aerosol and precipitation chemistry at a remote himalayan

Aerosol Science and Technology 36 441ndash456 (2002)cdeg 2002 American Association for Aerosol ResearchPublished by Taylor and Francis0278-6826=02=$1200 C 00

Aerosol and Precipitation Chemistry at a RemoteHimalayan Site in Nepal

Arun B Shrestha1 Cameron P Wake2 Jack E Dibb2 and Sallie I Whitlow1Department of Hydrology and Meteorology Babar Mahal Kathmandu Nepal2Climate Change Research Center Institute for the Study of Earth Oceans and the SpaceUniversity of New Hampshire Durham New Hampshire

As a part of a year long aerosol measurement program in theNepal Himalayas a series of 12 h samples of aerosol and event-based samples of precipitation were collected in postmonsoon1996 premonsoon 1997 and monsoon 1997 Results show thatdiurnal variations in the regional valley wind system are responsi-ble for variations in the major ion (Na+ NH+

4 K+ Mg2+ Ca2+Cliexcl NOiexcl

3 and SOiexcl24 ) concentrations in the aerosolOn time scales

greater than a day changes in atmospheric circulation and associ-ated changes in the precipitation regime have an important effecton aerosol ion concentrations

Periods of high major ion concentrationin the aerosol generallycoincide with similar periods in precipitation chemistry althougha linear relationshipbetween concentrations in these 2 media is notrobust due to limited data High scavenging ratio values are foundfor all species except SO2iexcl

4 Nitrate Cliexcl and NH+4 are enriched in

the precipitation compared to the aerosol implying the presenceof gaseous acidic species (HNO3 and HCl) and gaseous NH3 in theair Ammonium was also enriched with respect to SO2iexcl

4 in aerosolduring a dry episode in the monsoon season This may representrelatively local sources of NH3 (from neighboring villages)and wasnot scavenged due to the lack of precipitation

Empirical orthogonal function (EOF) analysis clearly shows thepresence of 2 dominant pollutant transport mechanisms for thepremonsoon and monsoon seasons (ie valley wind system andmonsoon circulation) Although physically different these 2 trans-port mechanisms follow similar transport pathways and trans-port aerosol into the Himalayas from similar source regions Fur-ther EOF analysis suggests a southerly maritime signal in theaerosol during monsoon and a more distant westerly maritime sig-

Received 22 August 2000 accepted 31 July 2001This study was partially funded by the Global Change System for

Analysis Research and Training (START)We thank the villagers of Phortse and the residents of Jiri for their

support of the sampling program Dr R W Talbot and Dr B D Keimprovided a detailed review on this manuscript We also thank DHMfor providing meteorological equipment and other logistical supportMost of all we are thankful to our observer Dorzee Sherpa Phortsefor his hard work Two anonymous reviewers provided some valuablesuggestions that helped improve the manuscript

Address correspondence to Arun B Shrestha Department ofHydrologyand MeteorologyPO Box 406 Babar Mahal KathmanduNepal E-mail abswlinkcomnp

nal during premonsoon Our results support the potential for us-ing glaciochemical records from the Himalayas to investigate vari-ations in the strength of past monsoon circulation and westerlydisturbances

INTRODUCTIONIce cores retrieved from Greenland and Antarctica contain

atmospheric chemistry records ranging from seasons to hun-dreds of thousand of years providing insight into climatic vari-ations over these periods (De Angelis et al 1987 Delmas andLegrand 1989 Jouzel et al 1990 Mayewski et al 1993 19941996) Tropical and sub-tropical glaciers do not contain recordsas far back in the past however they may provide informa-tion concerning past climatic changes in regions where such in-formation is virtually nonexistent (eg Thompson et al 1984Lyons et al 1991) The vast glaciated regions in the Himalayascontain valuable information on monsoon variations over thepast several thousand years Quantitative estimates of past at-mospheric chemical composition based on glaciochemical datadepend greatly on the knowledge of the relationship betweenthe chemistry of the precipitation and that of the airmass withinwhich it forms and deposition These relationships are sitespeci c so an interpretation of glaciochemical data from theHimalayas requires knowledge of the relationship betweenaerosol and precipitation chemistry in the region

Studies on the chemical composition of aerosol and precipi-tation in the Himalayas have been limited to date on both tem-poral and spatial scales A short-term sampling program in theNguzompa glacier basin near Mt Everest showed low ion bur-dens in both aerosol and snow (Wake et al 1994) A 2 weeksampling project in Hidden Valley in the Himalayas of west-ern Nepal showed that temporal variations in aerosol concen-trations were related to the in ux of air masses by monsooncirculation (Shrestha et al 1997) Recently a year-long sam-pling of atmospheric aerosol (with sampling intervals of 5 days)was conducted at a remote Himalayan site and a rural Middle-Mountainsite in Nepal (Shrestha et al 2000)This study showed

441

442 A B SHRESTHA ET AL

marked seasonal variations in the aerosol concentrations dueto changes in circulation and associated precipitation scaveng-ing As part of the same program aerosol samples were col-lected over 6 and 12 h sample intervals at the Himalayan siteduring 3 different seasons premonsoon monsoon and post-monsoonAerosol samplingwas complementedby precipitationsamples This paper investigates the variations in aerosol con-centrations during the intensive sampling programs and theirconnection with meteorological variables and transport mecha-nisms A preliminary attempt is made to quantify the relation-ship between aerosol concentrations and concentrations in theprecipitation

EXPERIMENTAL

The SiteThe sampling site was located near the remote village of

Phortse in the Khumbu region of the Himalayas in Nepal

(a)

Figure 1 Location map of Phortse sampling station The arrows show predominant daytime local wind directions (Continued)

(Figure 1) The sampling site stands at an elevation of 4100 masl at 27plusmn510 N latitude and 86plusmn450 E longitude and is situatedabout 350 m above the village close to where a northwest tosoutheast trending ridge merges with the mountain slope Theridge continues down to the con uence of the Imja Khola andDudh Kosi rivers at an elevation of 3550 m asl A meteoro-logical station was established 50 m below the aerosol samplingsite on the same ridgeAlthoughthe predominantwind directionover the valley is southwesterly observations of the local windreveal that wind near the sampling station blows predominantlyfrom a south-south-east direction (across the ridge) Thereforethe sampling site is well protected from sources of pollution inthe village The village of Phortse consists of about 60 perma-nent houses There are small patches of farmland around the vil-lage where potato and buckwheat are cultivated between Apriland September There are also some forested areas dominatedby juniper and different varieties of pine A small airstrip atSyangboche is located at a distance of about a 4 h walk

AEROSOL AND PRECIPITATION IN NEPAL 443

(b)

Figure 1 (Continued)

(5 km aerial distance) and a slightly larger airport Lukla isa 3 day walk from the Phortse village The nearest road headJiri is at a distance of a 7 day walk (ca 70 km aerial distance)

Aerosol SamplesThe sampling described in this paper was part of a year-long

aerosol sampling program at Phortse During 3 different periodsin postmonsoon premonsoon and monsoon seasons regular5 day integrated sampling was replaced by intensive samplingprograms during which samples were collected over time in-tervals of 6 to 12 h These intensive sampling programs wereconducted between September 24 to 29 1996 March 11 to 291997 and August 11 to 31 1997 corresponding to postmon-soon premonsoon and monsoon periods The rst samplingperiod was relatively short due to a pump breakdown There-fore only data for the premonsoon and monsoon samples arediscussed in detail

Aerosol samples were collectedon 2 sup1m pore size 90 mm di-ameterZe uor Te on lters (Gelman Sciences) The lters wereplaced facing downwards in a cylindrical polyethylene protec-tive cover The air was drawn by 24 high volume pumps pow-

ered by a combination of photovoltaic cells and batteries Thevolume of air sampled was measured by an inline ow meterCorrections for ambient temperature and pressure allowed forconversion of the measured volumes to standard cubic meters(scm) For the mean ow rate of 040 scm hiexcl1 the velocity atthe face of the lter was suf ciently high so that the ef ciencyof collection for particles as small as 03 sup1m is gt99 (Liuet al 1984) The ow velocity at the opening of the protectivecover was calculated to be 3 cm siexcl1 according to which thecutoff for large particles as given by the sedimentation velocityof particles is estimated at 8 sup1m (Barrie 1985 Davidson 1989Warneck 1988)

Care was taken to minimize contamination both in the lab-oratory and in the eld The lter cartridges were loaded andpacked in clean plastic bags in a class 100 clean lab at the Uni-versityof New Hampshire The 90 mm cartridgeswere packed inclean plastic bags and transported in air-tight containers Earliersampling in the Himalayas showed an abundance of NHC

4 in theaerosol and it was thought that postsampling reaction with NH3

in the ambient air might have created a positive artifact for NHC4

(Haagenson and Shapiro 1979 Hayes et al 1980 Silvente and


Legrand 1993) To avoid contamination by ambient air as wellas evaporation of sampled species during transport and storage lters were removed from their holders and were individuallyplaced in precleaned air-tight glass bottles immediately aftersampling at Phortse The lters were placed with the exposedside of the lters facing inside of the bottles The glass bottleswere then placed in air-tight dark containers The lters wereremoved from the glass bottles just prior to chemical analysisLoading and unloading of samples was carried out wearing anonparticulatingclean suit hood face mask and plastic glovesBlank lters were handled in the same manner as the samplesAfter sampling lters were brought to Kathmandu where theywere stored for several days The lters were then shipped tothe Climate Change Research Center (CCRC) at the Universityof New Hampshire by air cargo for extraction of the lter andlaboratory analysis

The lters were analyzed for major water soluble ionicspecies (NaC NHC

4 KC Mg2C Ca2C Cliexcl NOiexcl3 and SO2iexcl

4 )at the CCRC laboratory In order to analyze the major ion con-centration the sample and blank lters were wetted with 05 mlultrapuremethanolThe solublecomponentswere thenextractedwith three 5 ml aliquots of deionized Milli-Q water Major ionconcentrations in the aqueous extracts were determined by ionchromatographyusinga Dionexmodel 4000 ionchromatographCations were analyzed using a CS12 column 22 mM MSA elu-ent and CSRS suppresser and anions were analyzed using anAS11 column 66 mM NaOH eluent and MMS suppresser Thesample loop was 250 sup1l and the sample running time was ap-proximately 9 min

Ten eld blanksand5 lab blankswere collectedWe de ne thedetection limit for major inorganic ions as the standard deviationof all the blanks divided by the mean volume of all the samples(after Talbot et al (1986)) The detection limits are presentedin Table 1 Mean blank values were deducted from the samplevalues

The mountainous terrain is characterized by distinct diur-nal variations in the local wind pattern Generally the wind isup-slope (valley wind) during the daytime and down-slope(mountain wind) during the nighttime (Nakajima 1976

Table 1Detection limits of aerosol and precipitation samples from Phortse

Period NaC NHC4 KC Mg2C Ca2C Cliexcl NOiexcl

3 SO2iexcl4

Aerosol (neq scmiexcl1)Sep 1996 220 003 018 020 108 049 062 045Mar 1997 022 095 014 007 035 019 015 100Aug 1997 021 041 002 005 046 008 ltsnzv 007

Precipitation (sup1eq kgiexcl1)Sep 1996 004 002 ltsnzv 003 004 017 ltsnzv ltsnzvMar 1997 ltsnzv ltsnzv ltsnzv 001 007 001 ltsnzv ltsnzvAug 1997 ltsnzv 004 002 000 005 000 ltsnzv ltsnzv

snzv D speci ed nonzero value

Hindman 1995a b) While this is valid for all seasons the tim-ing of the diurnal wind variation depends on the geometry of thevalley and season Initially 3 samples were taken daily in thebeginning of the postmonsoon sampling at 0800 to 1400 1400to 1800 and 1800 to 0800 h (local standard time) Major ionconcentrations in the 2 daytime samples were similar for mostspecies except SO2iexcl

4 The lack of signi cant and consistent dif-ferencesbetween the 2daytimesamplescombinedwith thestudyof local wind patterns indicates that the mountain wind domi-nates only after 1700 h and the daytime samples were affectedprimarily by valley wind In the sampling programs thereafteronly 1 daytime (0800ndash1800 h) and 1 nighttime (1800ndash0800 h)sample was collected daily

Precipitation SamplesPrecipitation samples were collected in 500 and 1000 ml

polypropylene jars with an ori ce diameter of 150 mm (largerjars were used for collectingsolid precipitation)The precleanedjars were packed and sealed in clean bags (Clean Room Prod-ucts New York) Prior to exposure the jars with caps on wereremoved from plastic bags and mounted on a 15 m high standNext the cap was removed from the jar and placed in the origi-nal plastic bag Immediately after sample collection the jar wassealed tightly by the cap to avoid dry deposition and was placedin air-tight dark containers Storing in the container allowedsolid precipitation to melt Between 6 to 12 h after sampling theprecipitation samples were transferred to 30 ml polypropylenebottles Depending on the amount of precipitation 2ndash4 aliquotswere prepared for each precipitation event To avoid changesin concentrations due to bioconsumption half of the aliquotswere treated with ultrapure chloroform (1 of the aliquot vol-ume Galloway (1978)) The polypropylene bottles were thenplaced in dark air-tight containers Samples were frozen soonafter reaching Kathmandu and until they were dispatched toCCRC

Precipitationsampleswere alsoanalyzedat theCCRClabora-tory using a Dionex ion chromatographAmmonium concentra-tions were determined as average values of aliquots treated withchloroform while Cliexcl values were determined from untreated


samples For other species average values of all samples forthe precipitation event were used Standard deviations of mul-tiple samples representing the same precipitation event wereadequately low (lt5 of the mean) The amount of precipita-tion was determined by the readings from a manual precipita-tiongaugeDetection limits determinedas described in Shresthaet al (1997) are presented in Table 1 The rst precipitationsam-ple from the premonsoon sampling program is suspect becauseof unrealistically high (gt200 times the average) concentrationsfor all species except NHC

4 and NOiexcl3 and were therefore not

included in further analyses The same procedure of samplingsolid and liquid precipitation and missing initial small amountsof precipitation may have an impact on the data

Meteorological DataA meteorological station established and operated in con-

junction with the aerosol sampling program included a thermo-hygrograph an actinograph an anemometer a rain gauge andan altimeter barometer Melting was required to measure theamount of solid precipitationGeneral weather conditionsat thesite such as cloud cover and movement visibility fog mistetc were observed and noted twice daily at standard meteoro-logical observation times for Nepal (0845 and 1745 h) duringthe sampling program Synoptic weather maps prepared by theDepartment of Hydrology and Meteorology Nepal were ac-quired to obtain informationon synopticscale circulationduringthe sampling programs Wind direction data was not obtainedduring the 1997 monsoon due to instrument malfunction Forthat period data from a permanent meteorological station atDingboche (about 5 km northeast of Phortse) is used The winddata from Dingboche was well correlated with wind data fromPhortse during other sampling periods

RESULTS AND DISCUSSION

Diurnal Variations in Aerosol ConcentrationThe nighttime samples generally show lower concentrations

than daytime samples We present events for dominant ionicspecies in aerosol (ie NHC

4 NOiexcl3 and SO2iexcl

4 ) when consecu-tivedaytime and nighttimesamples were taken (Figure 3) Thereare 19 (out of 28) such events with conspicuous differences indaytime and nighttime concentrations for all species Amongthese daytime samples show higher concentrations thannighttime for 12 events while nighttime samples show higherconcentrations for the remaining 7 events In the premonsoonsamplingprogram only during the period between March 1920and 2122 did nighttime samples for SO2iexcl

4 show considerablyhigher concentrations compared to daytime samples (Figure 3left)

Apart from wind direction precipitation plays an importantrole in variations in aerosol concentrations (Figures 2 and 4)Precipitation was generally nocturnal in Phortse (67 of thetotal precipitationevents) Precipitationscavengingplays an im-portant role in reducing aerosol concentration in the Himalayas

(Shrestha et al 1997 2000) which could have resulted in theobserved diurnal variations in aerosol concentration (Figure 3)However there are several instances when nocturnal precipita-tion did not occur and nighttime concentrationsare still reduced(eg March 1516 1718 and 2122) Besides precipitationthe diurnal variation in regional scale wind systems (Shresthaet al 2000) could also effect the aerosol concentrations Pollu-tants transported to the site by regional scale valley wind sys-tems during the daytime are replaced by relatively cleaner airfrom aloft by the mountain wind system in the nighttime Thenortheasterly wind direction during the night on March 15 and20 indicates the presence of mountain wind at the samplingsite (Figure 4) The differences in daytime and nighttime sam-ples even when there was no precipitation event suggests thatthe role of diurnal changes in the wind pattern is the dominantcontrol on the diurnal uctuation of aerosol concentrationsCo-incidentally after March 23 in the premonsoon samples therewas no precipitation and the mountain wind system also ceased(Figure 4)

In contrast nighttime concentrationsare substantiallyhigherthan daytime concentrations from August 2122 to 2324 and onAugust 2728 for most species (Figure 3) Wind direction datafrom Dingboche shows the presence of mountain wind duringthis period (Figure 4) Mountain winds occur over a wide areain the Khumbu region (Hindman et al 1996) and the observa-tion of low level cloud movement at Phortse did show diurnalchanges in the local winds Occasionally during the early mon-soon the Phortse station experienced higher levels of pollutioncompared to a lower elevation station at Jiri This was likelydue to the intrusion of a polluted monsoon layer (Shrestha et al2000) Higher nighttime concentrations suggest that pollutantsare brought to the sites during the night by the mountain wind owing down the valley from northeast of the sampling siteThe nocturnal mountain wind could readily transport air massesfrom the polluted layer to the Phortse site causing nighttimerises in major ion concentrations during the periods when pre-cipitation did not occur An airborne measurement of aerosolspecies and meteorological parameters is planned in the nearfuture in the Nepal Himalayas which may provide valuable in-sight into the existence of a pollution layer overriding relativelyclean air masses (Hindman)

Multiday Variations in Aerosol ConcentrationPremonsoon concentrations of NHC

4 KC Mg2C Ca2CNOiexcl

3 and SO2iexcl4 are 2 to 10 times greater than the monsoon

concentrations (Figure 2) on average The differences in aerosolconcentrations between the 2 sampling periods (premonsoonand monsoon) are a part of greater seasonal variations The dis-cussion here focuses on concentration variations within eachsampling season

Concentrationvariationsatmultidaytimescales (several sam-ples) are characteristic of the premonsoon and monsoon con-centration series (Figure 2) At these time scales (several daysto a week) precipitation events were generally associated with


Figure 2 Time series of water solublemajor ions in the aerosol (solid lines) and precipitation (stippled bars) measured at Phortseduring premonsoon 1997 (left) and monsoon 1997 (right) seasons Note the difference in vertical scales Short breaks in aerosoldata are due to a gap while changing lters Longer data gaps are when samples are below detection unless indicated as ldquomdrdquo(missing data) Data offscale is labeled in the plot

Fig

ure

3D

aytim

ean

dni

ghtt

ime

conc

entr

atio

nsof

maj

orio

nic

spec

ies

inae

roso

lsam

ples

colle

cted

atPh

orts

eD

ays

with

outc

onse

cutiv

eda

yan

dni

ghts

ampl

esar

eno

tpre

sent

edin

the

gur

eD

aytim

ean

dni

ghtti

me

prec

ipita

tion

even

tsar

ein

dica

ted

byldquod

rdquoan

dldquon

rdquoin

(a)

and

(d)

whe

reas

ldquodnrdquo

indi

cate

spr

ecip

itatio

nev

ents

cove

ring

both

the

dayt

ime

and

the

follo

win

gni

ght

Not

eth

edi

ffer

ence

inve

rtic

alsc

ales

447

Fig

ure

4M

eteo

rolo

gica

ldat

aco

llect

edin

Phor

tse

duri

ngpr

emon

soon

199

7(l

eft)

and

mon

soon

199

7(r

ight

)H

ere

we

use

win

ddi

rect

ion

data

from

Din

gboc

hefo

rmon

soon

199

7as

ther

eis

now

ind

dire

ctio

nda

tafr

omPh

orts

efo

rtha

tper

iod

Poin

tsin

win

ddi

rect

ion

plot

sin

dica

teho

urly

data

whi

lea

solid

line

isw

eigh

ted

mea

nw

ith

5sm

ooth

ing

fact

or

448


decreases in aerosol concentrations For example during thepremonsoon sampling program concentrations of most ionicspecies decreased dramatically following precipitation eventsin March (ie March 13ndash14 1718 and 1920) (Figure 4)However the precipitation of March 23 was of low intensityand apparently did not signi cantly alter aerosol concentrationsIn addition the circulation pattern that is generally northwest-erly in premonsoon (Figure 5a) and is relatively cleaner anddoes not support advective transport from the south shiftedto a southwesterly ow between March 20 and 25 (Figure 5b)(Mayewski et al 1983) This shift in large-scale circulation iscapable of incorporating more pollutants from distant southerlyand southwesterly sources After March 25 the 500 mb circula-tion reestablished the pattern similar to that at the beginning ofthe month ( gure not shown)

The monsoon samples also display signi cant variations inmajor ion concentrations on time scales of a week The mon-soon samples can be divided into 2 parts August 8ndash20 andAugust 20ndash28 The rst period was characterized by low con-centrations close to or lt1 neq scmiexcl1 for all species except NaC

and Cliexcl (Figure 2 right) The sample of August 1920 (night)shows a large increase in concentrations of all species exceptCliexcl although all ion concentrations remain signi cantly lowerthan those during the premonsoon period until the end of thesampling period when they returned to the lower level charac-teristic of the rst half of the monsoon sampling (August 8ndash20)The sequence of high and low concentrations of atmosphericaerosol on weekly time scales in the monsoon sampling periodmatches a similar sequenceof variations in the amount of precip-itationPrecipitationrecords for the monsoon samplingprogramshow 2 distinct periods a wet period between August 8 and 17and a substantially drier period between August 18 and 27 Thelast days of the monsoonsamplingprogram (August 28ndash30) shiftback into a wet periodAnalysis of synopticweathermaps showsthat the monsoon trough shifted southward during that period(Figures 5c and d) This caused air masses reaching the site tochange trajectories from southeasterly to easterly The shift inair-mass trajectory elongated the travel distance from the Bayof Bengal such that the moisture got depleted in the earlier partof the path It is suggested that the high aerosol concentrationswere caused by incorporationof the pollutants into the relativelydry air masses in its path

Aerosol-Precipitation RelationshipConcentrations of chemical constituents in precipitation col-

lected during the premonsoon and monsoon intensive samplingprograms are shown in Figure 2 As with the aerosol samplesthe concentrations of major ions in the precipitation were highduring the premonsoon season as well as in the 2 isolated pre-cipitation samples of August 21 and 2122 and low in the rsthalf of the monsoon sampling In addition the concentrationsof water soluble species in the precipitation in the postmonsoonsampling were similar to those during the rst half of the mon-soon sampling (limited in time and samples) Figure 2 shows

that periods with high aerosol concentrations generally corre-spond to similar periods in the precipitation concentration Asa preliminary step to quantify the relationship between concen-trations in the precipitationand in the aerosol scavenging ratioswere calculated according to

W Dfrac12aCs

Ca

where Cs D concentrations in precipitation (ng giexcl1) Ca Dconcentration of aerosol (ng miexcl3) and frac12a D density of air(764 g miexcl3 after correcting for the standard temperature andpressure) (Davidson et al 1993)

The mean scavenging ratios for most species were signif-icantly higher than those found in Greenland (eg Davidsonet al 1993) or in the Swiss Alps (Baltensperger et al 1993)(Table 2) The scavenging ratio for SO2iexcl

4 although relativelylow compared to scavenging ratios of other species is com-parable to values found in Greenland (350 Baltensperger et al(1993)) The scavengingratios are highlyvariable for all speciesThe variability may be attributed to differences in precipitationformation gas and aerosol-phase atmospheric concentrationsandor the extent of below cloud scavenging between individ-ual events as samples were collected during different weatherconditions and seasons Our data set does not allow assessmentof the relative importance of the different processes control-ling relationships between aerosol and precipitation composi-tion However sample-to-sample variations in aerosol concen-trations are not re ected in corresponding precipitation eventsAlthough there is a high correlation between concentrations ofmost species in the aerosol and in the precipitationthis relation-ship is biased by a small number of samples with considerablyhigherconcentrationsthanaverage concentrationsin aerosol andin the precipitation ( gure not shown) Table 3 shows that majorion concentrations for most species in aerosol covary with thosein precipitationwhen concentrationsare averaged over episodesof high and low concentrations and over the 2 sampling seasons(premonsoon and monsoon) While these results show that theconcentrationincrease (decrease) inonemedium is re ectedby asimilar increase (decrease) in the other the magnitude of changeis not consistent and the relationship is not linear Since the re-lationship signi cantly improved by averaging over longer timeperiods it is possible that the relationship will further improveand approach a linear relationship as more data are collected inthe future This would provide greater con dence when usingglaciochemical data for interpreting atmospheric compositionof the past as the glaciochemical data provides concentrationsat seasonal or annual resolution

The mean scavenging ratio of NOiexcl3 is relatively high In the

aerosol NOiexcl3 concentrations were equivalent to only one third

of the NHC4 concentrations (Table 2) while in precipitation the

NOiexcl3 concentrations were greater than NHC

4 (Figures 6andashb)The relative increase in NOiexcl

3 concentrations was greater in theperiod of higher aerosol concentrations (eg during premon-soon) This suggests the presence of gaseous HNO3 in the air

Fig

ure

5G

eopo

tent

ialh

eigh

tsat

500

mb

leve

lT

heda

shed

arro

ws

indi

cate

tent

ativ

eis

obar

ictr

ajec

tory

ofai

rm

asse

sar

rivi

ngat

the

site

450


Table 2Scavenging ratios for water soluble major ions


3 SO2iexcl4

Mean 1393 1811 2436 2580 1805 2582 2502 487S Dev 1593 2351 2092 2277 1873 1870 2007 326

(largely unsampled on our lters) especially when the air wasmore pollutedRelativelyhigh scavenging ratios for NOiexcl

3 couldresult from the scavenging of gas-phase HNO3 by precipitationSimilarly our results show enhancement of Cliexcl with respect toNaC in the precipitationChloride was in large de cit relative toCliexcl in aerosol and there is no correlation between these species(Figure 6c) In contrast NaC and Cliexcl were not only linearlyrelated in the precipitation but Cliexcl was also enriched relativeto NaC (Figure 6d) The best t line shows that CliexclNaC inprecipitation was close to the mean sea-water ratio between Cliexcl

and NaC (1165 Drever (1988)) Such enrichment of Cliexcl withrespect to NaC in the following precipitation could be due to2 factors

(1) Presence and scavenging of gaseous HCl in air can causeenrichment of Cliexcl in the precipitationThe NaCCliexcl ratio in theprecipitation will approach the average sea water ratio if both ofthese components originated as marine aerosol and underwentchemical fractionationby the reaction with for example H2SO4

as they traveled to the site Evidence of fractionation of sea saltparticles by H2SO4 was found in Hidden Valley where similarenrichment of Cliexcl in the precipitation was observed (Shresthaet al 1997)

(2) Air masses transported to the site by more vigorous circu-lation are likely to be less fractionated due to shorter travel timeSince air masses causing precipitation represent stronger circu-lation enrichment of Cliexcl in the precipitation could be becauseof the bias of sampling such air masses When only the aerosoldata during the precipitation events in Phortse were examined

Table 3Event-based geometric means of selected major ionic species in aerosol and precipitation

NHC4 SO2iexcl

4 Ca2C

Events Aerosol Precipitation Aerosol Precipitation Aerosol Precipitation

Mar 13ndash19 1556 (4) 1292 (5) 5335 (3) 591 (4) 967 (4) 2127 (4)Mar 19ndash23 5805 (2) 2110 (2) 15279 (1) 3322 (1) 2970 (2) 9293 (2)Mar 24ndash25 24177 (1) nd 3393 (4) nd 18601 (1) ndPremonsoon average 3868 (3) 1521 (4) 7374 (2) 1051 (2) 2201 (3) 3478 (3)Aug 8ndash19 035 (8) 044 (8) 080 (8) 040 (8) 031 (8) 015 (8)Aug 20ndash28 718 (5) 2066 (3) 340 (5) 649 (3) 113 (5) 2067 (5)Aug 29 636 (6) 100 (6) 120 (7) 042 (7) 060 (7) 055 (6)Monsoon average 172 (7) 061 (7) 141 (6) 060 (6) 060 (6) 035 (7)

Here eventsare based on periodsof distinct high and low aerosol concentrationsValues in parenthesesindicate the rankof the mean concentration for this data set There was no precipitation during the Mar 24ndash25 event and this is indicatedby ldquondrdquo Ranks for precipitation concentrationswere assumed to be the same as for aerosol for this event

the relationship between Cliexcl and NaC improved remarkably(r D 083) and the ratio betweenCliexcl and NaC was 06Althoughthis ratio is only half of the mean sea-water ratio it is signi -cantly higher than when all aerosol data is considered Furtherwhen only aerosol data during monsoon precipitationevents areconsidered a similar ratio is obtained while the correlation issigni cantly lower (r D 063)

It is common to observe enrichment of SO2iexcl4 relative to NHC

4in precipitationdue to SO2 oxidationin clouds (eg Calvert et al1985) On the contrary our samples display relative enrichmentof NHC

4 in the precipitation with respect to SO2iexcl4 (Figures 6e

and f and 7a and b (August 9ndash13)) The amount of NHC4 is

more than the sum of SO2iexcl4 and NOiexcl

3 in the precipitation (Fig-ure 7) Similar enrichment was also observed in Hidden Valleyalthough enrichment of NHC

4 occurred to a much higher extent(Shrestha et al 1997) Poor scavengingof SO2iexcl

4 could have beenresponsible for the relative enrichment of NHC

4 A low scaveng-ing ratio of SO2iexcl

4 compared to all other species in our samplessupports thisHowever different scavengingrates shouldnot oc-cur if SO2iexcl

4 and NHC4 are internally mixed Southerly air masses

that bring in precipitation travel over populated areas forestsand cultivated lands in the lower mountain and plains in Nepalbefore reaching the Phortse site These air masses could accu-mulate additional gaseous NH3 emitted from animal manureand fertilizers As there are no industrial activities in this in-termediate region (ca 50 km from the site) additional sourcesof SO2iexcl

4 are insigni cant Low SO2 concentrations measured inrural and semiurban sites support the hypothesis that there are no


Figure 6 Relationshipsbetween water solublemajor ions in aerosol (left) and precipitation (right) Concentrationsin the aerosolare in neq scmiexcl1 and in the precipitation in sup1eq kgiexcl1 Solid lines represent common relationships between species and dottedlines represent the best t line


Figure 7 NHC4 SO2iexcl

4 and NHC4 (SO2iexcl

4 +NOiexcl3 ) ratios in the air and NHC

4 SO2iexcl4 ratios in the precipitation for the premonsoon

sampling (a) and the monsoon sampling (b)


signi cant SO2 (and thus SO2iexcl4 ) sources in Nepal (eg

Carmichael et al 1995) While concentrations of all specieswere elevated in the period between August 20 and 28 this pe-riod is also characterized by enrichment of NHC

4 with respectto SO2iexcl

4 in the aerosol (Figure 7b) supporting the presence ofNH3 in the ambient air

Interspecies Relationships Sources and TransportMultivariate analysis was performed on the aerosol data to

investigate interspecies relations with respect to sources andtransport mechanisms We used multivariate empirical orthogo-nal function (EOF) analysis rather than repeated simple linearregression or multiple linear regression because of its abilityto assess joint behavior of several species (variates) EOF de-composition provides objective representations of multivariatedata by analyzing the covariance structure of its variates (Pixotoand Oort 1992 Mayewski et al 1997 Meeker et al 1997) Thisanalysis was performed on the premonsoon samples and themonsoon samples Aerosol concentrationsbelow detectionwererepresented by the detection limits

The signs and magnitudes of variance explained by EOF1shows that 6 of the major ions (excluding NaC and Cliexcl) arestrongly loaded in both premonsoon and monsoon samples(Table 4) The rst EOF is interpreted to represent well-mixedair masses High loading of 6 species indicates long travel dis-tances from source regions (Heidam1984Mayewski et al 1997Meeker et al 1997) Thermal advection (valley winds) repre-sents the dominant transport mechanism during the premonsoonseasonwhilemonsoonair masses carry pollutantsto the site dur-ing the monsoon season (Shrestha et al 2000) While separateprocesses drive these 2 transport mechanisms their transportpathways are similar in the sense that both ow from the southThis implies that the valley wind system due to thermal advec-

Table 4Result of EOF analysis for aerosol species

Variance explained

Premonsoon Monsoon

EOF1 EOF2 EOF3 EOF1 EOF2 EOF3Species (63) (17) (10) (54) (17) (13)

NaC 21 iexcl55 iexcl10 0 iexcl60 7NHC

4 80 0 8 52 7 36KC 72 3 14 73 0 21Mg2C 93 0 0 90 iexcl2 iexcl3Ca2C 82 0 0 66 iexcl1 iexcl15Cliexcl iexcl7 iexcl69 17 0 iexcl63 1NOiexcl

3 75 0 0 60 iexcl2 iexcl17SO2iexcl

4 90 0 0 90 2 0

Values in parentheses indicate variance explained by the respectiveEOFs Negative signs indicate that the species is loaded opposite in theEOF

tion is relatively extensive and the source regions are similar tothose of large-scale monsoon circulationhence it causes similarvariations in aerosol concentrations

Sodium and Cliexcl dominate the second EOF for both seasonssuggesting that sourcestransport of NaC and Cliexcl are differ-ent from those of other species (Table 4) Premonsoon EOF1explains 21 of variance in NaC while a majority of NaC vari-ance is explained by EOF2 (55) This implies that NaC hasdual sources The loading on EOF1 may support an associationwith a crustal source The crustal source of NaC is insigni -cant during the monsoon season High loading of NaC and Cliexcl

on EOF2 for both seasons is a surprising result Joint loadingof these 2 species is indicative of marine aerosols It is natu-ral to expect some in uence of marine aerosol in the monsoonseason Crustal sources of NaC and Cliexcl (evaporites) from aridregions of central Asia transported by the westerly circulationis precluded because of the lack of Ca2C and Mg2C loadingin this EOF (Wake et al 1994) The occasional precipitationduring winter and premonsoon has its moisture source at theMediterranean Sea (westerly disturbance Mani 1981 Nayava1980) It is possible that the EOF2 during premonsoon is re- ecting the distant maritime origin of moisture transported by awesterly ow Although Mg2C is also a major component of seasalt it is absent in the second EOF possibly because its crustalfraction is much larger than the sea salt fraction

Unlike EOF1 and EOF2 EOF3 is different for the 2 seasonsand shows mixed loading both in sign and magnitude (Table 4)The monsoon loading of NHC

4 and KC may represent a biomassburning signalThe premonsoon is a season when biomass burn-ing is common in the low-lands and hills of Nepal Similar load-ing of Cliexcl could be because a fraction of NHC

4 is associatedwith Cliexcl This can happen if NHC

4 is in excess after completelyneutralizing NOiexcl

3 and SO2iexcl4 and there is gaseous HCl present in

the ambient air (Harrison and Pio 1983)

SUMMARY AND CONCLUSIONSIntensive aerosol sampling during postmonsoon premon-

soon and monsoon seasons at Phortse in the Nepal Himalayasshowed diurnal variations and variations at the time scaleof several samples (ie a week) in the concentrations of wa-ter soluble inorganic chemical species Generally higher con-centrations in the daytime and lower concentrationsin the night-time samples re ect the diurnal variation in the regionalvalley wind systems and also the in uence of scavenging ofparticles by predominantly nighttime precipitation events Thein uence of the former is dominant Additionally distinct vari-ations in aerosol concentrationsoccurring at weekly time scaleswere related to precipitation events This time-scale cleans-ing of atmospheric constituents by precipitation resulted in de-creased concentrations However these precipitation eventsalso coincided with considerable changes in circulationchanges in the source regions associated with the changes inthe circulation may therefore also be affecting the aerosolconcentrations


Event-based precipitation sampling showed that precipita-tion chemistry in general mimics variability in aerosol chem-istry While the aerosol concentrations are not re ected in cor-responding precipitation events on a sample-by-sample basisthe relationship between the ionic concentrations in the 2 me-dia improves as the samples are averaged over several days toa season This supports the potential for using glaciochemicaltime series from the Himalayas to infer the atmospheric compo-sition of the past in that region Scavenging ratios were found tobe highly variable and the mean values were relatively high formost species exceptSO2iexcl

4 High scavengingratios for NOiexcl3 sug-

gest enrichment in precipitation due to the presence of gaseousHNO3 in the air Similarly Cliexcl is also enriched in precipita-tion compared to aerosol and in precipitation it shows a closerelationship with NaC This is attributed to the scavenging ofHCl present in the ambient air likely of the same source asNaC (ie sea salt particles fractionated due to reaction withH2SO4 or HNO3) In addition low fractionationof precipitatingair masses due to vigorous circulation is also likely responsiblefor the apparent enhancementof Cliexcl with respect to NaC in pre-cipitation Enrichment of NHC

4 in the precipitation with respectto SO2iexcl

4 is attributed to higher solubility of gaseous NH3 com-pared to SO2iexcl

4 Furthermore similar enrichment in the aerosolduring a dry episode in the monsoon sampling is attributed tothe presence of NH3 in the ambient air from areas in the neigh-borhood of the sampling site The present data sets are too lim-ited to suggest the observed relationships as conclusive Con-tinued observations of synchronous aerosol and precipitationchemistry in the future may substantiate such relationships en-abling us to better interpret the glaciochemical data from theHimalayas

EOF analysis shows that there is a dominant transport mech-anism controlling aerosol concentrations of all species exceptNaC and Cliexcl both during premonsoon and monsoon seasonsWhile large-scale thermal winds and monsoon circulation arethe predominant transport mechanism operating in the premon-soon and monsoonseasons respectivelyremarkable similaritiesbetween EOF1 for both seasons suggest that both transportmechanisms have similar pathways and incorporate pollutantsfrom similar source regions EOF2 during the monsoon repre-sents marine contributions to the aerosol transported bysoutheasterly monsoon circulation while during premonsoonEOF2 likely depicts distant maritime sources transported bywesterly disturbances These results are promising concerningthe use of glaciochemicaldata from the Himalayas to infer varia-tions in continental contribution to the atmospheric compositionof the regions and variation in relative strengths of monsoon andwesterly circulation

REFERENCESBaltensperger U Schwikowski M Gaggeler H W Jost D T Beer J

Siegenthaler U Wagenbach D Hoffmann H J and Synal H A (1993)Transfer of Atmospheric Constituents into an Alpine Snow Field AtmosEnviron 271881ndash1890

Barrie L A (1985) Atmospheric Particles Their Physical and Chemical Char-acteristics and Deposition Processes Relevant to the Chemical Compositionof Glaciers Annals of Glaciology 7100ndash108

Calvert J K Lazrus A Kok G L Heikes B G Walega J H Lind Jand Cantrell C A 1985 Chemical Mechanisms of Acid Generation in theTroposphere Nature 31727ndash35

Carmichael G R Ferm M Adhikary S Ahmed J Mohan M HongM S Chen L Fook L Liu C M SoedomoM Tran G Sukomasnk KZhao D Arndt R and Chen L L (1995) Observed Regional Distributionof Sulfer Dioxide in Asia Water Air and Soil Pollution 852289ndash2294

Davidson C I (1989) Mechanisms of Wet and Dry Deposition of AtmosphericContaminants to Snow Surfaces Dahlem Workshop on the EnvironmentalRecord in Glaciers March 1988 Dahlem

Davidson C I Jaffrezo J-L Mosher B W Dibb J E Borys R DBodhaine B A Rasmussen R A Boutron C F Gorlach U CachierH Ducret J Colin J-L Heidam N Z Kemp K and Hillamo R (1993)Chemical Constituents in the Air and Snow at Dye 3 Greenland II Analysisof Epidodes in April 1989 Atmos Environ 27A2723ndash2738

De Angelis M Barkov N I and Petrov V N (1987) Aerosol Concentrationsover the Last Climatic Cycle (160 kyr) from an Antarctic Ice Core Nature325318ndash321

Delmas R J and Legrand M (1989) Long-Term Changes in the Concentra-tions of Major Chemical Compounds (Solubleand Insoluble) along Deep IceCores In The Environmental Record in Glaciers and Ice Sheets edited byJ H Oeschger and C C Langway John Wiley amp Sons pp 319ndash341

Drever J I (1988) The Geochemistry of Natural Waters Prentice HallEnglewood Cliffs NJ

Galloway J N (1978) The Collection of Precipitation for Chemical AnalysisTellus 3071ndash82

Haagenson P and ShapiroM A (1979) IsentropicTrajectories forDerivationof Objectively Analyzed Meteorological Parameters NCAR Technical NoteNCARTN-149 C STR Atmospheric Quality Division National Center forAtmospheric Research

Harrison R M and Plo C A (1983) Major Ion Composition and Chemi-cal Associations of Inorganic Atmosphreic Aerosols Environ Sci Technol17169ndash174

Hayes D Snetsinger K Ferry G Oberbeck V and Farlow N (1980)Reactivity of Stratospheric Aerosols to Small Amounts of Ammonia in theLaboratory Environment Geophys Res Lett 7974ndash976

Heidam N Z (1984) The Components of the Arctic Aerosol Atmos Environ18329ndash343

Hindman E E (1995a) Air Motions in the Kumbu Himalaya and PossibleSoaring Flights Technical Soaring XIX3ndash8

Hindman E E (1995b) Himalayan Banner Clouds and Thermals SeventhConference on Mountain Meteorology American Meteorological SocietyBoston MA

Hindman E E (1998) Personal Communication City College of New YorkNew York

Hindman E E Upadhyay B P Maharjan S and Aryal D (1996) A Studyof the Meteorology and Combustion Particle Transport in the Dudh KosiValley of the Khumbu Himalaya Report from TU-CCNY Project CooperativeStudies in Mountain Meteorology Transport in Mountain-Induced Air owsTribhuvan UniversityCity College of New York

Jouzel J Petit J R and Raynaud D (1990) Palaeoclimatic Information fromIce Cores The Vostok Records Earth Sciences 81349ndash355

Liu B Y H Pui D Y H and Rubow K L (1984) Characteristics ofAir Sampling Filter Media In Aerosols in the Mining and Industrial WorkEnvironments Vol 3 Instrumentation edited by V A Marple and B Y HLiu Ann Arbor Science Ann Arbor MI

Lyons W B Wake C and Mayewski P A (1991) Chemistry of Snow fromHigh Altitude MidLow Latitude Glaciers In Seasonal Snowpacks edited byT D Davies Vol G 28 Springer-Verlag Berlin pp 359ndash383

Mani A (1981) The Climate of Himalaya In The Himalaya Aspects ofChanges edited by J S Lall and A D Moddie Oxford University PressNew Delhi pp 3ndash15


Mayewski P A Lyons W B and Ahmad N (1983) Chemical Compositionof a High Altitude Fresh Snowfall in the Ladakh Himalayas Geophys ResLett 10105ndash108

Mayewski P A Meeker L D Morrison M C Twickler M S WhitlowS I Ferland K K Meese D A Legrand M R and Steffensen J P(1993) Greenland Ice Core ldquoSignalrdquo Characteristics An Expanded View ofClimate Change J Geophys Res 9812839ndash12847

Mayewski P A Meeker L D Twickler M S Whitlow S Yang Q andLyons W B (1997) Major Features and Forcing of High-Latitude NorthernHemisphere Atmospheric Circulation using a 110000 Year-Long Glacio-chemical Series J Geophys Res 10226345ndash26366

Mayewski P A Meeker L D Whitlow S Twickler M S MorrisonM C Bloom eld P Bond G C Alley R B Gow A J Grootes PM Meese D A Ram M Taylor K C and Wumkes W (1994) Changesin Atmospheric Circulation and Ocean Ice Cover over the North Atlanticduring the Last 41000 Years Science 2631747ndash1751

Mayewski P A TwiklerMSWhitlowS I Meeker LD YangQ ThomasJ Kreutz K Grootes P M Morse D L Steig E J Waddlington E DSaltzman E D Whung P-Y and Tylor K C (1996) Climate Changeduring the Last Deglaciation in Antarctica Science 2721636ndash1638

Meeker L D Mayewski P A Twikler M S and Whitlow S I (1997)A 110000-Year History of Change in Continental Biogenic Emission andRelated Atmospheric Circulation Inferred from the Greenland Ice SheetProject Ice Core J Geophys Res 10226489ndash26504

Nakajima C (1976) Movement and Development of the Clouds over KhumbuHimal in Winter Seppyo (Journal of the Japanese Society of Snow and Ice)3889ndash92

Nayava J L (1980) Rainfall in Nepal The Himalayan Review 121ndash18

Pixoto J P and Oort A H (1992) Physics of Climate American Institute ofPhysics New York

Shrestha A B Wake C P and Dibb J E (1997) Chemical Compositionof Aerosol and Snow in the High Himalaya during the Summer MonsoonSeason Atmos Environ 312815ndash2826

Shrestha A B Wake C P Dibb J E Wayewski P A Whitlow S ICarmichael G R and Ferm M (2000) Seasonal Variations in AerosolConcentrations and Compositions in the Nepal Himalaya Atmos Environ343349ndash3363

Silvente E and Legrand M (1993) Ammonium to Sulphate Ratio in Aerosoland Snow of Greenland and Antarctic Regions J Geophys Res 687ndash690

Talbot R W Harris R C Browell E V Gregory G L Sebacher D I andBeck S M (1986)Distributionand Geochemistry ofAerosols in the TropicalNorth Atalantic Troposphere Relationship to Saharan Dust J Geophys Res915173ndash5182

Thompson L G Mosley-Thompson E Grootes P M Pourchet M andHastenrath S (1984) Tropical Glaciers Potential for Ice Core PaleoclimaticReconstructions J Geophys Res 89 4638ndash4646

Wake C P Dibb J E Mayewski P A Zhongqin L and Zichu X (1994)The Chemical Composition of Aerosols over the Eastern Himalayasand Tibetan Plateau during Low Dust Periods Atmos Environ 28695ndash704

Warneck P (1988) Chemistry of Natural Atmosphere Academic Press IncSan Diego pp 757


marked seasonal variations in the aerosol concentrations dueto changes in circulation and associated precipitation scaveng-ing As part of the same program aerosol samples were col-lected over 6 and 12 h sample intervals at the Himalayan siteduring 3 different seasons premonsoon monsoon and post-monsoonAerosol samplingwas complementedby precipitationsamples This paper investigates the variations in aerosol con-centrations during the intensive sampling programs and theirconnection with meteorological variables and transport mecha-nisms A preliminary attempt is made to quantify the relation-ship between aerosol concentrations and concentrations in theprecipitation

EXPERIMENTAL

The SiteThe sampling site was located near the remote village of

Phortse in the Khumbu region of the Himalayas in Nepal

(a)

Figure 1 Location map of Phortse sampling station The arrows show predominant daytime local wind directions (Continued)

(Figure 1) The sampling site stands at an elevation of 4100 masl at 27plusmn510 N latitude and 86plusmn450 E longitude and is situatedabout 350 m above the village close to where a northwest tosoutheast trending ridge merges with the mountain slope Theridge continues down to the con uence of the Imja Khola andDudh Kosi rivers at an elevation of 3550 m asl A meteoro-logical station was established 50 m below the aerosol samplingsite on the same ridgeAlthoughthe predominantwind directionover the valley is southwesterly observations of the local windreveal that wind near the sampling station blows predominantlyfrom a south-south-east direction (across the ridge) Thereforethe sampling site is well protected from sources of pollution inthe village The village of Phortse consists of about 60 perma-nent houses There are small patches of farmland around the vil-lage where potato and buckwheat are cultivated between Apriland September There are also some forested areas dominatedby juniper and different varieties of pine A small airstrip atSyangboche is located at a distance of about a 4 h walk


(b)




















3 SO2iexcl4































Fig

ure

3D

aytim

ean

dni

ghtt

ime

conc

entr

atio

nsof

maj

orio

nic

spec

ies

inae

roso

lsam

ples

colle

cted

atPh

orts

eD

ays

with

outc

onse

cutiv

eda

yan

dni

ghts

ampl

esar

eno

tpre

sent

edin

the

gur

eD

aytim

ean

dni

ghtti

me

prec

ipita

tion

even

tsar

ein

dica

ted

byldquod

rdquoan

dldquon

rdquoin

(a)

and

(d)

whe

reas

ldquodnrdquo

indi

cate

spr

ecip

itatio

nev

ents

cove

ring

both

the

dayt

ime

and

the

follo

win

gni

ght

Not

eth

edi

ffer

ence

inve

rtic

alsc

ales

447

Fig

ure

4M

eteo

rolo

gica

ldat

aco

llect

edin

Phor

tse

duri

ngpr

emon

soon

199

7(l

eft)

and

mon

soon

199

7(r

ight

)H

ere

we

use

win

ddi

rect

ion

data

from

Din

gboc

hefo

rmon

soon

199

7as

ther

eis

now

ind

dire

ctio

nda

tafr

omPh

orts

efo

rtha

tper

iod

Poin

tsin

win

ddi

rect

ion

plot

sin

dica

teho

urly

data

whi

lea

solid

line

isw

eigh

ted

mea

nw

ith

5sm

ooth

ing

fact

or

448








W Dfrac12aCs

Ca










Fig

ure

5G

eopo

tent

ialh

eigh

tsat

500

mb

leve

lT

heda

shed

arro

ws

indi

cate

tent

ativ

eis

obar

ictr

ajec

tory

ofai

rm

asse

sar

rivi

ngat

the

site

450




3 SO2iexcl4

Mean 1393 1811 2436 2580 1805 2582 2502 487S Dev 1593 2351 2092 2277 1873 1870 2007 326








NHC4 SO2iexcl

4 Ca2C



































Variance explained

Premonsoon Monsoon





4 90 0 0 90 2 0






























































(b)




















3 SO2iexcl4































Fig

ure

3D

aytim

ean

dni

ghtt

ime

conc

entr

atio

nsof

maj

orio

nic

spec

ies

inae

roso

lsam

ples

colle

cted

atPh

orts

eD

ays

with

outc

onse

cutiv

eda

yan

dni

ghts

ampl

esar

eno

tpre

sent

edin

the

gur

eD

aytim

ean

dni

ghtti

me

prec

ipita

tion

even

tsar

ein

dica

ted

byldquod

rdquoan

dldquon

rdquoin

(a)

and

(d)

whe

reas

ldquodnrdquo

indi

cate

spr

ecip

itatio

nev

ents

cove

ring

both

the

dayt

ime

and

the

follo

win

gni

ght

Not

eth

edi

ffer

ence

inve

rtic

alsc

ales

447

Fig

ure

4M

eteo

rolo

gica

ldat

aco

llect

edin

Phor

tse

duri

ngpr

emon

soon

199

7(l

eft)

and

mon

soon

199

7(r

ight

)H

ere

we

use

win

ddi

rect

ion

data

from

Din

gboc

hefo

rmon

soon

199

7as

ther

eis

now

ind

dire

ctio

nda

tafr

omPh

orts

efo

rtha

tper

iod

Poin

tsin

win

ddi

rect

ion

plot

sin

dica

teho

urly

data

whi

lea

solid

line

isw

eigh

ted

mea

nw

ith

5sm

ooth

ing

fact

or

448








W Dfrac12aCs

Ca










Fig

ure

5G

eopo

tent

ialh

eigh

tsat

500

mb

leve

lT

heda

shed

arro

ws

indi

cate

tent

ativ

eis

obar

ictr

ajec

tory

ofai

rm

asse

sar

rivi

ngat

the

site

450




3 SO2iexcl4

Mean 1393 1811 2436 2580 1805 2582 2502 487S Dev 1593 2351 2092 2277 1873 1870 2007 326








NHC4 SO2iexcl

4 Ca2C



































Variance explained

Premonsoon Monsoon





4 90 0 0 90 2 0






































































3 SO2iexcl4































Fig

ure

3D

aytim

ean

dni

ghtt

ime

conc

entr

atio

nsof

maj

orio

nic

spec

ies

inae

roso

lsam

ples

colle

cted

atPh

orts

eD

ays

with

outc

onse

cutiv

eda

yan

dni

ghts

ampl

esar

eno

tpre

sent

edin

the

gur

eD

aytim

ean

dni

ghtti

me

prec

ipita

tion

even

tsar

ein

dica

ted

byldquod

rdquoan

dldquon

rdquoin

(a)

and

(d)

whe

reas

ldquodnrdquo

indi

cate

spr

ecip

itatio

nev

ents

cove

ring

both

the

dayt

ime

and

the

follo

win

gni

ght

Not

eth

edi

ffer

ence

inve

rtic

alsc

ales

447

Fig

ure

4M

eteo

rolo

gica

ldat

aco

llect

edin

Phor

tse

duri

ngpr

emon

soon

199

7(l

eft)

and

mon

soon

199

7(r

ight

)H

ere

we

use

win

ddi

rect

ion

data

from

Din

gboc

hefo

rmon

soon

199

7as

ther

eis

now

ind

dire

ctio

nda

tafr

omPh

orts

efo

rtha

tper

iod

Poin

tsin

win

ddi

rect

ion

plot

sin

dica

teho

urly

data

whi

lea

solid

line

isw

eigh

ted

mea

nw

ith

5sm

ooth

ing

fact

or

448








W Dfrac12aCs

Ca










Fig

ure

5G

eopo

tent

ialh

eigh

tsat

500

mb

leve

lT

heda

shed

arro

ws

indi

cate

tent

ativ

eis

obar

ictr

ajec

tory

ofai

rm

asse

sar

rivi

ngat

the

site

450




3 SO2iexcl4

Mean 1393 1811 2436 2580 1805 2582 2502 487S Dev 1593 2351 2092 2277 1873 1870 2007 326








NHC4 SO2iexcl

4 Ca2C



































Variance explained

Premonsoon Monsoon





4 90 0 0 90 2 0



















































































Fig

ure

3D

aytim

ean

dni

ghtt

ime

conc

entr

atio

nsof

maj

orio

nic

spec

ies

inae

roso

lsam

ples

colle

cted

atPh

orts

eD

ays

with

outc

onse

cutiv

eda

yan

dni

ghts

ampl

esar

eno

tpre

sent

edin

the

gur

eD

aytim

ean

dni

ghtti

me

prec

ipita

tion

even

tsar

ein

dica

ted

byldquod

rdquoan

dldquon

rdquoin

(a)

and

(d)

whe

reas

ldquodnrdquo

indi

cate

spr

ecip

itatio

nev

ents

cove

ring

both

the

dayt

ime

and

the

follo

win

gni

ght

Not

eth

edi

ffer

ence

inve

rtic

alsc

ales

447

Fig

ure

4M

eteo

rolo

gica

ldat

aco

llect

edin

Phor

tse

duri

ngpr

emon

soon

199

7(l

eft)

and

mon

soon

199

7(r

ight

)H

ere

we

use

win

ddi

rect

ion

data

from

Din

gboc

hefo

rmon

soon

199

7as

ther

eis

now

ind

dire

ctio

nda

tafr

omPh

orts

efo

rtha

tper

iod

Poin

tsin

win

ddi

rect

ion

plot

sin

dica

teho

urly

data

whi

lea

solid

line

isw

eigh

ted

mea

nw

ith

5sm

ooth

ing

fact

or

448








W Dfrac12aCs

Ca










Fig

ure

5G

eopo

tent

ialh

eigh

tsat

500

mb

leve

lT

heda

shed

arro

ws

indi

cate

tent

ativ

eis

obar

ictr

ajec

tory

ofai

rm

asse

sar

rivi

ngat

the

site

450




3 SO2iexcl4

Mean 1393 1811 2436 2580 1805 2582 2502 487S Dev 1593 2351 2092 2277 1873 1870 2007 326








NHC4 SO2iexcl

4 Ca2C



































Variance explained

Premonsoon Monsoon





4 90 0 0 90 2 0





























































Fig

ure

4M

eteo

rolo

gica

ldat

aco

llect

edin

Phor

tse

duri

ngpr

emon

soon

199

7(l

eft)

and

mon

soon

199

7(r

ight

)H

ere

we

use

win

ddi

rect

ion

data

from

Din

gboc

hefo

rmon

soon

199

7as

ther

eis

now

ind

dire

ctio

nda

tafr

omPh

orts

efo

rtha

tper

iod

Poin

tsin

win

ddi

rect

ion

plot

sin

dica

teho

urly

data

whi

lea

solid

line

isw

eigh

ted

mea

nw

ith

5sm

ooth

ing

fact

or

448








W Dfrac12aCs

Ca










Fig

ure

5G

eopo

tent

ialh

eigh

tsat

500

mb

leve

lT

heda

shed

arro

ws

indi

cate

tent

ativ

eis

obar

ictr

ajec

tory

ofai

rm

asse

sar

rivi

ngat

the

site

450




3 SO2iexcl4

Mean 1393 1811 2436 2580 1805 2582 2502 487S Dev 1593 2351 2092 2277 1873 1870 2007 326








NHC4 SO2iexcl

4 Ca2C



































Variance explained

Premonsoon Monsoon





4 90 0 0 90 2 0




































































W Dfrac12aCs

Ca










Fig

ure

5G

eopo

tent

ialh

eigh

tsat

500

mb

leve

lT

heda

shed

arro

ws

indi

cate

tent

ativ

eis

obar

ictr

ajec

tory

ofai

rm

asse

sar

rivi

ngat

the

site

450




3 SO2iexcl4

Mean 1393 1811 2436 2580 1805 2582 2502 487S Dev 1593 2351 2092 2277 1873 1870 2007 326








NHC4 SO2iexcl

4 Ca2C



































Variance explained

Premonsoon Monsoon





4 90 0 0 90 2 0





























































Fig

ure

5G

eopo

tent

ialh

eigh

tsat

500

mb

leve

lT

heda

shed

arro

ws

indi

cate

tent

ativ

eis

obar

ictr

ajec

tory

ofai

rm

asse

sar

rivi

ngat

the

site

450




3 SO2iexcl4

Mean 1393 1811 2436 2580 1805 2582 2502 487S Dev 1593 2351 2092 2277 1873 1870 2007 326








NHC4 SO2iexcl

4 Ca2C



































Variance explained

Premonsoon Monsoon





4 90 0 0 90 2 0
































































3 SO2iexcl4

Mean 1393 1811 2436 2580 1805 2582 2502 487S Dev 1593 2351 2092 2277 1873 1870 2007 326








NHC4 SO2iexcl

4 Ca2C



































Variance explained

Premonsoon Monsoon





4 90 0 0 90 2 0














































































Variance explained

Premonsoon Monsoon





4 90 0 0 90 2 0





























































aerosol and precipitation chemistry at a remote himalayan

Documents