass trajectory clusters

Figure 4 Monthly wind sector relative frequency () for meteoro-logical stations (a) ECSF and (b) Cerro del Consuelo

We first evaluated potential contributory sources to sea saltconcentration at the receptor site for each season For thispurpose the two hybrid receptor models were used as shownin Fig 5 In accordance with a sensitivity analysis done withback trajectories starting at different altitudes these func-tions were applied to 3180 m the altitude of the MS Cerro delConsuelo on top of the highest peak in the catchment Tra-jectories starting at lower altitudes have greater uncertaintybecause local flows are driven by the complex topographyand cannot be reproduced Those starting at higher altitudesprovided no further information as they have coincidentalsource areas

Figure 5 shows the spatial distribution of potential sourcescalculated by the PSCF (a b) and the CWT (c) for DJFMAM JJA and SON at 3180 m starting height at the recep-tor When we compare the spatial distribution of sources be-tween the two models (PSCF and CWT) similar locationsin the Atlantic and Pacific oceans are indicated The high-est likelihood (concentrations above 5eminus9 for Fig 5a and band above 5eminus9 for Fig 5c) is an equatorial Pacific locationwhich points to stronger sources of sea salt in that regioncontributing to the high concentration at the receptor site

Strong sources of sea salt are expected from either thePacific or the Atlantic To judge from the high probabilitythat the concentration stems from the oceans the results ofthe PSCF (90th percentile concentration threshold Fig 5b)and the CWT (Fig 5c) performed best in discriminating be-tween potential geographic sources that contributed to mod-

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Summer (DJF)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Summer (DJF)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

90e-09

80e-09

70e-09

60e-09

50e-09

40e-09

30e-09

20e-09

10e-09

00e+09

Figure 5 Seasonal sea salt source maps according to (a) PSCF withconcentration threshold at 75th percentile (b) PSCF with concen-tration threshold at 90th percentile and (c) CWT the back trajectorystarting height is 3180 m at the receptor

erate and high sea salt concentrations over southern EcuadorIn contrast when using the 75th percentile as the concen-tration threshold (Fig 5a) the PSCF only detected the trans-port pathways for sea salt irrespective of the intensity of thesource contribution to the concentration

Seasonal source contributions that had the greatest impactie responsible for high sea salt concentrations at the recep-

tor site occurred between September and February (Septem-ber October and November (SON) and December Januaryand February (DJF)) During SON the equatorial Pacific wasthe dominant source while in DJF both the Pacific and At-lantic sources contributed to the concentration Yet the Pa-cific sources still appeared stronger as indicated by the largenumber of high values in that area Furthermore chlorine-containing compounds related to sea salt and biomass burn-ing were likely co-linearly transported during that season asshown by the high concentration over the northern portionof South America during DJF and the coincidence of thebiomass-burning season in that area (Fig 9) On the otherhand during austral autumn (MAM) and winter (JJA) themodels identified no relevant potential sources of high seasalt concentration

After locating the potential geographic sources of sea salttrajectory cluster analysis was applied to identify the mainrepresentative air mass transport patterns and thus the trans-port pathways of sea salt (Fig 6) Here the easterlies weredominant In the air mass transport fast flowing east trajec-tories dominated (from approximately 83 to 97 of the tra-jectories in DJF and JJA respectively) and slower-movingtrajectories from the west appeared rather sporadically (be-tween approximately 28 and 17 ) The occurrence of bow-shaped trajectories was common (Fig 6 MAM and DJF)and characterized the coastal wind system associated withthe Humboldt current (Bendix et al 2008a Emck 2007)

Westerlies mostly evolved during SON and DJF and to alesser extent during MAM Meanwhile north easterlies wereabsent during the austral winter (JJA) following the displace-ment of the ITCZ to the north The eastern clusters exhib-ited no seasonal pattern because they represent the prevail-ing wind directions throughout the year

Table 3 summarizes the mean sea salt concentration oversouthern Ecuador related to each cluster object reaching thereceptor site for each season High concentrations of seasalt are associated with westerly and north-easterly trajec-tories mainly occurring between September and May (Ta-ble 3) whereas easterly air masses that passed over the At-lantic and continental South America before arriving at thereceptor site showed intermediate to lower concentrations Inaddition to the cluster-associated mean sea salt concentra-tion values in parenthesis in Table 3 describe the proportion(in percentage) that each cluster contributes to the total con-centration during the study period Cluster C2 representingthe north easterlies was associated with the highest contri-butions in DJF and MAM In SON cluster C2 representedthe easterlies and was likewise associated with the highestcontributions SON is the main biomass-burning season inthe Brazilian Amazon which likely contributed to the overallbudget Furthermore easterliesrsquo transport from September toMay was associated with approximately 75ndash80 of the totalconcentration The remaining 15ndash20 were attributed to airflows passing over the Pacific before reaching the receptorsite These highly loaded seasonal Pacific flows took place in

-100 -90 -80 -70 -60 -50 -40 -30

Spring (MAM)

-100 -90 -80 -70 -60 -50 -40 -30

Summer (JJA)

-100 -90 -80 -70 -60 -50 -40 -30

Autumn (SON)

-100 -90 -80 -70 -60 -50 -40 -30

Winter (DJF)

ClusterC1C2C3C4C5C6

223 371

231 225

85 186 309

135 52

Figure 6 Seasonal mean back-trajectory clusters (C1ndashC6)

Table 3 Mean sea salt concentration and percentage of total concentration for each season at the receptor site associated to the meantrajectory clusters (C1ndashC6) in the Andes of southern Ecuador The percentage contribution of the mean clusters to the total concentration isshown in parenthesis

Mean sea salt concentration (kg kgminus1) and relative total concentration ()

C1 C2 C3 C4 C5 C6

Summer (DJF) 512E-09(1442) 518E-09(4376) 381E-09(2299) 586E-09(684) 457E-09(57) 578E-09(628)Autumn (MAM) 190E-09(3034) 380E-09(3642) 135E-09(176) 283E-09(404) 385E-09(518) 382E-09(643)Winter (JJA) 172E-09(2128) 183E-09(2433) 164E-09(32) 153E-09(2083) 181E-09(2077) 152E-09(96)Spring (SON) 216E-09(2363) 233E-09(2836) 270E-09(1777) 653E-09(1956) 418E-09(707) 347E-09(361)

the Southern Hemispherersquos late spring and summer as east-erlies weaken due to the southward shift of the ITCZ At-lantic air masses contributed to the concentration constantlyover the year also in austral winter (JJA) when the Pacificflows were negligible However transport from the Pacificclearly dominated the high peaks at the end and the begin-ning of each year The sea salt concentration associated withthe easterly clusters was much weaker However due to itshigh frequency it persistently contributed to the transportfrom the Atlantic with likely additions from Amazon firesSimilar seasonal patterns were also identified in the measure-ments as illustrated in Fig 2 the most clearest of which oc-curred at the highest station Cerro del Consuelo (Fig 2a)That means that the observed patterns in the measurementscan be explained by the large-scale atmospheric circulationpatterns

Figure 7 depicts the sea salt concentration along the clus-ter mean trajectories and the trajectory height for each sea-son The north equatorial Atlantic was a great source of sea

salt during DJF (Fig 7 column 1) according to the sea saltconcentration along the trajectory clusters Nonetheless theconcentration rapidly decreased as the air masses travelledover the continent Compared to westerly air masses east-erly air masses were lower in elevation which increased theprobability that they were loaded with aerosols from surfaceemissions Those air masses then ascended abruptly as theyapproached the Andean range In comparison the equatorialPacific is a less significant source for sea salt Because of itsvicinity to the receptor site and because the air masses spentmost of the time over the ocean the concentration did notsink significantly over time (Fig 7 C4ndashC6) The concentra-tion peaks in C1 and C2 were due to sea salt intrusions fromthe Caribbean Sea and canalized by the Andean cordillera asdepicted in Fig 8a The season DJF was also characterizedby frequent forest and agricultural fires in northern SouthAmerica which likely contributed chlorinated compoundsfrom biomass burning to the budget as well (Fig 9)

DJ F MAM SONJ J A

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

Time step Time step Time stepTime step

Figure 7 Seasonal plots of sea salt concentration and pressure level along mean back-trajectory clusters The colours of the sea salt concen-tration lines match those in the trajectory clusters in Fig 6

20degS

20degN

20degS

20degN

20degS

20degN

20degS

20degN

80degW 60degW 40degW 80degW 60degW 40degW

015 075 135 195 255

kg kg-1 (10^8)

05 15 25 35 45

Relative Cl mass fraction(permil)

Figure 8 (a) Seasonal maps of sea salt concentration calculated from the MACC reanalysis model (b) Seasonal maps of biomass-burningCl mass fraction relative to sea salt Cl mass ratio to NOx was estimated based on emission factors from Ferek et al (1998)

Figure 9 Seasonal maps of NOx fluxes caused by biomass burning Values were calculated from the MACC reanalysis model

A similar situation also occurred in MAM (Fig 7 column2) where C1ndashC3 are north-easterly air masses and C4ndashC6represent westerly pathways In C2 the intrusion of sea saltfrom the Caribbean was also present but less pronouncedthan in DJF (Fig 8)

For SON (Fig 7 column 3) most of the budget was trans-ported from the Atlantic and the continent clusters C1 C2and C3 Because this period coincided with the Brazilianbiomass-burning season (see Fig 9) a considerable quantityof sodium and chloride from the Atlantic (Fig 8a) and fromfire emissions were probably transported to the receptor site

5 Discussion

In this study we examined potential sources of sodium andchloride for the southern escarpment of the Ecuadorian An-des The investigation analysed chemical ion concentrationsin rain and OP samples along an altitudinal gradient usingback-trajectory statistical analysis and sourcendashreceptor mod-elling

We first explored the distribution of sodium and chlorideinputs by rain and OP in relation to altitude Overall com-parisons between the MSs reveal a difference in the temporalvariation of the concentration of sodium and chloride in rainand OP depending on elevation and exposure The highestMS Cerro del Consuelo displays a distinct seasonal pattern

which is otherwise lacking or less pronounced at the remain-ing MSs The largest sodium and chloride concentration oc-curred between September and February concomitantly withthe southward migration of the ITCZ and the more frequentnorth easterlies (see Fig 2)

Chloride was consistently a larger portion of the concen-tration than sodium which agrees with findings by Tardyet al (2005) Their study investigated the chemical compo-sition of rainwater in the Amazon and found that in placescloser to the Atlantic Ocean chloride concentrations werehigher than sodium whereas this ratio inverted for locationsfurther away from the Atlantic coast The observed excess inchloride concentration compared to sodium at our study sitemeans that marine sources have a significant impact on theoverall sea salt presence This influence is also demonstratedby the Na+ Clminus molar ratios in samples from the MSs alongthe altitudinal gradient Higher altitudes are exposed to syn-optic circulation and stations there register Na+ Clminus ratioscloser to that in marine air masses This indicates a gradientof sea salt inputs with relation to terrain height

Common transport histories were identified on the basisof the PFA As evinced by their prevalence in the first twocomponents sodium and chloride concentration are very rel-evant in both rain and OP The fact that they exclusively loadto factor 1 or 2 based on their location in multidimensionalspace suggests the likely origin of sodium and chloride is sea

salt from either the Pacific or the Atlantic Ocean Biomass-burning seems to play a minor role in the transport of sodiumand chloride since nitrate and sulfate did not load to thesame factor Although emissions of chloride from fires maybe recognizable (Akagi et al 2011) they are likely irrelevantcompared to sea salt (Fabian et al 2009) Calcium and mag-nesium (crustal material) loading to factor 2 in OP at TS1(Table 2) reveals the effect local winds have on emissions atlower elevations The influence of chloride-containing dustblown from the LojandashZamora road (Fig 1) and plumes fromlocal biomass-burning fires are the most likely causes of thehigh chloride concentrations at ECSF Findings by Yokouchiet al (2002) Spanos (2002) Hildemann et al (1991) Akagiet al (2011) and Harrison and Pio (1983) substantiate thisassumption According to these studies chloride not onlystems from sea spray but is also emitted from natural andanthropogenic terrestrial sources ie dust biomass-burningand biogenic forest emissions

Concluding from the ion concentration time series (Fig 2)and PFA (Table 2) the distribution of chloride and its sourcescan differ between valleys and mountain tops In the for-mer the concentration is influenced by local winds and emis-sions while in the latter the concentration most likely de-pends on synoptic air-mass transport and emissions from dis-tant sources This difference in atmospheric circulation be-tween valleys and mountain tops is clearly depicted in Fig 4in which only the mountain tops are strongly influenced bythe synoptic circulation

The potential sources defined by PSCF and CWT (Fig 5)as well as the cluster-concentration statistics (Table 3) con-cur with the occurrence of the highest sodium and chlorideconcentration between September and February at the mostelevated MSs ndash El Tiro and Cerro del Consuelo ndash in the studyarea (Fig 2a b and Table 3) This corroborates the conclu-sions reached in previous paragraphs that the transport athigher elevations is more synoptically driven Thus medium-to long-range transport (reproduced by back-trajectory mod-elling) has more of an effect in areas of high elevation thanat lower slopes and valleys which are more affected by localtransport

The results of the PSCF and CWT sourcendashreceptor models(Fig 5a b c) indicate the areas that contribute to the high-est concentration at the receptor site ie the equatorial Pa-cific in the vicinity of the coast of Ecuador and northern Peruas well as the north equatorial Atlantic and the CaribbeanSea Nonetheless according to the spatial distribution of thesea salt concentration illustrated in Fig 8a the PSCF andCWT models seem to overestimate the contribution of theequatorial Pacific which exhibits a lower sea salt concentra-tion than the Atlantic Ocean Analysis of the sea salt con-centration along the trajectory clusters reveal a comparablebehaviour wherein the clusters passing over the Pacific con-tain a lower sea salt concentration However the concentra-tion remains quite stable contrary to the easterly trajectoriespassing over the continent where wet scavenging is much

more pronounced Those drier conditions over the equato-rial Pacific were clearly seen in DJF where the concentra-tion among clusters C4 C5 and C6 even increase as the airmasses approach the receptor site (Fig 7)

The CWT (Fig 5c) model delivered the best results in thatit successfully differentiates the source hot spots over theoceans from those areas of moderate contribution over thecontinent In contrast the PSCF (Fig 5a b) is less successfulin making this distinction As already reported by Hsu et al(2003) and Stohl (1996) a drawback of the PSCF methodis that the high and extreme values above a defined thresh-old get similar probabilities which hamper their distinctionThus PSCF results are heavily influenced by the choice ofan arbitrary threshold concentration Pekney et al (2006) re-ported that the selection of the threshold value relies on theevaluated concentration time series The authors found thatfor low background values and high concentration peaks the90th percentile threshold performs better while the 75th per-centile is more appropriate for concentration time series withless variability In our case the quite strong seasonal vari-ations in the sea salt concentration explain why the PSCFperformed better with the 90th percentile threshold (Fig 5b)rather than the 75th percentile (Fig 5a)

To perform a cluster analysis the trajectories weregrouped together and six dominant pathways were identi-fied (C1ndashC6 in Fig 6) In general over the entire observa-tion period the eastern clusters originating on the equato-rial and south equatorial Atlantic predominate (gt 51 ofthe trajectories) However when seasonally linking the maintransport pathways (C1ndashC6) to the sea salt concentration atthe receptor site (see Table 3) we notice that those path-ways do not have the highest impact on the sea salt con-centration in southern Ecuador Even if easterly and south-easterly transport prevail larger sea salt loads are trans-ported from the north equatorial Atlantic the Caribbean Seaand the equatorial Pacific The north easterlies originatingfrom the north equatorial Atlantic and Caribbean Sea oc-curred approximately 295 of the time and accounted foraround 565 of the concentration over southern EcuadorThe westerlies from the equatorial Pacific were much lessfrequent (asymp 93 ) and accounted for 26 of the concen-tration That means despite the barrier effect of the Andesand the low frequency of occurrences of western pathwaysPacific sea salt sources still play a relevant role in transport-ing sea salt to the receptor site Together equatorial Pacificand north equatorial Atlantic sources accounted for around824 of the total sea salt concentration Furthermore largequantities were added solely from the equatorial Pacific andthe Caribbean Sea in a short period of time (asymp 16 of trajec-tories) contributing up to 467 of the total concentrationwhich stress the importance of these sources to the atmo-spheric sea salt budget over southern Ecuador

Nevertheless in light of the sea salt concentration alongthe seasonal trajectory clusters (Fig 7) and sea saltrsquos spatialdistribution (Fig 8a) the significance of Pacific sea salt re-

Table 4 Comparison of the mean concentration of Na+ and Clminus in precipitation in this study with data from other sites in the Amazon basinThe values represent volume-weighted means expressed in micro eq Lminus1

Na+ Clminus Reference

South Ecuador (RBSF) 780 960 This studyCentral Amazon (Lake Calado) 240 460 Williams et al (1997)

Central Amazon (Balbina) 380 520 Pauliquevis et al (2012)Northeast Amazon 1660 1690 Forti et al (2000)

Eastern Amazon (Belem) 1890 1950 Mortatti (1995)

mains questionable The concentration of sea salt in the equa-torial Pacific is less than that in the Atlantic Therefore theformerrsquos influence may be overestimated even if the con-centration decay over the Pacific is much less pronounced asover continental South America On the one hand the greaterfrequency of the north easterlies and on the other hand thehigher sea salt concentration in the Atlantic are good reasonsthat justify a greater influence of the Caribbean Sea and thenorth equatorial Atlantic with respect to the atmospheric seasalt budget over southern Ecuador

Regarding the addition of salt from biomass burning to thechloride budget at the study area based on the co-occurrenceof high NOx concentration from biomass burning and highsea salt concentration during the main sea salt transport sea-son (SON and DJF in Figs 8a and 9) it is very likelythat sea salt is indeed enriched by biomass-burning chlo-ride However this assumption is not corroborated by ourfield samples from southern Ecuador In the concentration ofsodium and chloride from rain and OP samples we did notfind any correlation between nitrate and sulfate the productsof biomass burning (Fabian et al 2009 Makowski Gian-noni et al 2013 2014) and sodium and chloride (Table 2)Furthermore the mass fraction of biomass-burning sodium(maximum 007 ) and chloride (maximum values in the or-der of 35 see Fig 8b) are very small and therefore playonly a limited role in the sodium and chloride transport tosouth Ecuador

6 Conclusions

Sodium and chloride ions exhibited different concentrationsin rain and OP along the altitudinal gradient of interest tothis study Their concentration levels and temporal variabil-ity in the highest and more exposed MSs presented a strongerseasonality linked to global circulation patterns and thus agreater influence from sea salt confirmed by Na+ Clminus mo-lar ratios similar to those from marine air masses Similarseasonal patterns were observed by modelling at a largerscale using MACC sea salt concentration data and ERA-Interim air mass back trajectories confirming the influenceof the medium- to long-range transport at higher elevationsIn contrast MSs situated at lower altitudes were influencedby the mountain-valley wind systems and local aerosols

According the sea salt transport analysis by back-trajectory modelling for medium- to long-range sources theCaribbean Sea the north equatorial Atlantic and equatorialPacific play an important role in the transport of sea salt tosouthern Ecuador Here the Caribbean and north equatorialAtlantic sources have the greatest impact Equatorial Pacificsources on the other hand are less significant they are sea-sonally driven with the greatest contributions occurring whenthe ITCZ migrates further south in austral late spring (SON)and summer (DJF) In total the north equatorial Atlantic andequatorial Pacific contribute to 565 and 26 of the totalconcentration in southern Ecuador respectively which rep-resents an important addition to the total atmospheric sea saltbudget

A comparison of the sodium and chloride concentrationsat our area of investigation with those at other sites furthereast substantiates the important role played by the identifiedsources (Caribbean Sea north equatorial Atlantic and equa-torial Pacific oceans) on the sea salt transport to our studyarea (Table 4) Even if concentrations in southern Ecuadorare lower than in forests close to the Atlantic they clearly ex-ceed those concentrations measured in the central BrazilianAmazon thousands of kilometres to the east despite being lo-cated further from the Atlantic coast However whether thehigher sodium and chloride availability observed in south-ern Ecuador makes this tropical ecosystem less salt deprivedthan other similar ecosystems in the western Amazon is stillan open question deserving of investigation

7 Data availability

The MACC sea salt and biomass-burning NOx fluxesare available at httpappsecmwfintdatasetsdatamacc-reanalysislevtype=ml and httpappsecmwfintdatasetsdatacams-gfas respectively The ERA Interimreanalysis data can be accessed at httpappsecmwfintdatasetsdatainterim-full-dailylevtype=sfc The pre-cipitation chemistry data and climate data from in-situmeteorological stations are accessible at the Platform forBiodiversity and Ecosystem Monitoring and Research inSouth Ecuador httpwwwtropicalmountainforestorgdata_predocmd=showall

Appendix A Tables

Table A1 The 6-year average ion concentration (2004ndash2009) in rain and OP precipitation volume and electrical conductivity at MSs alongan altitudinal gradient in southern Ecuador

Site Collector Elev P pH eC NH4minus Ca+ Clminus PO4

minus Mg+ NO3minus K+ Na+ SO4

(m) (mm) (microS cmminus1) (mg Lminus1)

C del Consuelo OP 3180 10559 54 12 055 017 042 0085 0059 082 015 022 1C del Consuelo Rain 3180 47265 53 35 018 01 029 016 005 011 0098 014 026ECSF OP 1960 726 5 13 017 042 086 0098 0073 014 023 022 044ECSF Rain 1960 14248 53 43 015 011 043 013 0048 0077 016 018 024El Tiro OP 2825 7541 6 20 08 019 045 012 007 14 03 031 16El Tiro Rain 2825 14237 54 54 022 014 036 015 0047 012 015 024 039TS1 OP 2660 3415 53 52 025 012 05 016 0054 016 012 025 032TS1 Rain 2660 279 54 4 018 0077 036 013 0047 0084 013 018 026

Acknowledgements We thank the German Academic ExchangeService (DAAD) for funding the PhD thesis of S Makowski Gian-noni (ref no A0898222) and the German Research Foundation(DFG) for funding this work in the scope of the Research UnitRU816 (funding no BE 178015-1) We also thank Thorsten Petersfor providing meteorological data We are grateful to Giulia FCuratola Fernaacutendez and Tim Appelhans for their valuable help Wealso thank the foundation Nature amp Culture International (NCI)Loja and San Diego for logistic support Last but not least weare very thankful to Jeffrey Reid from the US Naval ResearchLaboratory for his valuable comments on the manuscript

Edited by R KrejciReviewed by J Reid and two anonymous referees

References

Akagi S K Yokelson R J Wiedinmyer C Alvarado M JReid J S Karl T Crounse J D and Wennberg P O Emis-sion factors for open and domestic biomass burning for usein atmospheric models Atmos Chem Phys 11 4039ndash4072doi105194acp-11-4039-2011 2011

Beck E Bendix J Kottke I Makeschin F and Mosandl R(Eds) Gradients in a tropical mountain ecosystem of Ecuadorvol 198 of Ecological Studies Springer BerlinHeidelbergBerlin Germany doi101007978-3-540-73526-7 2008

Bendix J Rollenbeck R and Reudenbach C Diurnal patternsof rainfall in a tropical Andean valley of southeastern Ecuadoras seen by a vertically pointing Kndashband Doppler radar Int JClimatol 26 829ndash846 doi101002joc1267 2006

Bendix J Rollenbeck R Goettlicher D Nauss T and FabianP Seasonality and diurnal pattern of very low clouds in a deeplyincised valley of the eastern tropical Andes (South Ecuador) asobserved by a costndasheffective WebCam system Meteorol Appl15 281ndash291 doi101002met 2008a

Bendix J Rollenbeck R Richter M Fabian P and Emck PClimate in Gradients in a Tropical Mountain Ecosystem ofEcuador edited by Beck E Bendix J Kottke I MakeschinF and Mosandl R chap 1 Springer BerlinHeidelberg ecolog-ical Edn 63ndash74 doi101007978-3-540-73526-7_8 2008b

Bendix J Beck E Brauning A Makeschin F Mosandl RScheu S and Wilcke W (Eds) Ecosystem services biodiver-sity and environmental change in a tropical mountain ecosys-tem of South Ecuador vol xxx Springer Berlin-Heidelbergdoi101007978-3-642-38137-9 2013

Benedetti a Morcrette J-J Boucher O Dethof A EngelenR J Fisher M Flentje H Huneeus N Jones L KaiserJ W Kinne S Mangold a Razinger M Simmons A Jand Suttie M Aerosol analysis and forecast in the EuropeanCentre for Medium-Range Weather Forecasts Integrated ForecastSystem 2 Data assimilation J Geophys Res 114 D13205doi1010292008JD011115 2009

Bobbink R Hicks K and Galloway J Global assessment of ni-trogen deposition effects on terrestrial plant diversity a synthe-sis Ecol Appl 20 30ndash59 doi10189008-11401 2010

Carslaw D C and Ropkins K openair ndash An R package for airquality data analysis Environ Modell Softw 27-28 52ndash61doi101016jenvsoft201109008 2012

Clarke N Zlindra D Ulrich E Mosello R Derome JDerome K Koumlnig N Loumlvblad G Draaijers G HansenK Thimonier A and Waldner P sampling and analysisof deposition Part XIV in Manual on methods and criteriafor harmonized sampling assessment monitoring and analysisof the effects of air pollution on forests chap 14 UNECEICP Forests Hamburg available at httpicp-forestsnetpageicp-forests-manual (last access 8 August 2016) 2010

Dee D P Uppala S M Simmons A J Berrisford P PoliP Kobayashi S Andrae U Balmaseda M A Balsamo GBauer P Bechtold P Beljaars A C M van de Berg L Bid-lot J Bormann N Delsol C Dragani R Fuentes M GeerA J Haimberger L Healy S B Hersbach H Hoacutelm E VIsaksen L Karingllberg P Koumlhler M Matricardi M McNallyA P Monge-Sanz B M Morcrette J-J Park B-K PeubeyC de Rosnay P Tavolato C Theacutepaut J-N and Vitart F TheERA-Interim reanalysis configuration and performance of thedata assimilation system Q J Roy Meteor Soc 137 553ndash597doi101002qj828 2011

Dentener F Drevet J Lamarque J F Bey I Eickhout BFiore a M Hauglustaine D Horowitz L W Krol M Kul-shrestha U C Lawrence M Galy-Lacaux C Rast S Shin-dell D Stevenson D Van Noije T Atherton C Bell NBergman D Butler T Cofala J Collins B Doherty REllingsen K Galloway J Gauss M Montanaro V MuumlllerJ F Pitari G Rodriguez J Sanderson M Solmon F Stra-han S Schultz M Sudo K Szopa S and Wild O Ni-trogen and sulfur deposition on regional and global scales Amultimodel evaluation Global Biogeochem Cy 20 GB4003doi1010292005GB002672 2006

Draxler R and Hess G An overview of the HYSPLIT_4 modelingsystem of trajectories dispersion and deposition Aust Meteo-rol Mag 47 295ndash308 1998

Dudley R Kaspari M and Yanoviak S P Lust for Salt inthe Western Amazon Biotropica 44 6ndash9 doi101111j1744-7429201100818x 2012

Emck P A climatology of South Ecuador Phd thesis Univer-sity of Erlangen available at httpwwwopusubuni-erlangendeopusvolltexte2007656 (last access 8 August 2016) 2007

Fabian P Rollenbeck R Spichtinger N Brothers LDominguez G and Thiemens M Sahara dust ocean sprayvolcanoes biomass burning pathways of nutrients into Andeanrainforests Adv Geosci 22 85ndash94 doi105194adgeo-22-85-2009 2009

Ferek R J Reid J S Hobbs P V Blake D R and LiousseC Emission factors of hydrocarbons halocarbons trace gasesand particles from biomass burning in Brazil J Geophys Res-Atmos 103 32107ndash32118 doi10102998JD00692 1998

Fleming Z L Monks P S and Manning A J Review Un-tangling the influence of air-mass history in interpreting ob-served atmospheric composition Atmos Res 104-105 1ndash39doi101016jatmosres201109009 2012

Forti M C Melfi A J Astolfo R and Fostier A-H Rain-fall chemistry composition in two ecosystems in the northeasternBrazilian Amazon (Amapaacute State) J Geophys Res 105 28895doi1010292000JD900235 2000

Galloway J N Townsend A R Erisman J W Bekunda M CaiZ Freney J R Martinelli L A Seitzinger S P and SuttonM A Transformation of the nitrogen cycle recent trends ques-

tions and potential solutions Science (New York NY) 320889ndash892 doi101126science1136674 2008

Griffith K T Ponette-Gonzaacutelez A G Curran L M and Weath-ers K C Assessing the influence of topography and canopystructure on Douglas fir throughfall with LiDAR and empiricaldata in the Santa Cruz mountains USA Environ Monitor As-sess 187 4486 doi101007s10661-015-4486-6 2015

Guelle W Schulz M Balkanski Y and Dentener F Influenceof the source formulation on modeling the atmospheric globaldistribution of sea salt aerosol J Geophys Res 106 27509doi1010292001JD900249 2001

Harrison R M and Pio C A Major ion composition and chemi-cal associations of inorganic atmospheric aerosols Environ SciTechnol 17 169ndash174 doi101021es00109a009 1983

Hildemann L M Markowski G R and Cass G RChemical composition of emissions from urban sources offine organic aerosol Environ Sci Technol 25 744ndash759doi101021es00016a021 1991

Homeier J Hertel D Camenzind T Cumbicus N L Ma-raun M Martinson G O Poma L N Rillig M C Sand-mann D Scheu S Veldkamp E Wilcke W Wullaert Hand Leuschner C Tropical andean forests are highly suscepti-ble to nutrient inputs-rapid effects of experimental N and p ad-dition to an ecuadorian montane forest PloS one 7 e47128doi101371journalpone0047128 2012

Hsu Y-K Holsen T M and Hopke P K Comparison of hybridreceptor models to locate PCB sources in Chicago Atmos Env-iron 37 545ndash562 doi101016S1352-2310(02)00886-5 2003

Inness A Baier F Benedetti A Bouarar I Chabrillat S ClarkH Clerbaux C Coheur P Engelen R J Errera Q Flem-ming J George M Granier C Hadji-Lazaro J HuijnenV Hurtmans D Jones L Kaiser J W Kapsomenakis JLefever K Leitatildeo J Razinger M Richter A Schultz M GSimmons A J Suttie M Stein O Theacutepaut J-N Thouret VVrekoussis M Zerefos C and the MACC team The MACCreanalysis an 8 yr data set of atmospheric composition AtmosChem Phys 13 4073ndash4109 doi105194acp-13-4073-20132013

Jaegleacute L Quinn P K Bates T S Alexander B and Lin J-TGlobal distribution of sea salt aerosols new constraints from insitu and remote sensing observations Atmos Chem Phys 113137ndash3157 doi105194acp-11-3137-2011 2011

Kaiser J W Heil A Andreae M O Benedetti A ChubarovaN Jones L Morcrette J-J Razinger M Schultz M GSuttie M and van der Werf G R Biomass burning emis-sions estimated with a global fire assimilation system basedon observed fire radiative power Biogeosciences 9 527ndash554doi105194bg-9-527-2012 2012

Kaspari M Yanoviak S P and Dudley R On the biogeographyof salt limitation a study of ant communities P Natl Acad SciUSA 105 17848ndash17851 doi101073pnas0804528105 2008

Kaspari M Yanoviak S P Dudley R Yuan M and Clay N ASodium shortage as a constraint on the carbon cycle in an inlandtropical rainforest P Natl Acad Sci USA 106 19405ndash19409doi101073pnas0906448106 2009

Keene W C Pszenny A A P Galloway J N and Haw-ley M E Sea-salt corrections and interpretation of constituentratios in marine precipitation J Geophys Res 91 6647doi101029JD091iD06p06647 1986

Kirchner M Fegg W Roumlmmelt H Leuchner M Ries L Zim-mermann R Michalke B Wallasch M Maguhn J Faus-Kessler T and Jakobi G Nitrogen deposition along differentlyexposed slopes in the Bavarian Alps Sci Total Environ 470ndash471 895ndash906 doi101016jscitotenv201310036 2014

Koehler B Corre M D Veldkamp E Wullaert H andWright S J Immediate and long-term nitrogen oxide emis-sions from tropical forest soils exposed to elevated nitrogen in-put Glob Change Biol 15 2049ndash2066 doi101111j1365-2486200801826x 2009

Lee A T K Kumar S Brightsmith D J and Marsden S JParrot claylick distribution in South America do patterns ofldquowhererdquo help answer the question ldquowhyrdquo Ecography 33 503ndash513 doi101111j1600-0587200905878x 2009

Lizcano D J and Cavelier J Caracteriacutesticas Quiacutemicas de sala-dos y haacutebitos alimenticios de la Danta de montantildea (Tapirus pin-chaque Roulin 1829) en los Andes Centrales de Colombia Mas-tozoologiacutea neotropical 11 193ndash201 2004

Lovett G and Kinsman J Atmospheric pollutant deposition tohigh-elevation ecosystems Atmos Environ 24 2767ndash27861990

Mahowald N M Artaxo P Baker A R Jickells T D OkinG S Randerson J T and Townsend A R Impacts of biomassburning emissions and land use change on Amazonian atmo-spheric phosphorus cycling and deposition Global BiogeochemCy 19 doi1010292005GB002541 2005

Makowski Giannoni S Rollenbeck R Fabian P and Bendix JComplex topography influences atmospheric nitrate depositionin a neotropical mountain rainforest Atmos Environ 79 385ndash394 doi101016jatmosenv201306023 2013

Makowski Giannoni S Rollenbeck R Trachte K and BendixJ Natural or anthropogenic On the origin of atmospheric sul-fate deposition in the Andes of southeastern Ecuador AtmosChem Phys 14 11297ndash11312 doi105194acp-14-11297-2014 2014

Malm W C Johnson C E and Bresch J F Application ofprincipal component analysis for purposes of identifying source-receptor relationships in Receptor methods for source appor-tionment edited by Pace T G Air pollution control associa-tion Pittsburgh PA 1986

Matson A L Corre M D and Veldkamp E Nitrogen cyclingin canopy soils of tropical montane forests responds rapidly toindirect N and P fertilization Glob Change Biol 20 3802ndash3813 doi101111gcb12668 2014

Matson P Lohse K A and Hall S J The Globalization of Nitro-gen Deposition Consequences for Terrestrial Ecosystems AM-BIO A J Hum Environ 31 113ndash119 doi1015790044-7447-312113 2002

Millero F Treatise on Geochemistry Elsevier doi101016B978-0-08-095975-700601-X 2014

Morcrette J-J Boucher O Jones L Salmond D Bechtold PBeljaars A Benedetti A Bonet A Kaiser J W RazingerM Schulz M Serrar S Simmons A J Sofiev M Suttie MTompkins A M and Untch A Aerosol analysis and forecastin the European Centre for Medium-Range Weather Forecasts In-tegrated Forecast System Forward modeling J Geophys Res114 D06206 doi1010292008JD011235 2009

Mortatti J Erosatildeo na Amazonia processos Modelos e balanccediloPhd thesis Satildeo Paulo Brazil 1995

Pauliquevis T Lara L L Antunes M L and Artaxo P Aerosoland precipitation chemistry measurements in a remote site inCentral Amazonia the role of biogenic contribution AtmosChem Phys 12 4987ndash5015 doi105194acp-12-4987-20122012

Pekney N J Davidson C I Zhou L and Hopke P KApplication of PSCF and CPF to PMF-Modeled Sources ofPM 25 in Pittsburgh Aerosol Sci Technol 40 952ndash961doi10108002786820500543324 2006

Pentildeuelas J Poulter B Sardans J Ciais P van der VeldeM Bopp L Boucher O Godderis Y Hinsinger P LlusiaJ Nardin E Vicca S Obersteiner M and Janssens I AHuman-induced nitrogen-phosphorus imbalances alter naturaland managed ecosystems across the globe Nat Commun 42934 doi101038ncomms3934 2013

Pett-Ridge J C Contributions of dust to phosphorus cycling intropical forests of the Luquillo Mountains Puerto Rico Biogeo-chemistry 94 63ndash80 doi101007s10533-009-9308-x 2009

Phoenix G K Hicks K W Cinderby S Kuylenstierna J C IStock W D Dentener F J Giller K E Austin A T LefroyR D B Gimeno B S Ashmore M R and Ineson P Atmo-spheric nitrogen deposition in world biodiversity hotspots theneed for a greater global perspective in assessing N depositionimpacts Glob Change Biol 12 470ndash476 doi101111j1365-2486200601104x 2006

Powell L L Powell T U Powell G V N and Bright-smith D J Parrots Take it with a Grain of Salt AvailableSodium Content May Drive Collpa (Clay Lick) Selection inSoutheastern Peru Biotropica 41 279ndash282 doi101111j1744-7429200900514x 2009

Pozzer A de Meij A Pringle K J Tost H Doering U M vanAardenne J and Lelieveld J Distributions and regional bud-gets of aerosols and their precursors simulated with the EMACchemistry-climate model Atmos Chem Phys 12 961ndash987doi105194acp-12-961-2012 2012

Reid J S Prins E M Westphal D L Schmidt C C Richard-son K A Christopher S A Eck T F Reid E A Cur-tis C A and Hoffman J P Real-time monitoring of SouthAmerican smoke particle emissions and transport using a cou-pled remote sensingbox-model approach Geophys Res Lett31 L06107 doi1010292003GL018845 2004

Richter M Beck E Rollenbeck R and Bendix J The studyarea in Ecosystemservices biodiversity and environmentalchange in a tropical mountain ecosystem of South Ecuadoredited by Bendix J Beck E Brauning A Makeschin FMosandl R Scheu S and Wilcke W chap 1 SpringerBerlinHeidelberg Berlin Germany doi101007978-3-642-38137-9_1 2013

Riuttanen L Hulkkonen M Dal Maso M Junninen H andKulmala M Trajectory analysis of atmospheric transport offine particles SO2 NOx and O3 to the SMEAR II station inFinland in 1996ndash2008 Atmos Chem Phys 13 2153ndash2164doi105194acp-13-2153-2013 2013

Robinson N H Newton H M Allan J D Irwin M HamiltonJ F Flynn M Bower K N Williams P I Mills G ReevesC E McFiggans G and Coe H Source attribution of Borneanair masses by back trajectory analysis during the OP3 projectAtmos Chem Phys 11 9605ndash9630 doi105194acp-11-9605-2011 2011

Rollenbeck R Bendix J Fabian P Boy J Wilcke W DalitzH Oesker M and Emck P Comparison of different tech-niques for the measurement of precipitation in tropical mon-tane rain forest regions J Atmos Ocean Tech 24 156ndash168doi101175JTECH19701 2007

Rollenbeck R Bendix J and Fabian P Spatial and tempo-ral dynamics of atmospheric water inputs in tropical moun-tain forests of South Ecuador Hydrol Process 25 344ndash352doi101002hyp7799 2011

Schemenauer R S and Cereceda P A proposed standard fog col-lector for use in high-elevation regions J Appl Meteorol 331313ndash1322 1994

Schulz M de Leeuw G and Balkanski Y Sea-salt aerosol sourcefunctions and emissions in Emissions of Atmospheric TraceCompounds Springer Netherlands 333ndash359 doi101007978-1-4020-2167-1_9 2004

Seibert P Kromb-Kolb H Baltensperger U Jost D andSchwikowski M Trajectory analysis of high-alpine air pollu-tion data in NATO challenges of modern society edited byGryning S-E and Millaacuten M M Springer US 595ndash596doi101007978-1-4615-1817-4_65 1994

Spanos T Environmetric modeling of emission sources for dryand wet precipitation from an urban area Talanta 58 367ndash375doi101016S0039-9140(02)00285-0 2002

Stohl A Trajectory statistics-A new method to establish source-receptor relationships of air pollutants and its application to thetransport of particulate sulfate in Europe Atmos Environ 30579ndash587 doi1010161352-2310(95)00314-2 1996

Talbot R W Andreae M O Berresheim H Artaxo P GarstangM Harriss R C Beecher K M and Li S M Aerosol Chem-istry During the Wet Season in Central Amazonia The Influenceof Long-Range Transport J Geophys Res 95 16955ndash169691990

Tanner E V J Vitousek P M and Cuevas E Experimen-tal investigation of nutrient limitation of forest growth onwet tropical mountains Ecology 79 10ndash22 doi1018900012-9658(1998)079[0010EIONLO]20CO2 1998

Tardy Y Bustillo V Roquin C Mortatti J and Victo-ria R The Amazon Bio-geochemistry applied to riverbasin management Appl Geochem 20 1746ndash1829doi101016japgeochem200506001 2005

Tipping E Benham S Boyle J F Crow P Davies J FischerU Guyatt H Helliwell R Jackson-Blake L Lawlor A JMonteith D T Rowe E C and Toberman H Atmosphericdeposition of phosphorus to land and freshwater EnvironSci Process Impacts 16 1608ndash1617 doi101039c3em00641g2014

Vet R Artz R S Carou S Shaw M Ro C-U Aas W BakerA Bowersox V C Dentener F Galy-Lacaux C Hou APienaar J J Gillett R Forti M C Gromov S Hara HKhodzher T Mahowald N M Nickovic S Rao P and ReidN W A global assessment of precipitation chemistry and depo-sition of sulfur nitrogen sea salt base cations organic acidsacidity and pH and phosphorus Atmos Environ 93 3ndash100doi101016jatmosenv201310060 2014

Vitousek P M Litterfall Nutrient Cycling and Nutrient Limita-tion in Tropical Forests 65 285ndash298 1984

Voigt C C Capps K A Dechmann D K N Michener R Hand Kunz T H Nutrition or detoxification why bats visit

mineral licks of the Amazonian rainforest PloS one 3 e2011doi101371journalpone0002011 2008

Wang F Li J Wang X Zhang W Zou B Neher D A and LiZ Nitrogen and phosphorus addition impact soil N2O emissionin a secondary tropical forest of South China Scientific reports4 5615 doi101038srep05615 2014

Wilcke W Leimer S Peters T Emck P Rollenbeck R Tra-chte K Valarezo C and Bendix J The nitrogen cycle oftropical montane forest in Ecuador turns inorganic under en-vironmental change Global Biogeochem Cy 27 1194ndash1204doi1010022012GB004471 2013

Williams M R Fisher T R and Melack J M Chemical compo-sition and deposition of rain in the central Amazon Brazil At-mos Environ 31 207ndash217 doi1010161352-2310(96)00166-5 1997

Wolf K Veldkamp E Homeier J and Martinson G O Ni-trogen availability links forest productivity soil nitrous ox-ide and nitric oxide fluxes of a tropical montane forestin southern Ecuador Global Biogeochem Cy 25 GB4009doi1010292010GB003876 2011

Wullaert H Geographie P Chemie F and Gutenberg-universitJ Response of nutrient cycles of an old-growth montane forestin Ecuador to experimental low-level nutrient amendments PhDthesis available at httpubmopushbz-nrwdevolltexte20102312 (last access 8 August 2016) 2010

Yokouchi Y Ikeda M Inuzuka Y and Yukawa T Strong emis-sion of methyl chloride from tropical plants Nature 416 163ndash165 doi101038416163a 2002

Yu H Chin M Yuan T Bian H Remer L A ProsperoJ M Omar A Winker D Yang Y Zhang Y Zhang Zand Zhao C The Fertilizing Role of African Dust in theAmazon Rainforest A First Multiyear Assessment Based onCALIPSO Lidar Observations Geophys Res Lett 42 1984ndash1991 doi1010022015GL063040 2015

Zeng Y and Hopke P A study of the sources of acid precip-itation in Ontario Canada Atmos Environ 23 1499ndash1509doi1010160004-6981(89)90409-5 1989

Abstract

Introduction

Study area

Data and methods

Sample collection materials and data

Statistical evaluation of sodium and chloride ion concentration in rain and OP

Back-trajectory and source--receptor analysis

Results

Sodium and chloride concentration

Spatial allocation of sources and general transport pathways

Discussion

Conclusions

Data availability

Appendix A Tables

Acknowledgements

References

In several tropical and temperate forests human interven-tion in the nitrogen and phosphorus cycles have has beendocumented (Mahowald et al 2005 Matson et al 2002Phoenix et al 2006 Tipping et al 2014 Yu et al 2015)Because nutrient availability regulates ecosystem processesand functions the changes currently affecting the nitrogenand phosphorus budgets are expected to have wide-reachingimpacts in forest ecosystem structure and diversity (Bobbinket al 2010 Homeier et al 2012 Matson et al 2014 Pentildeue-las et al 2013 Pett-Ridge 2009 Wang et al 2014 Wilckeet al 2013) The role of sea salt availability has very re-cently gained attentions as it has been found to conditionthe behaviour of herbivores in addition to affecting carboncycling and organic matter decomposition in tropical ecosys-tems (Dudley et al 2012 Kaspari et al 2008 2009 Pow-ell et al 2009 Voigt et al 2008) At the western rim ofthe Amazon forest in Peru Ecuador and Colombia there isevidence that herbivorous and frugivorous birds and mam-mals visit mineral licks to compensate for low sodium con-centration in plant and fruit tissues (Lee et al 2009 Liz-cano and Cavelier 2004 Powell et al 2009 Voigt et al2008) Furthermore some taxa of arthropod have reportedlybegun practicing geophagy to deal with salt scarcity in plants(Kaspari et al 2008) Yet despite its pantropical importancesalt availability has hitherto been overlooked in most biogeo-graphic and biogeochemical studies (Dudley et al 2012)

By far the most important source of continental sea saltdepositions are the oceans Sea salt scarcity in Amazonianrainwater increases along a gradient from the Atlantic coasttowards the Andean range which acts as a natural orographicbarrier to the west The concentration of both sodium andchloride in rainwater diminishes significantly with increas-ing distance from the Atlantic Ocean (Talbot et al 1990)Additionally the ratio between both concentrations invertsfrom Clminus gt Na+ close to the ocean to Clminus lt Na+ far fromthe ocean (Tardy et al 2005) Consequently tropical moun-tain forests on the eastern slopes of the Andes and tropicallowland forests at the western edge of the Amazon are ex-pected to suffer from sea salt deprivation whereas forestscloser to the Atlantic coast are subject to large sea salt depo-sition (Dudley et al 2012)

The tropical mountain forests at the eastern Andean slopesin southern Ecuador may likely represent an exception of thisgeneralized pattern because of their location in the Huan-cabamba depression an area where the Andes rarely exceed3600 m in altitude This allows the transport of Pacific airmasses rich in sea salt As a result the mountain forest mightbenefit not only from sea salt transported by easterly airmasses from the Atlantic but also by sea salt originating fromPacific air masses Depending on the strength of the contribu-tion to sodium and chloride deposition originating from thePacific the combined input from Atlantic and Pacific sourcescould result in a greater sodium and chloride availability thanthat found for the lowland forests on the western rim of theAmazon (Dudley et al 2012)

However little research has investigated the deposition ofatmospheric sodium and chloride in the tropical forests ofthe southern Ecuadorian Andes Furthermore any such re-search has yet to identify their sources In this context an in-vestigation of the deposition by occult precipitation (OP) isparticularly important because OP comprises an extremelyhigh proportion of total precipitation in tropical mountainforests OP is the water supplied to soil or vegetation by lightdrizzle wind-driven rain and fog andor clouds that conven-tional rain gauges cannot measure An exception is the workof Fabian et al (2009) who estimated the origin of the localsea salt deposition by visual interpretation of single back tra-jectories To our knowledge neither a comprehensive quan-titative investigation on sea salt sources nor any estimates oftheir contribution to the atmospheric deposition have beenconducted yet

As a consequence of the knowledge gaps regarding thesea salt sources of deposition in the Andes of southeast-ern Ecuador the aims of this study are as follows (1) tocharacterize sodium and chloride atmospheric deposition byrain and OP along an altitudinal gradient and at differenttopographic locations in a tropical mountain rainforest site(2) to identify potential Pacific Atlantic and continental ge-ographic sources of sea salt concentration over the Andesof southern Ecuador by applying back-trajectory statisticaltechniques and reanalysis data of atmospheric compositionand (3) to estimate the contribution of each source area to theatmospheric sea salt concentration in the Andes of southernEcuador

2 Study area

The study area is located at the northwestern edge of theAmazon basin (400prime S 7905primeW) at the southeastern An-des of Ecuador approximately a 100 km straight line distanceaway from the Pacific coast and around 2000 km from theclosest part of the Atlantic coast (Fig 1) The study area con-tains the San Francisco valley deeply incised into the east-ern slope of the main Andes range Since 2002 two succes-sive multidisciplinary research programs have investigatedthe Reserva Bioloacutegica San Francisco (RBSF) located on thenorthern slopes of the valley and some areas outside of thereserve (Beck et al 2008 Bendix et al 2013)

The Andes in this area are characterized by lower ele-vations and higher geomorphological complexity comparedto other parts of the mountain chain in northern EcuadorPeru and Colombia Since studies have shown that exposureand altitude affect deposition patterns (Griffith et al 2015Kirchner et al 2014 Lovett and Kinsman 1990 MakowskiGiannoni et al 2013) a precondition for the study of sea saltdeposition is to collect measurements along a large altitudi-nal gradient and at different slope exposures

The climate of the catchment is mainly determined bythe constant tropical easterlies However the trade winds

3 Data and methods

Cerro del Consuelo

El Tiro

PSCFij =mij

ln(Cij

k=1tij

Nsumk=1

)tijk (2)

4 Results

Na rain+

Cl rain-

Na OP+

Cl OP-

Rain OP

Cerro del Consuelo

NH4+ 005 014 081 000 088 009 015 008

El Tiro

minus007 minus006 064 003 082 034 028 022Ca2+

K+ 015 042 035 minus038 085 030 039 002Na+ 097 004 008 001 033 087 031 016SO4

minus004 005 059 027 003 012 009 068Ca2+

NH4+ 001 078 011 minus010 034 042 039 043

minusminus000 022 075 031 080 007 001 001

051015

Cerro del Consuelo

051015

El Tiro

051015

04-01 04-07 05-01 05-07 06-01 06-07 07-01 07-07 08-01 08-07 09-01 09-07

051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

070(b)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Summer (DJF)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Summer (DJF)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

90e-09

80e-09

70e-09

60e-09

50e-09

40e-09

30e-09

20e-09

10e-09

00e+09

-100 -90 -80 -70 -60 -50 -40 -30

Spring (MAM)

-100 -90 -80 -70 -60 -50 -40 -30

Summer (JJA)

-100 -90 -80 -70 -60 -50 -40 -30

Autumn (SON)

-100 -90 -80 -70 -60 -50 -40 -30

Winter (DJF)

ClusterC1C2C3C4C5C6

223 371

231 225

85 186 309

135 52

C1 C2 C3 C4 C5 C6

DJ F MAM SONJ J A

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

20degS

20degN

20degS

20degN

20degS

20degN

20degS

20degN

015 075 135 195 255

kg kg-1 (10^8)

05 15 25 35 45

5 Discussion

6 Conclusions

7 Data availability

Appendix A Tables

References

Abstract

Introduction

Study area

Data and methods




Results



Discussion

Conclusions

Data availability

Appendix A Tables

Acknowledgements

References

3 Data and methods

Cerro del Consuelo

El Tiro

PSCFij =mij

ln(Cij

k=1tij

Nsumk=1

)tijk (2)

4 Results

Na rain+

Cl rain-

Na OP+

Cl OP-

Rain OP

Cerro del Consuelo

NH4+ 005 014 081 000 088 009 015 008

El Tiro

minus007 minus006 064 003 082 034 028 022Ca2+

K+ 015 042 035 minus038 085 030 039 002Na+ 097 004 008 001 033 087 031 016SO4

minus004 005 059 027 003 012 009 068Ca2+

NH4+ 001 078 011 minus010 034 042 039 043

minusminus000 022 075 031 080 007 001 001

051015

Cerro del Consuelo

051015

El Tiro

051015

04-01 04-07 05-01 05-07 06-01 06-07 07-01 07-07 08-01 08-07 09-01 09-07

051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

070(b)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Summer (DJF)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Summer (DJF)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

90e-09

80e-09

70e-09

60e-09

50e-09

40e-09

30e-09

20e-09

10e-09

00e+09

-100 -90 -80 -70 -60 -50 -40 -30

Spring (MAM)

-100 -90 -80 -70 -60 -50 -40 -30

Summer (JJA)

-100 -90 -80 -70 -60 -50 -40 -30

Autumn (SON)

-100 -90 -80 -70 -60 -50 -40 -30

Winter (DJF)

ClusterC1C2C3C4C5C6

223 371

231 225

85 186 309

135 52

C1 C2 C3 C4 C5 C6

DJ F MAM SONJ J A

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

20degS

20degN

20degS

20degN

20degS

20degN

20degS

20degN

015 075 135 195 255

kg kg-1 (10^8)

05 15 25 35 45

5 Discussion

6 Conclusions

7 Data availability

Appendix A Tables

References

Abstract

Introduction

Study area

Data and methods




Results



Discussion

Conclusions

Data availability

Appendix A Tables

Acknowledgements

References

Cerro del Consuelo

El Tiro

PSCFij =mij

ln(Cij

k=1tij

Nsumk=1

)tijk (2)

4 Results

Na rain+

Cl rain-

Na OP+

Cl OP-

Rain OP

Cerro del Consuelo

NH4+ 005 014 081 000 088 009 015 008

El Tiro

minus007 minus006 064 003 082 034 028 022Ca2+

K+ 015 042 035 minus038 085 030 039 002Na+ 097 004 008 001 033 087 031 016SO4

minus004 005 059 027 003 012 009 068Ca2+

NH4+ 001 078 011 minus010 034 042 039 043

minusminus000 022 075 031 080 007 001 001

051015

Cerro del Consuelo

051015

El Tiro

051015

04-01 04-07 05-01 05-07 06-01 06-07 07-01 07-07 08-01 08-07 09-01 09-07

051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

070(b)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Summer (DJF)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Summer (DJF)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

90e-09

80e-09

70e-09

60e-09

50e-09

40e-09

30e-09

20e-09

10e-09

00e+09

-100 -90 -80 -70 -60 -50 -40 -30

Spring (MAM)

-100 -90 -80 -70 -60 -50 -40 -30

Summer (JJA)

-100 -90 -80 -70 -60 -50 -40 -30

Autumn (SON)

-100 -90 -80 -70 -60 -50 -40 -30

Winter (DJF)

ClusterC1C2C3C4C5C6

223 371

231 225

85 186 309

135 52

C1 C2 C3 C4 C5 C6

DJ F MAM SONJ J A

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

20degS

20degN

20degS

20degN

20degS

20degN

20degS

20degN

015 075 135 195 255

kg kg-1 (10^8)

05 15 25 35 45

5 Discussion

6 Conclusions

7 Data availability

Appendix A Tables

References

Abstract

Introduction

Study area

Data and methods




Results



Discussion

Conclusions

Data availability

Appendix A Tables

Acknowledgements

References

Cerro del Consuelo

El Tiro

PSCFij =mij

ln(Cij

k=1tij

Nsumk=1

)tijk (2)

4 Results

Na rain+

Cl rain-

Na OP+

Cl OP-

Rain OP

Cerro del Consuelo

NH4+ 005 014 081 000 088 009 015 008

El Tiro

minus007 minus006 064 003 082 034 028 022Ca2+

K+ 015 042 035 minus038 085 030 039 002Na+ 097 004 008 001 033 087 031 016SO4

minus004 005 059 027 003 012 009 068Ca2+

NH4+ 001 078 011 minus010 034 042 039 043

minusminus000 022 075 031 080 007 001 001

051015

Cerro del Consuelo

051015

El Tiro

051015

04-01 04-07 05-01 05-07 06-01 06-07 07-01 07-07 08-01 08-07 09-01 09-07

051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

070(b)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Summer (DJF)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Summer (DJF)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

90e-09

80e-09

70e-09

60e-09

50e-09

40e-09

30e-09

20e-09

10e-09

00e+09

-100 -90 -80 -70 -60 -50 -40 -30

Spring (MAM)

-100 -90 -80 -70 -60 -50 -40 -30

Summer (JJA)

-100 -90 -80 -70 -60 -50 -40 -30

Autumn (SON)

-100 -90 -80 -70 -60 -50 -40 -30

Winter (DJF)

ClusterC1C2C3C4C5C6

223 371

231 225

85 186 309

135 52

C1 C2 C3 C4 C5 C6

DJ F MAM SONJ J A

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

20degS

20degN

20degS

20degN

20degS

20degN

20degS

20degN

015 075 135 195 255

kg kg-1 (10^8)

05 15 25 35 45

5 Discussion

6 Conclusions

7 Data availability

Appendix A Tables

References

Abstract

Introduction

Study area

Data and methods




Results



Discussion

Conclusions

Data availability

Appendix A Tables

Acknowledgements

References

4 Results

Na rain+

Cl rain-

Na OP+

Cl OP-

Rain OP

Cerro del Consuelo

NH4+ 005 014 081 000 088 009 015 008

El Tiro

minus007 minus006 064 003 082 034 028 022Ca2+

K+ 015 042 035 minus038 085 030 039 002Na+ 097 004 008 001 033 087 031 016SO4

minus004 005 059 027 003 012 009 068Ca2+

NH4+ 001 078 011 minus010 034 042 039 043

minusminus000 022 075 031 080 007 001 001

051015

Cerro del Consuelo

051015

El Tiro

051015

04-01 04-07 05-01 05-07 06-01 06-07 07-01 07-07 08-01 08-07 09-01 09-07

051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

070(b)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Summer (DJF)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Summer (DJF)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

90e-09

80e-09

70e-09

60e-09

50e-09

40e-09

30e-09

20e-09

10e-09

00e+09

-100 -90 -80 -70 -60 -50 -40 -30

Spring (MAM)

-100 -90 -80 -70 -60 -50 -40 -30

Summer (JJA)

-100 -90 -80 -70 -60 -50 -40 -30

Autumn (SON)

-100 -90 -80 -70 -60 -50 -40 -30

Winter (DJF)

ClusterC1C2C3C4C5C6

223 371

231 225

85 186 309

135 52

C1 C2 C3 C4 C5 C6

DJ F MAM SONJ J A

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

20degS

20degN

20degS

20degN

20degS

20degN

20degS

20degN

015 075 135 195 255

kg kg-1 (10^8)

05 15 25 35 45

5 Discussion

6 Conclusions

7 Data availability

Appendix A Tables

References

Abstract

Introduction

Study area

Data and methods




Results



Discussion

Conclusions

Data availability

Appendix A Tables

Acknowledgements

References

Na rain+

Cl rain-

Na OP+

Cl OP-

Rain OP

Cerro del Consuelo

NH4+ 005 014 081 000 088 009 015 008

El Tiro

minus007 minus006 064 003 082 034 028 022Ca2+

K+ 015 042 035 minus038 085 030 039 002Na+ 097 004 008 001 033 087 031 016SO4

minus004 005 059 027 003 012 009 068Ca2+

NH4+ 001 078 011 minus010 034 042 039 043

minusminus000 022 075 031 080 007 001 001

051015

Cerro del Consuelo

051015

El Tiro

051015

04-01 04-07 05-01 05-07 06-01 06-07 07-01 07-07 08-01 08-07 09-01 09-07

051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

070(b)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Summer (DJF)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Summer (DJF)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

90e-09

80e-09

70e-09

60e-09

50e-09

40e-09

30e-09

20e-09

10e-09

00e+09

-100 -90 -80 -70 -60 -50 -40 -30

Spring (MAM)

-100 -90 -80 -70 -60 -50 -40 -30

Summer (JJA)

-100 -90 -80 -70 -60 -50 -40 -30

Autumn (SON)

-100 -90 -80 -70 -60 -50 -40 -30

Winter (DJF)

ClusterC1C2C3C4C5C6

223 371

231 225

85 186 309

135 52

C1 C2 C3 C4 C5 C6

DJ F MAM SONJ J A

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

20degS

20degN

20degS

20degN

20degS

20degN

20degS

20degN

015 075 135 195 255

kg kg-1 (10^8)

05 15 25 35 45

5 Discussion

6 Conclusions

7 Data availability

Appendix A Tables

References

Abstract

Introduction

Study area

Data and methods




Results



Discussion

Conclusions

Data availability

Appendix A Tables

Acknowledgements

References

Rain OP

Cerro del Consuelo

NH4+ 005 014 081 000 088 009 015 008

El Tiro

minus007 minus006 064 003 082 034 028 022Ca2+

K+ 015 042 035 minus038 085 030 039 002Na+ 097 004 008 001 033 087 031 016SO4

minus004 005 059 027 003 012 009 068Ca2+

NH4+ 001 078 011 minus010 034 042 039 043

minusminus000 022 075 031 080 007 001 001

051015

Cerro del Consuelo

051015

El Tiro

051015

04-01 04-07 05-01 05-07 06-01 06-07 07-01 07-07 08-01 08-07 09-01 09-07

051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

070(b)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Summer (DJF)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Summer (DJF)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

90e-09

80e-09

70e-09

60e-09

50e-09

40e-09

30e-09

20e-09

10e-09

00e+09

-100 -90 -80 -70 -60 -50 -40 -30

Spring (MAM)

-100 -90 -80 -70 -60 -50 -40 -30

Summer (JJA)

-100 -90 -80 -70 -60 -50 -40 -30

Autumn (SON)

-100 -90 -80 -70 -60 -50 -40 -30

Winter (DJF)

ClusterC1C2C3C4C5C6

223 371

231 225

85 186 309

135 52

C1 C2 C3 C4 C5 C6

DJ F MAM SONJ J A

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

20degS

20degN

20degS

20degN

20degS

20degN

20degS

20degN

015 075 135 195 255

kg kg-1 (10^8)

05 15 25 35 45

5 Discussion

6 Conclusions

7 Data availability

Appendix A Tables

References

Abstract

Introduction

Study area

Data and methods




Results



Discussion

Conclusions

Data availability

Appendix A Tables

Acknowledgements

References

051015

Cerro del Consuelo

051015

El Tiro

051015

04-01 04-07 05-01 05-07 06-01 06-07 07-01 07-07 08-01 08-07 09-01 09-07

051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

070(b)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Summer (DJF)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Summer (DJF)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

90e-09

80e-09

70e-09

60e-09

50e-09

40e-09

30e-09

20e-09

10e-09

00e+09

-100 -90 -80 -70 -60 -50 -40 -30

Spring (MAM)

-100 -90 -80 -70 -60 -50 -40 -30

Summer (JJA)

-100 -90 -80 -70 -60 -50 -40 -30

Autumn (SON)

-100 -90 -80 -70 -60 -50 -40 -30

Winter (DJF)

ClusterC1C2C3C4C5C6

223 371

231 225

85 186 309

135 52

C1 C2 C3 C4 C5 C6

DJ F MAM SONJ J A

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

20degS

20degN

20degS

20degN

20degS

20degN

20degS

20degN

015 075 135 195 255

kg kg-1 (10^8)

05 15 25 35 45

5 Discussion

6 Conclusions

7 Data availability

Appendix A Tables

References

Abstract

Introduction

Study area

Data and methods




Results



Discussion

Conclusions

Data availability

Appendix A Tables

Acknowledgements

References

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Summer (DJF)

-100-90 -80 -70 -60 -50 -40 -30 -20

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Spring (SON)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Summer (DJF)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Autumn (MAM)

-100-90 -80 -70 -60 -50 -40 -30

-30-25-20-15-10-5051015

Winter (JJA)

90e-09

80e-09

70e-09

60e-09

50e-09

40e-09

30e-09

20e-09

10e-09

00e+09

-100 -90 -80 -70 -60 -50 -40 -30

Spring (MAM)

-100 -90 -80 -70 -60 -50 -40 -30

Summer (JJA)

-100 -90 -80 -70 -60 -50 -40 -30

Autumn (SON)

-100 -90 -80 -70 -60 -50 -40 -30

Winter (DJF)

ClusterC1C2C3C4C5C6

223 371

231 225

85 186 309

135 52

C1 C2 C3 C4 C5 C6

DJ F MAM SONJ J A

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

20degS

20degN

20degS

20degN

20degS

20degN

20degS

20degN

015 075 135 195 255

kg kg-1 (10^8)

05 15 25 35 45

5 Discussion

6 Conclusions

7 Data availability

Appendix A Tables

References

Abstract

Introduction

Study area

Data and methods




Results



Discussion

Conclusions

Data availability

Appendix A Tables

Acknowledgements

References

-100 -90 -80 -70 -60 -50 -40 -30

Spring (MAM)

-100 -90 -80 -70 -60 -50 -40 -30

Summer (JJA)

-100 -90 -80 -70 -60 -50 -40 -30

Autumn (SON)

-100 -90 -80 -70 -60 -50 -40 -30

Winter (DJF)

ClusterC1C2C3C4C5C6

223 371

231 225

85 186 309

135 52

C1 C2 C3 C4 C5 C6

DJ F MAM SONJ J A

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

20degS

20degN

20degS

20degN

20degS

20degN

20degS

20degN

015 075 135 195 255

kg kg-1 (10^8)

05 15 25 35 45

5 Discussion

6 Conclusions

7 Data availability

Appendix A Tables

References

Abstract

Introduction

Study area

Data and methods




Results



Discussion

Conclusions

Data availability

Appendix A Tables

Acknowledgements

References

DJ F MAM SONJ J A

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC6

0 56 112 168 224

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC5

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC3

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC4

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC2

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

-1(E-9)

PressureC1

20degS

20degN

20degS

20degN

20degS

20degN

20degS

20degN

015 075 135 195 255

kg kg-1 (10^8)

05 15 25 35 45

5 Discussion

6 Conclusions

7 Data availability

Appendix A Tables

References

Abstract

Introduction

Study area

Data and methods




Results



Discussion

Conclusions

Data availability

Appendix A Tables

Acknowledgements

References

20degS

20degN

20degS

20degN

20degS

20degN

20degS

20degN

015 075 135 195 255

kg kg-1 (10^8)

05 15 25 35 45

5 Discussion

6 Conclusions

7 Data availability

Appendix A Tables

References

Abstract

Introduction

Study area

Data and methods




Results



Discussion

Conclusions

Data availability

Appendix A Tables

Acknowledgements

References

5 Discussion

6 Conclusions

7 Data availability

Appendix A Tables

References

Abstract

Introduction

Study area

Data and methods




Results



Discussion

Conclusions

Data availability

Appendix A Tables

Acknowledgements

References

6 Conclusions

7 Data availability

Appendix A Tables

References

Abstract

Introduction

Study area

Data and methods




Results



Discussion

Conclusions

Data availability

Appendix A Tables

Acknowledgements

References

6 Conclusions

7 Data availability

Appendix A Tables

References

Abstract

Introduction

Study area

Data and methods




Results



Discussion

Conclusions

Data availability

Appendix A Tables

Acknowledgements

References

Appendix A Tables

References

Abstract

Introduction

Study area

Data and methods




Results



Discussion

Conclusions

Data availability

Appendix A Tables

Acknowledgements

References

References

Abstract

Introduction

Study area

Data and methods




Results



Discussion

Conclusions

Data availability

Appendix A Tables

Acknowledgements

References

Abstract

Introduction

Study area

Data and methods




Results



Discussion

Conclusions

Data availability

Appendix A Tables

Acknowledgements

References

Abstract

Introduction

Study area

Data and methods




Results



Discussion

Conclusions

Data availability

Appendix A Tables

Acknowledgements

References

Abstract

Introduction

Study area

Data and methods




Results



Discussion

Conclusions

Data availability

Appendix A Tables

Acknowledgements

References

Documents

Atmospheric salt deposition in a tropical mountain ... · Sea salt (NaCl) has recently been proven to be of the utmost importance for ecosystem functioning in Amazon lowland forests