Air particulate matter pollution in Ulaanbaatar, Mongolia: determination of composition, source contributions and source locations

Atmospheric Pollution Research

www.atmospolres.com

Air

particulate

matter

pollution

in

Ulaanbaatar,

Mongolia:

determination

of

composition,

source

contributions

and

source

locations

Perry

K.

Davy

,

Gerelmaa

Gunchin

,

Andreas

Markwitz

,

William

J.

Trompetter

,

Bernard

J.

Barry

,

Dagva

Shagjjamba

,

Sereeter

Lodoysamba

1

InstituteofGeologicalandNuclearSciences,NationalIsotopeCentre,P.O.Box31–312,LowerHutt,NewZealand

2

NationalUniversityofMongolia,IkhSurguuliingudamj–1,Ulaanbaatar,Mongolia

ABSTRACT

Ulaanbaatar,thecapitalcityofMongoliaissubjecttohighairparticulatematterpollutionepisodesduringwinterand duringduststormeventsinspringandautumnthathavesevereimplicationsforthehealthoftheexposedpopulation. Thispaperpresentstheresultsoffine(PM2.5)andcoarse(PM10–2.5)particulatemattermonitoringinUlaanbaatarfrom 2004to2008andreceptormodellingtodeterminethesourcescontributingtoparticulatematterpollution.IonBeam Analysiswasusedtodetermineelementalconcentrationsinthetwosizefractionsandblackcarbonwasmeasured withalightreflectancedevice.Masscontributionstoambientparticleconcentrationsfromemissionsourceswere estimatedbypositivematrixfactorisationandairmassback–trajectoryanalysiswasusedtoassessprobablesource locations.Theresultsshowthatcrustalmattersourcesaretheprimarycontributorstothecoarseparticlefraction. Combustionsources(coalcombustion,biomassburning,andmotorvehicles)dominatethefinefractionofparticulate matterintheUlaanbaatarairshed,primarilyfromlocalemissionsourcesbutforestfirestothenorthcanbea significantcontributortobiomassburningconcentrationsattimes.Analysisofseasonaldifferencesshowedthatcoal combustionprocesseswerelargelyresponsibleforfineparticleairpollutionepisodesduringwinter.Temporaltrends showanincreaseinthecoalcombustioncontributionsoverthemonitoringperiod.Wesuggestthatthisislinkedto theincreaseintheUlaanbaatarpopulationandaconsequentincreaseintheuseofcoalforpowergenerationand domesticheatingpurposes.

Keywords: Airpollution Particulatematter IBA Positivematrixfactorisation Receptormodelling Mongolia ArticleHistory: Received:27May2010 Revised:25August2010 Accepted:09September2010 CorrespondingAuthor: PerryK.Davy Tel:+64Ͳ4Ͳ570Ͳ4688 Fax:+64Ͳ4Ͳ570Ͳ4657 EͲmail:[email protected] ©Author(s)2011.ThisworkisdistributedundertheCreativeCommonsAttribution3.0License. doi:10.5094/APR.2011.017

1.

Introduction

This paperpresentstheresults ofair particulatematter

samplingfromOctober2004toApril2008atUlaanbaatar,the

capitalcityofMongolia(Figure1).ThemonitoringinUlaanbaatar

waspartofawiderprogrammerunundertheauspicesofthe

InternationalAtomicEnergyAgency(IAEA)(Markwitz2005;Hopke

etal.,2008).Ulaanbaatar(population1million)issituatedina

largedomevalleysurroundedbymountainsandsubjecttofreezing

conditions and temperature inversions during winter months

(December–February). This work presents PM2.5 and PM10–2.5 monitoringresultsandreceptormodellingusingpositivematrix

factorisation(PMF)toapportionsourcecontributionstoeachof

thesizefractions.AirparticulatematterpollutioninUlaanbaatar

canbeextremeattimeswithbothPM2.5orPM10–2.5exceeding 200μgmо3.

Anumberofemissioninventorieshavebeenpreparedforthe

Ulaanbaatarairshed(NAMHEM,2007;Guttikunda,2007)inorder

to quantifysourceemissionsandcontributions toparticulate

matterairpollutioninthecity.However,somesourcessuchas

fugitivedustfromduststormsandlocaldustgeneratingactivities

oremissionsfromdomesticuseoffuelsforheatingandcooking

aredifficulttoquantifyperseortheemissionsfactorsandactivity

estimateshaveahighuncertaintyassociated(>50%).Thereceptor

modellingapproach fillsthese knowledgegapsand provides

informationonthekeysourcescontributingtoparticulatematter

concentrationsinUlaanbaatar.

2.

Experimental

2.1.Samplecollectionandanalyticalmethods

PM2.5andPM10–2.5sampleswerecollectedonatwo–day–per– weekbasis(MondaysandWednesdays)fromOctober2004toApril

2008attheNationalUniversityofMongoliaairqualitymonitoring

site,4kilometerseastofthecentreofUlaanbaatarasshownin

Figure1b(latitude47914deg,longitude106972deg,1300m

abovesealevel).Thefineandcoarseparticlesizefractionswere

collectedwithaGENTsampler(Maenhautetal.,1993).The

samplerconsistsofaPM10impactor–typesizeselectiveinletand stackedfilterunitassemblyconnectedtoapumpandgasmeter.

Thestackedfilterunit(SFU)ismadeupoftwofiltersinseries,the

topfilter(polycarbonate)collectstheparticulatesizefraction

between10micronsand2.5micronsinaerodynamicdiameter

(PM10–2.5orcoarsefraction),thebottomfilter(polycarbonate) collectsparticulatematter2.5micronsandlessinaerodynamic

diameter(PM2.5orfinefraction).TheperformanceoftheSFUand thesizeselectiveinlethasbeenassessedagainstotherparticle

Figure1.Mapshowing(a)Ulaanbaatar,Mongoliaand(b)locationofmonitoringsitein Ulaanbaatar(Ì).Source:GoogleEarth(www.google.com).

werecollectedfora24hourperiod,butonsomedaysitwasnot

possibletocollectthesampleovertheentire24hoursasthehigh

particulatematterconcentrationsleadtofilterclogging.Thus,the

samplerwasoperatedalternatingtimeonandtimeoff(e.g.,1hon

followedby1hoff)overthecourseofthe24hoursperiodto

providearepresentativesampleforthatday.Gravimetricanalysis

was carried out at the National University of Mongolia,

Ulaanbaatarandresultsreportedasmicrogramsofparticulate

matterpercubicmetreofairsampled(LODa5μgmо3;uncertainty

estimatedat±10%).

IonBeamAnalysis(IBA)wasusedtomeasuretheconcenͲ

trationsofelementswithatomicnumberaboveneoninparticulate

matteronthesize–resolvedfiltersamplesfromUlaanbaatar.IBA

measurementsforthisstudywerecarriedoutattheNewZealand

IBAFacilityoperatedbytheInstituteofGeologicalandNuclear

Sciences(GNS)inGracefield,LowerHutt(Trompetteretal.,2005).

Blackcarbon(BC)concentrationsonfiltersweredetermined

bylightreflectionusingaM43DDigitalSmokeStainReflectometer.

Blackcarbonhasbeenstudiedextensivelybutitisstillnotclearto

whatextentitiseitherelementalcarbon[graphitic,C(0)],high

molecularweightrefractoryweightorganicspeciesoracomͲ

binationofboth(Jacobsonetal.,2000).Forcombustionsources

suchaspetrolanddieselfuelledvehiclesandbiomassburning

elemental(orgraphiticcarbon)andorganiccarboncompounds

(OC)aretheprincipleaerosolcomponentsemitted(Jacobsonet

al.,2000;Fineetal.,2001;Watsonetal.,2002;Salmaetal.,2004).

Theabsorptionandreflectionofvisiblelightonparticlesinthe

atmosphereorcollectedonfiltersisdependentontheparticle

concentration,density,refractiveindexandsize.Foratmospheric

particles,BCisthemosthighlyabsorbingcomponentinthevisible

lightspectrumwithverymuchsmallerabsorptioncomingfrom

soil,sulphateandnitrate(Horvath,1993;Horvath,1997).Hence,to

thefirstorderitisassumedforthepurposesofthisstudythatall

thelightabsorptionbyparticlesonfiltersisduetoBC.

2.2.Receptormodelling

ReceptormodellingandapportionmentofPMmassbyPMF

wasperformedusingthePMF2program(Paatero,1997).With

PMF, sources are constrained to have non–negative species

concentrations,nosamplecanhaveanegativesourcecontribution

anderrorestimatesforeachobserveddatapointareusedas

point–by–pointweights.Thisfeatureisadistinctadvantage,inthat

itcanaccommodatemissingandbelowdetectionlimitdatathatis

acommonfeatureofenvironmentalmonitoringresults(Songet

al.,2001).Thesignal–to–noiseratioforanindividualelemental

measurementcanhaveasignificantinfluenceonareceptormodel

andmodellingresults.Fortheweakest(closesttodetectionlimit)

speciesthevariancemaybeentirelyduetonoise(Paateroand

Hopke,2002).PaateroandHopke(2003)stronglysuggestdown– weightingordiscardingnoisyvariables(asmeasuredbytheir

signal–to–noiseratio)thatarealwaysbelowtheirdetectionlimitor

speciesthathavealotofuncertaintyintheirmeasurements

relativetothemagnitudeoftheirconcentrations.Thereforethe

datawerescreenedbytheirsignal–to–noiseratio(S/Nratio).

VariableswithverylowS/Nratios(ч0.2)wereexcludedfromthe

analysis, while weak variables (0.2чS/Nч2) were down– weighted.Rotationalfreedominsolutionswereexploredand

controlledusingFPEAK(PaateroandHopke,2002)andobserving

theeffectontheQvalues(chi–squared),G–vectorplotsand

residualplots(Paateroetal.,2005).Onlythosesolutionsthatcould

berelatedtophysicalsourceswereconsideredacceptable.

2.3.Backtrajectoriesandpotentialsourcecontributionfunction

analysis

TheHYSPLIT(HYbrid Single–ParticleLagrangianIntegrated

Trajectory)modelisthenewestversionofacompletesystemfor

computingsimpleairparceltrajectoriestocomplexdispersionand

depositionsimulations(DraxlerandRolph,2003).Usingaglobal

meteorologicaldatasetHYSPLITcomputestheadvectionofasingle

pollutantparticle,orsimplyitstrajectory.Forthepurposeofthis

study,HYSPLIThasbeenusedtocalculatethe96–hourback

trajectoriesofairparcelsforsampledaysofinterestinorderto

examine long–range atmospheric transport processes and

determine potential particulate matter source locations. The

potentialsourcecontributionfunction(PSCF)analysiscalculates

theprobablesourcelocationsforthosesourcesconsideredtobe

impacting at a receptor site from long–range transport of

pollutants.Ifatrajectoryendpointliesintheijthcell,theairparcel

assumestocollectPMemittedinthecellandtransportsalongthe

trajectory to the monitoring site. PSCFij is the conditional

probabilitythatanairparcelthatpassedthroughtheijthcellhada

highconcentrationuponarrivalatthemonitoringsite:

ܲܵܥܨ௜௝ൌ

݉௜௝

݊௜௝ (1)

wherenijisthetotalnumberofendpointsthatfallintheijthcell,

mijisthenumberofendpointsthatexceededthethreshold

criterion(inthisstudy,theaverageconcentrationofeachspecies

wasusedforthethresholdcriterion).

PSCFdescribesthespatialdistributionofprobablegeographiͲ

calsourcelocations.GridcellswhichhavehighPSCFvaluesarethe

potentialsourceareaswhoseemissionsarelikelytocontributeto

particulatematterconcentrationsatthemonitoringsite.Fora

secondaryaerosolpollutant,ahighPSCFareamayalsoinclude

areaswheresecondaryparticleformationisenhanced(Begumet

al.,2005).

3.

Results

and

Discussion

3.1.ParticulatematterconcentrationsinUlaanbaatar

Coarseandfineparticleconcentrationswerecombinedto

provideameasureofambientPM10concentrationsatUlaanbaatar (Figure2a).Particulatematterconcentrationswerefoundtobe

dominatedbythecoarsefraction(Figure2b)whichwerehigher

duringspring(March–May)andautumn(September–NovemͲ

ber).Fineparticleconcentrationspeakedduringwinter(December

–February).Notethatnomonitoringwascarriedoutduringthe

summer(April–August)of2005.

Particulatematterconcentrationswerefoundtobeextremely

highattimeswithmaximumPM10concentrationsmeasuredat 1100μgmо3andmanydayswhereconcentrationswereabove

300μgmо3,however,theremaybesomeuncertaintyintheGENT

gravimetric mass at such extreme concentrations (i.e.

>600μgmо3)duetothepotentialforfilterclogging.Thecoarse

particlefractiongenerallydominatedPM10buthighPM2.5values wererecordedduringwintermonthswithPM2.5valuesoftenover 100μgmо3.Withsuch highparticulatematterconcentrations

presentintheUlaanbaatarurbanareatherearelikelytobe

significanthealthimplicationsfortheexposedpopulation.

3.2.PM102.5dataanalysis

Atotalof224(91%oftotalsamplescollected)particulate

mattersampleswereincludedinthereceptormodellinganalyses

ofUlaanbaatarcoarseparticledata.Whilemorethan25elemental

constituentsweredetectedabovetheirrespectiveLOD,16species

wereselectedforPMFmodelling(PM102.5,BC,Na,Mg,Al,Si,P,S, Cl,K,Ca,Ti,Mn,Fe,Cu,Zn)asotherelementswereonlypresentin

afewsamplesand/ortheS/Nratiowaslowerthan0.2.Table1

presents the mean, standard deviation, median, maximum,

minimum,andnumberofsamplesaboveLODandS/Nratiofor

individualspecies.

Differentsourcenumbers(3Ͳ9)wereusedforderivingPMF

solutionsandtheeffectsontheQvalues(andQvalueasafunction

ofFPEAK),residualmatricesandGͲspaceplotswereobservedto

determinetheoptimalnumberofprimarycontributingsourcesto

ambientcoarseparticleconcentrations.Themeasuredcoarsemass

wasincludedasanindependentvariableinPMFmodellingto

circumventtheneedformultiplelinearregressiontoapportion

mass(KimandHopke,2006).Thederivedsourceprofileswere

comparedtomeasuredsourceprofilesand/orthosereportedin

other studies to ensure they made physical sense. A fourͲ

sourcemodelwasfoundtobethemostappropriatesolution

(FPEAK=Ͳ0.2)which,onaverageexplained86%ofthegravimetric

mass.Figure3showsthecomparison(r2=0.88)betweenpredicted

(PMF) PM102.5 and measured (gravimetric) PM102.5 concenͲ trations.Theaveragemasscontributionsofeachofthesourcesto

ambientcoarseparticleconcentrationsareshowninTable2and

theelementalsourceprofilesarepresentedinFigure4.

The first source profile was identified as airborne soil

originatingfromcrustalmatterastheprofileisdominatedbyAl

andSialongwithCa,TiandFe.TheSi:Alratiowasfoundtobe

2.6:1,typicalofcrustalmattercomposition(aluminosilicates)(Lide,

1992).Asecondcrustalmattersource,labelledasSoil2,hasalso

beenidentifiedasoriginatingfromcrustalmatterandtheprimary

distinctionbetweenthetwosoilprofilesisthattheSoil2source

hasasignificantlyhigherBCcomponent.Wesuggestthatthe

differencebetweenthetwocrustalmattersourcesismostlikely

Figure2.ParticulatematterconcentrationsatUlaanbaatar:(a)PM10;(b)PM2.5;(c)PM102.5.

Table1.SummarystatisticsforPM102.5andelementalconcentrationsatUlaanbaatar

ngmͲ3 ArithmeticMean StdDev Median Maximum Minimum LOD #Samples>LODa S/N

PM102.5 143000 122000 110000 955000 5000 Ͳ Ͳ Ͳ BC 4401 3394 3550 25617 483 300 224 3.1 Na 843 829 614 4434 0 400 173 0.3 Mg 936 769 736 4781 70 72 195 0.5 Al 5524 4832 3800 23436 191 40 223 3.4 Si 13579 10877 10696 61060 507 30 224 26.0 P 211 285 130 2087 0 61 189 0.6 S 1458 1659 965 13831 129 30 224 15.1 Cl 331 244 264 1284 34 27 224 6.0 K 1760 1441 1340 7813 72 25 224 18.0 Ca 4604 3443 3616 20023 257 25 224 20.6 Ti 298 324 197 2981 7 17 224 8.1 Mn 80 66 60 364 3 14 224 3.1 Fe 3338 2707 2493 14214 116 16 224 1.0 Cu 21 38 9 372 0 18 126 0.4 Zn 99 180 58 2415 0 16 211 2.5 a Limitofdetection

Figure3.Plotofpredicted(PMF)PM102.5massconcentrationsversus

measured(gravimetric)PM102.5massconcentrations.

Table2.AveragesourcemasscontributionstoPM102.5atUlaanbaatar

Source PM10Ͳ2.5massμgmͲ3 Soil1 42.5(3.0)a Combustion 11.9(0.7) Soil2 54.5(4.0) Roaddust 13.9(1.9) a Standarddeviation

thesourcelocation,withtheSoil2sourceoriginatingmorelocally

inUlaanbaatarwherethereislikelytobeagreaterconcentration

ofsettledcombustionparticlesandcoaldustsmixedintothe

crustalmatter,hencethehigherpresenceofBCinthesource

profile.Analysisshowsthatthetwosoilsourcesarecorrelated

(r2=0.8).However,athreefactorPMFsolutionwhichessentially

combinedthetwocrustalmattersourcesyieldsunsatisfactory

results,primarilyduetoahighQvalueandsignificantskewsinthe

residualplots(morehighlypositiveornegativeforanumberof

elements) particularly for Al, Si, K, Ca and Fe, the major

contributingelementstocrustalmatter.

Interestingly,thethirdfactorhasbeenidentifiedasacoal

combustionsourcecontributionforthecoarsefractionandthis

containsthemajorityofblackcarbonandasignificantsulphur

component.Thetime–seriesplotsofthepersamplemasscontribͲ

utionspresentedinFigure5showthatthePM10Ͳ2.5coalburning

Figure4.ElementalsourceprofilesderivedbyPMFforPM102.5samples

(predicted±SD).

sourcecontributionspeakedduringwinterindicatingthatitwas

likelytobeduetoemissionsfromlocalcoalcombustionsources.

Whiletheemissionofparticlesfromcombustionsourcesare

primarilyinthesubͲmicrometerrange(Hedbergetal.,2002)there

islikelytobesomeagglomeration,condensationorimpactionof

thefinefractionontocoarserparticles(and/orontothecoarse

filteroftheSFU)toprovideameasurablecontributiontothe

coarsefractionsample.Thefourthsourcehasbeenidentifiedasa

localroaddustcomponentasthisprofilecontainsBCandmostof

thezincalongwithelementstypicalofcrustalmatter.Mongoliais

understoodtohavephasedouttheuseofleadedpetroloverthe

pastfewyears(UNEP,2008)andleadwasonlydetectedabovethe

LOD(a100ngmͲ3)inafewofthe coarseparticulatematter

samples.FurtherdiscussionisprovidedinSection3.4.

Figure5 showsthedominance ofcrustalmattersource

contributionstothecoarseparticlefractionatUlaanbaatarwith

combinedmaximumconcentrationsupto460μgmо3.Figure6

showsthatthecrustalmattercontributionsarehighestduring

springandautumnwithsignificantcontributionsduringwinter

fromtheSoil2crustalmattersource.Thetemporalvariationsin

crustal mattercontributions areconsistent with precipitation

patternsinUlaanbaatarasmostrainfallsduringsummeranda

consequentsuppressionofdustgeneration.Spring,autumnand

winterinMongoliaaregenerallydrywithspringandautumnthe

windiestseasons(Xuanetal.,2004). y = 0.79x R2 _{= 0.74} 0 200 400 600 800 1000 0 200 400 600 800 1000 PM10-2.5 Gravimetric mass (Pg m -3 ) PM 10 -2 .5 P M F m ass (P g m -3)

Figure5.ContributiontimeseriesofresolvedsourcesforPM102.5atUlaanbaatar.

Figure6.SeasonalsourcemasscontributionstoPM10Ͳ2.5atUlaanbaatar.

3.3.PM2.5dataanalysis

Atotalof232fineparticulatemattersampleswereincludedin

thereceptormodellinganalysesofUlaanbaatardata.Whileatotal

of25elementalconstituentsweredetectedabovetheirrespective

LOD,15specieswereselectedforPMFmodelling(PM2.5,BC,Na, Mg,Al,Si,P,S,Cl,K,Ca,Ti,Mn,Fe,Cu,andZn)asotherelements

wereonlypresentinafewsamplesand/ortheS/Nratiowaslow.

Table3presentsthemean,standarddeviation,median,maximum,

minimum,andnumberofsamplesaboveLODandS/Nratiofor

individualspecies.

ThesourceprofilesforthePM2.5datawerederivedinthe samewayasforthePM10Ͳ2.5data.AsevenͲsourcemodelwas found to offer the most appropriate solution after rotation

(FPEAK=Ͳ0.6).Figure7showsthecomparisonbetweenmodelled

(PMF)PM2.5andmeasuredPM2.5concentrations(r2=0.89).The averagemasscontributionsofeachofthesourcestoambientfine

particleconcentrationsareshowninTable4andtheelemental

sourceprofilesarepresentedinFigure8.Thederivedsource

profileswerecomparedtomeasuredsourceprofilesand/orthose

reportedinotherstudiestoensuretheymadephysicalsense. Table3.SummarystatisticsforPM2.5andelementalconcentrationsatUlaanbaatar

ngmͲ3

ArithmeticMean StdDev Median Maximum Minimum LOD #Samples>LODa

S/N PM2.5 46000 51000 28 000 400 000 6 000 Ͳ Ͳ Ͳ BC 7290 10454 4 242 94 206 680 300 232 2.75 Na 290 425 112 2 642 0 500 88 0.13 Mg 326 276 235 2 083 31 80 200 0.62 Al 1150 1224 745 7 627 0 44 227 0.81 Si 2305 1740 1 871 10 554 129 33 232 17.20 P 79 106 42 773 0 62 157 0.24 S 1969 3978 900 40 079 125 31 232 17.55 Cl 139 133 88 849 12 28 232 2.14 K 324 239 243 1 558 35 22 232 7.00 Ca 789 559 652 3 194 50 19 236 11.92 Ti 37 33 28 156 0 15 191 0.27 Mn 15 14 11 65 0 11 178 0.18 Fe 523 388 416 2 150 26 10 232 0.62 Cu 10 29 3 373 0 13 105 0.21 Zn 44 54 30 400 0 14 213 0.55 a Limitofdetection 0 10 20 30 40 50 60 70 80 S o il 1 Combu s ti o n S o il 2 R o ad du st M a ss co nt ri b u ti o n t o P M 10-2. 5 (P g m -3) Winter Spring Summer Autumn

Figure7.Plotofpredicted(PMF)PM2.5massconcentrationsversus

measured(gravimetric)PM2.5massconcentrations.

Table4.AveragesourcemasscontributionstoPM2.5atUlaanbaatar

Source PM2.5massμgmͲ3 Soil 5.0(0.4)a Coalcombustion1 12.3(0.8) Coalcombustion2 11.6(1.8) Motorvehicles 1.9(0.6) Biomassburning 1.1(0.2) Roaddust 3.0(0.4) Zinc 0.6(0.3) a Standarddeviation

Thefirstsourceprofilewasidentifiedasthefinefractionof

airbornesoiloriginatingfromcrustalmatterastheprofilecontains

themajorityofAlandSiwithanSi:Alratioofapproximately2.5:1.

Thesecondfactor,andhighestsourcecontributor,wasidentified

asacoalburningsourceasitwasdominatedbyblackcarbonwith

a significant sulphur component as found in otherreceptor

modellingstudies(Mukaietal.,1993;Mukaietal.,2001;Songet

al.,2006).Thethirdfactorwasalsoattributedtocoalcombustion

sourcesbutwithalessersulphurcomponentandsomedifferences

in minor elemental species present in the source profile

composition.Wesuggesttherearetwodistinctcoalcombustion

sourcetypespresentintheUlaanbaatarairshedwithdiffering

combustion characteristics. Further discussion is provided in

Section3.4.Thefourthsourcehasbeenidentifiedasoriginating

from motor vehicles as a similar separation of elemental

componentsassociatedwithfineparticlemotorvehicleemissions

hasbeenreportedelsewhere(Mizohataetal.,1995;Schaueretal.,

2006;Wahlinetal.,2006).Thefifthsourcerepresentsbiomass

burningcontributionstoPM2.5ambientmassconcentrationsdue totheassociationofblackcarbonandpotassiumintheprofile.

WoodburningforcookingisusedallyearinUlaanbaatarbutthere

wasalsoevidencethatforestfireeventstothenorthoftheregion

can contribute at certain times. The sixth factor has been

attributedtofinerfractionofairborneroaddustandtheseventh

sourceextractedfromthePMFanalysiscontainedadistinctZn

componenttotheelementalsignature.Itwouldappearfromthe

PSCFanalysisthatthisoriginatestothewestofUlaanbaatar.

ThetemporalvariationsonsourcecontributionstoPM2.5are presentedinFigure9.Notethatthescalesforthecoalcombustion

sourcesareanorderofmagnitudehigherthantheotherplots.

Thetemporalvariationinsourcecontributionsshowsthatcoal

combustion peaks during winter (December–February) in

Ulaanbaatar. Seasonal contributions by PM2.5 sources are presentedinFigure10.PeakcontributionsfromthePM2.5crustal matterandroaddustsourcesoccurredduringspringandautumn

commensuratewithdriestseasonsasdidthebiomassburning

source.

3.4.Discussion

Crustalmattersources.ThefactoranalysisofPM10Ͳ2.5andPM2.5 fractionshasshownthatcrustalmattersourcesaresignificant

contributors to air particulate matter concentrations at

Ulaanbaatar,primarilydominatingthecoarsefraction.Wehave

identifiedtwocoarsefractioncrustalmatterelementalprofiles

(Soil1andSoil2)andsuggestthattheyoriginatefromdifferent

sourceareasbasedontheBCcontentandotherelementsinthe

profile.ThecrustalmatterelementalprofilesderivedbyPMFwere

compared to actual soil samples collected in and around

Ulaanbaatar (Markwitz et al., 2008). Markwitz et al. (2008)

collected crustal matter samples from sand dunes, open

countrysideandurbanlocations(samplelocationsanddescriptions

areprovidedinthereference).Thecrustalmattersampleswere

then compressed into discs and elemental concentrations

determinedbyIBAatGNS.Duetothenatureofthesamples,BC

wasnotmeasuredinthatstudy.Figure11presentselemental

profiles(asthepercentageofthetotalmeasuredelementalmass)

forelementscommontothecrustalmattersamplescollectedby

Markwitzetal.(2008)andthosederivedfromthecoarseandfine

PMFanalysisinthisstudy.

Figure8.ElementalsourceprofilesderivedbyPMFforPM2.5samples(predicted±SD).

y = 0.71x R2 _{= 0.78} 0 100 200 300 400 500 0 100 200 300 400 500 PM2.5 Gravimetric mass (Pg m -3 ) PM 2. 5 PM F m a ss (P g m -3)

Figure9.ContributiontimeseriesofresolvedsourcesforPM2.5atUlaanbaatar(notethedifferenceinscaleforcoalcombustionsources).

Figure10.SourcecontributionstoPM2.5byseasonatUlaanbaatar.

Figure11showsthatthesandsampleshavethehighestSi:Al

ratioandalowerironandcalciumcontentinaccordancewiththe

highersilicacontentofsand.Thecityandcountrycrustalmatter

samples have a similar composition to each other. The air

particulatemattercrustalmatterprofilesderivedbyPMFallhavea

lowerSi:Alratiothanthecrustalmattersourcesamples,most

likelyduetoparticlesizefractionationofthemineralsmakingup

crustalmatterwithaluminosilicatespresentinthefinefractions.

TheseresultsindicatethattheSoilsourcesarelargelyderivedfrom

thetransportoffineloessratherthansandasobservedbyZhang

etal.(2003).

MasscontributionstoPM10Ͳ2.5fromtheSoil1component weregenerallyhigherthanSoil2andprobablyrepresentdust

source areas from the Ulaanbaatar, surrounding countryside

environsandlongͲrangetransport.TheSoil2dustsourcehasbeen

identifiedasmorelocalisedtotheUlaanbaatarurbanareaduetoa

significantBCcomponentandthepresenceofPandCuinthe

elemental soil source profile. Analyses of air mass back– trajectoriesusingHYSPLIT(DraxlerandRolph,2003)indicatea

distinctwesterlydirectionalcomponentforsampledayswhere

there was a significant Soil 2 contribution but little or no

contributionfromtheSoil1source.Westofthemonitoringsite,

acrossUlaanbaatar,aretwolargecoalfiredpowerstationsas

indicatedinFigure12.Associatedwiththepowerstationsarelarge

stockpilesofcoalandslagdumpsthatarelikelytobesignificant

sourcesoffugitivedustemissionsandmayaccountfortheBC

observedintheSoil2elementalprofile.Otherurbansourceareas

offugitive dusts such as construction, excavations andland

clearance,unsealedyardsandroadsarealsolikelytocontributeto

themix. Figure11.ElementalprofilesforcrustalmattersourcesamplesandthosederivedforcoarseandfineairparticulatemattersamplesbyPMF. 0 5 10 15 20 25 30 35 So il Co a l c om b us ti on 1 Co a l c om b us ti on 2 R oa d dus t M o to r v e h ic le s B io m ass bu rni ng _Zin c M ass co nt ri b u ti o n (P g m -3) Winter Spring Summer Autumn 0 20 40 60 80 100 Na Al Si S Cl K Ca Ti V Mn Fe Cu Zn P er ce n ta g e o f el em e n ta l m a ss Sand City Country PM10-2.5 Soil 1 PM10-2.5 Soil 2 PM2.5 Soil

Figure12.AerialviewofUlaanbaatarindicatingmonitoringsitelocation()andcoalfiredpowerstationstothewest(source:Googlemaps2008).

Air–massbacktrajectoriesforpeakSoil1andSoil2source

contributions indicate source areas to the west and north

respectively.NorthwestwindspeakduringspringinMongoliadue

tothebreakdownintheSiberianHighPressurezoneafterwinter

(Xuanetal.,2004).Xuanetal.(2004)foundthattheAsianDust

Storm events originate in the deserts of southern

Mongolia/northernChina.Otherstudieshaveshownsourceareas

in western and northwestern China and trajectories across

Mongoliafromthenorthwesttosoutheast(seeforexampleZhang

etal.,2003;Xuanetal.,2004;Songetal.,2006;Wangetal.,2006;

Kim et al., 2007). One peak dust event in Ulaanbaatar as

determinedfromthePMFanalysishasbeencorroboratedwith

satellite imagesand otherobservations in eastern Asiathat

indicatelongͲrangetransportmechanisms.Forexample,theSoil1

sourceshowedapeakconcentrationon6March2006(180μgmо3)

andsatelliteimagery(NASAEarthObservatory)for9March2006

shows the dust storm spreading outover eastern China as

presentedinFigure13.

Itseemsevidentfromthisstudythatsignificantloadingof

dustswerepresentinair–massesarrivingatUlaanbaatarfromthe

westandnorthwestandthatthecrustalmatter,particularlythe

PM2.5fraction(peakconcentrationsatUlaanbaatar20–30μgmо3) canbetransportedconsiderablyfurthertothesoutheastand

arriveatcitiesaffectedbyAsianDustStormeventssuchasBeijing

(Zhangetal.,2005;Songetal.,2006).Figure14showsthePSCF

analysisfor thePM2.5 soil source indicating potentialsource locationsthatincludethedesertsinthesouthofMongolia(Gobi

Desert),centralKazakhstan(BetpaqdalaDesert)tothenorthwest

ofUlaanbaatar,andpossiblytheTaklamakanDesert(Wangetal.,

2008)inChinatothewestofMongolia.

Combustion sources. Combustion sourceswere found to be

significantsourcecontributorstoairparticulatematter,primarily

to PM2.5, at Ulaanbaatar. Cold temperatures and inversion conditionsduringthewinterseverelylimitthedispersionofair

pollutantinUlaanbaatarandthereforehighconcentrationsof

combustionderivedPM2.5arerecordedintheUlaanbaatarairshed. In 2006/2007 approximately 4900000tonnes of coal and

413000tonnesofwoodwereburntinUlaanbaatar(NAMHEM,

2007).Analysisoftemporalvariationsinsourcecontributions

showedthatcoalcombustionsourcesdominateduringwinterdue

theuseofcoalforheatingandpowergeneration.Air–massback– trajectoryanalysesindicatethatpeakcontributionsfromtheCoal

combustion1sourceoccurduringwesterlywindspossiblydueto

emissionsfromthecoalfiredpowerstationstothewest(shownon

theaerialphotographinFigure12)directlyimpactingatthe

monitoringsite.TheCoalcombustion1sourceprofileisdominated

byBCandS(mostlikelypresentassulphatespecies)repreͲ

sentativeofhightemperaturecoalcombustioncharacteristicof

powerstations(Polissaretal.,2001).TheCoalcombustion2

sourcerepresentsemissionsfromcoalfiresusedfordomestic

heating as a significant proportion of the urban population

(120000households)inandaroundUlaanbatarliveinGers(48%)

(traditionalMongolianhouse)orsmallbrickdwellings(52%),both

ofwhichareheatedbysimplesolidfuelappliancesoropenfires

(NAMHEM,2007).Assuch,thesourceprofileisrepresentativeof

lowͲtemperaturecoalcombustionwithsignificantashandcrustal

mattercomponents(Na,Mg,Al,Si)(Zhengetal.,2005).Bothcoal

combustionsourcespeakduringwinterduetodemandforheating

andthesignificanttemperatureinversionconditionseffectively

trappingpollutantsintheUlaanbaatarairshed.SourcecontriͲ

butionstoPM2.5fromcoalcombustionduringwinterappeartobe increasingannually.Someofthevariationisprobablydueto

interannualclimatevariations,however,asignificantincreasein

thePM2.5coalcombustioncontributionsover2007/2008winter periodsuggeststhattheincreasemaybeduetocoalusage.Figure

15presentstheaveragemasscontributionstoPM2.5fromcoal combustionsourcesforeachwinter(December–February)since

monitoringbeganin2004.Furthermonitoringisrequiredtoseeif

thetrendcontinues.

Figure13.Dustcloud(tandiscolouration)spreadingoutoverEasternchina andKoreanPeninsulaon9March2006(source:http://earthobservatory. nasa.gov/NaturalHazards).

Figure14.PSCFanalysisforPM2.5soilsourcecontributionsshowingpotentialsourcelocations.

Figure15.Averagewinter(December–February)contributionstoPM2.5

fromcoalcombustionsourcesinUlaanbaatar.

The biomass burning source originates primarily from

domesticcookingactivitiesinUlaanbaatarwithpeakcontributions

fromforestfirestothenorthduringspringandautumn.Forest

firesareacommonoccurrenceduringthedryspringandautumn

monthsinMongolia (Nyamjavetal.,2007). Air–mass back– trajectoryanalysisforpeakbiomassburningeventsshowthat

sourceemissionsoriginatenorthofUlaanbaatar.Forexample,a

peakinbiomassburningsourcecontributiontoPM2.5occurredon 20April2006(4μgmо3).Theback–trajectorypresentedinFigure

16showstheairmassarrivingatUlaanbaatarfromthenorthand

satelliterecordsshowsignificantfiresburninginsouthernRussia

andnorthernMongoliaoverthesameperiod(Figure17).

Figure17presentsthesatelliteimageforApril23,2006,

showingthatlargefireswereburninginthefoothillsofthe

mountainsthatseparateRussia(north)fromMongolia(south).The

Moderate Resolution Imaging Spectroradiometer (MODIS) on

NASA’sTerrasatellitecapturedtheimageofthefires(actively

burningareasoutlinedinred)andtheiraccompanyinglargeburn

scars(darkbrown).Thesmokewasblowingsoutheastwardacross

Mongolia(fromhttp://earthobservatory.nasa.gov/NaturalHazards/

view.php?id=16493).

ThezincsourcewasfoundtobeaminorcontributortoPM2.5 inUlaanabaatar.InitiallytheassociationofzincandBCinthe

profilesuggestedthatitmay be dueto twostrokeengine

emissions (Begum et al., 2004). However, few vehicles in

Ulaanbaatararethoughttobepoweredbytwostrokeenginesand

nootherlocalzinccontainingparticulatematteremissionsource

hasbeenidentifiedasyet.ThePSCFanalysisforthezincsource

presentedinFigure18indicatesthatthepotentialsourceareafor

thezincsourcewasnorthwesternChinacentredonUrumchi,the

localprovincialcapital.Thezincsourcemaybeduetolong–range

transportofemissionsfromheavyindustryinthearea(e.g.zinc

miningandsmelting). Figure16.AirmassbackͲtrajectory(24hr)forApril202006(Source:NOAA, HYSPLIT). 0 10 20 30 40 50 60 70 80 90 100 2004/2005 2005/2006 2006/2007 2007/2008 Winter period PM 2. 5 m a ss co nt ri b u ti o n P g m -3

Figure17.SatelliteimageofforestfiresburninginsouthernRussiaand northernMongoliaon23April2006(source:NASAEarthObservatoryat http://earthobservatory.nasa.gov/NaturalHazards/view.php?id=16493).

4.

Conclusions

ThisstudypresentedtheresultsoffouryearsofPM10Ͳ2.5and PM2.5 monitoring data collected at Ulaanbaatar, Mongolia. Particulatematterconcentrationswerefoundtobeextremelyhigh

at times with maximum PM10 concentrations measured at 1100μgmо3andmanydayswhereconcentrationswereabove

300μgmо3.ThecoarseparticlefractiongenerallydominatedPM10 buthighPM2.5valueswererecordedduringwintermonthswith PM2.5valuesoftenover100μgmо3.Withsuchhighparticulate

matterconcentrationspresentintheUlaanbaatarurbanareathere

arelikelytobesignificanthealthimplicationsfortheexposed

population.Elementalconcentrationswithinparticulatematter

collectedontofiltersweredeterminedbyIBAandfactoranalysis

performedusingPositiveMatrixFactorisation.

Four source contributors were identified for the coarse

particlefractionandsevensourcesderivedforthefineparticle

fraction.Majorsourcescontributingtoparticleconcentrations

werefoundtobecrustalmatter(primarilythecoarsefraction)and

coal combustion sources dominated the fine fraction during

winter.Crustalmattersourceswerefoundtobegeneratedboth

locally(urbandustsources)andtransportedtoUlaanbaatarfrom

thewestandnorthwest,particularlyduringthedryandwindy

monthsofspringandautumn.Duststormeventsinthisregionare

wellknowntogeneratetheAsianDusteventsthataffectmany

partsofeasternandsoutheasternAsiaduringspringandcanbe

transportedoutoverthePacific.

SourcecontributionstoPM2.5weredominatedbyemissions fromcoalcombustionsourceswithtwodistinctsourcetypes,the

first identified as a high–temperature combustion source

originatingfrompowerstationemissionsinthewesternpartof

Ulaanbaatar and the other from coal combustion used for

domesticheatingduringwinter.Themonitoringdatasuggeststhat

masscontributionsfromcoalcombustionsourcesareincreasing.It

wasalsofoundthat,attimes,forestfiresinnorthernMongoliaand

southernRussiacanimpactonairqualityinUlaanbaataranda

long–rangetracesourceofzinccontainingPM2.5particleswas identifiedasoriginatinginnorthwesternChina.

Acknowledgements

ThisprojectwassupportedinpartbytheInternationalAtomic

Energy Agency through the Regional Cooperation Agreement

Programme“RAS/7/015ͲCharacterizationandSourceIdentification

ofParticulateAirPollutionintheAsiaRegion”.

Figure18.PSCFanalysisforthePM2.5zincsourceshowingapotentialsourcelocationinnorthwesternChina(Urumchi)tothewestofMongolia.

References

Begum,B.A.,Kim,E.,Jeong,C.ͲH.,Lee,D.ͲW.,Hopke,P.K.,2005.Evaluation ofthepotentialsourcecontributionfunctionusingthe2002Quebec forestfireepisode.AtmosphericEnvironment39,3719Ͳ3724.

Begum,B.A.,Kim,E.,Biswas,S.K.,Hopke,P.K.,2004.Investigationof sourcesofatmosphericaerosolaturbanandsemiͲurbanareasin Bangladesh.AtmosphericEnvironment38,3025Ͳ3038.

Draxler,R.R.,Rolph,G.D.,2003.HYSPLIT(HYbridSingleͲParticleLagrangian IntegratedTrajectory)Model,NOAAAirResourcesLaboratory,Silver Spring,MD.

Fine,P.M.,Cass,G.R.,Simoneit,B.R.T.,2001.Chemicalcharacterizationof fineparticleemissionsfromfireplacecombustionofwoodsgrownin thenortheasternUnitedStates.EnvironmentalScienceandTechnology 35,2665Ͳ2675.

Guttikunda,S., 2007. Urban air pollution analysis for Ulaanbaatar, Mongolia.WorldBankConsultantReport,WashingtonDC,USA. Hedberg,E.,Kristensson,A.,Ohlsson,M.,Johansson,C.,Johansson,P.A.,

Swietlicki,E.,Vesely,V.,Wideqvist,U.,Westerholm,R.,2002.Chemical and physical characterization of emissions from birch wood combustioninawoodstove.AtmosphericEnvironment36,4823Ͳ4837. Hopke,P.K.,Cohen,D.D.,Begum,B.A.,Biswas,S.K.,Ni,B.,Pandit,G.G.,

Santoso,M.,Chung,Y.S.,Davy,P.,Markwitz,A.,Waheed,S.,Siddique, N.,Santos,F.L.,Pabroa,P.C.B.,Seneviratne,M.C.S.,Wimolwattanapun, W.,Bunprapob,S.,Vuong,T.B.,Hien,P.D.,Markowicz,A.,2008.Urban airqualityintheAsianregion.ScienceoftheTotalEnvironment404, 103Ͳ112.

Hopke,P.K.,Xie,Y.,Raunemaa,T.,Biegalski,S.,Landsberger,S.,Maenhaut, W.,Artaxo,P.,Cohen,D.,1997.CharacterizationoftheGentstacked filterunitPM10sampler.AerosolScienceandTechnology27,726Ͳ735. Horvath,H.,1997.Experimentalcalibrationforaerosollightabsorption

measurementsusingtheintegratingplatemethod Ͳsummaryofthe data.JournalofAerosolScience28,1149Ͳ1161.

Horvath,H.,1993.Atmosphericlightabsorption Ͳareview.Atmospheric EnvironmentͲPartAGeneralTopics27A,293Ͳ317.

Jacobson,M.C.,Hansson,H.C.,Noone,K.J.,Charlson,R.J.,2000.Organic atmosphericaerosols:reviewandstateofthescience.Reviewsof Geophysics38,267Ͳ294.

Kim,E.,Hopke,P.K.,2006.Characterizationoffineparticlesourcesinthe GreatSmokyMountainsarea.ScienceoftheTotalEnvironment368, 781Ͳ794.

Kim,J.Y.,Ghim,Y.S.,Song,C.H.,Yoon,S.C.,Han,J.S.,2007.Seasonal characteristicsofairmassesarrivingatGosan,Korea,usingfineparticle measurementsbetweenNovember2001andAugust2003.Journalof GeophysicalResearchͲAtmospheres112,art.no.D07202.

Lide,D.R.,1992.CRCHandbookofChemistryandPhysics,CRCPressInc. Maenhaut,W.,Francois,F.,Cafmeyer,J.,1993.The“Gent”StackedFilter

UnitSamplerfortheCollectionofAtmosphericAerosolsinTwoSize Fractions:DescriptionandInstructionsforInstallationandUse.CoͲ ordinatedResearchProgramme:CRPE4.10.08.Vienna,International AtomicEnergyAgency.

Markwitz, A.,2005.AnoverviewoftheRCA/IAEAactivities inthe AustralasianRegionusingnuclearanalysistechniquesformonitoring airpollution.InternationalJournalofPIXE15,271Ͳ276.

Markwitz,A.,Barry,B.,Shagjjamba,D.,2008.PIXEanalysisofsandandsoil fromUlaanbaatarandKarakurum,Mongolia.NuclearInstrumentsand MethodsinPhysicsResearch,SectionB:BeamInteractionswith MaterialsandAtoms266,4010Ͳ4019.

Mizohata,A.,Ito,N.,1995.Quantitativedeterminationoftheairborne particulatemattersourcesbymotorvehiclesusingTTFA.TaikiKankyo Gakkaishi30,243Ͳ55.

Mukai,H.,Furuta,N.,Fujii,T.,Ambe,Y.,Sakamoto,K.,Hashimoto,Y.,1993. CharacterizationofsourcesofleadintheurbanairofAsiausingratios ofstableleadisotopes.EnvironmentalScienceandTechnology27, 1347Ͳ1356.

Mukai,H.,Tanaka,A.,Fujii,T.,Zeng,Y.,Hong,Y.,Tang,J.,Guo,S.,Xue,H., Sun,Z.,Zhou,J.,Xue,D.,Zhao,J.,Zhai,G.,Gu,J.,Zhai,P.,2001. Regionalcharacteristicsofsulfurandleadisotoperatiosinthe atmosphereatseveralChineseurbansites.EnvironmentalScienceand Technology35,1064Ͳ1071.

NAMHEM,2007.EmissionsInventoryforUlaanbaatar.NationalAgencyfor Meteorology, Hydrology, Environmental Monitoring, Mongolian Government.

Nyamjav,B.,Goldammer,J.G.,Uibrig,H.,2007.Theforestfiresituationin Mongolia.InternationalForestFireNews36,46Ͳ66.

Paatero,P.,Hopke,P.K.,Begum,B.A.,Biswas,S.K.,2005.Agraphical diagnosticmethodforassessingtherotationinfactoranalyticalmodels ofatmosphericpollution.AtmosphericEnvironment39,193Ͳ201. Paatero,P.,HopkeP.K.,2003.DiscardingordownweightinghighͲnoise

variablesinfactoranalyticmodels.AnalyticaChimicaActa490,277Ͳ 289.

Paatero,P.,Hopke,P.K.,Song,X.H.,Ramadan,Z.,2002.Understandingand controllingrotationsinfactoranalyticmodels.Chemometricsand IntelligentLaboratorySystems60,253Ͳ264.

Paatero,P.,Hopke,P.K.,2002.Utilizingwinddirectionandwindspeedas independent variables in multilinear receptor modeling studies. ChemometricsandIntelligentLaboratorySystems60,25Ͳ41.

Paatero,P.,1997.LeastsquaresformulationofrobustnonͲnegativefactor analysis.ChemometricsandIntelligentLaboratorySystems37,23Ͳ35. Polissar,A.V.,Hopke,P.K.,Poirot,R.L.,2001.Atmosphericaerosolover

Vermont:chemicalcompositionandsources."EnvironmentalScience andTechnology35,4604Ͳ4621.

Salma,I.,Chi,X.,Maenhaut,W.,2004.Elementalandorganiccarbonin urbancanyonandbackgroundenvironmentsinBudapest,Hungary. AtmosphericEnvironment38,27Ͳ36.

Schauer,J.J.,Lough,G.C.,Shafer,M.M.,Christensen,W.F.,Arndt,M.F., DeMinter,J.T.,Park,J.S.,2006.Characterizationofmetalsemitted frommotorvehicles.ResearchReportHealthEffectsInstitute133,1Ͳ 88.

Song,Y.,Zhang,Y.,Xie,S.Zeng,L.,Zheng,M.,Salmon,L.G.,Shao,M., Slanina,S.,2006.SourceapportionmentofPM2.5inBeijingbypositive matrixfactorization.AtmosphericEnvironment40,1526Ͳ1537. Song,X.H.,Polissar,A.V.,Hopke,P.K.,2001.Sourcesoffineparticle

compositioninthenortheasternUS.AtmosphericEnvironment35, 5277Ͳ5286.

Trompetter,W.J.,Markwitz,A.,Davy,P.,2005.Airparticulateresearch capabilityattheNewZealandionbeamanalysisfacilityusingPIXEand IBAtechniques.InternationalJournalofPIXE15,249Ͳ255.

UNEP,2008.Target2008:GlobalEliminationofLeadedPetrol.AReportof thePartnershipforCleanFuelsandVehicles(PCFV)UnitedNations EnvironmentProgram,Nairobi,Kenya.

Wahlin,P.,Berkowicz,R.,Palmgren,F.,2006.Characterisationoftraffic generated particulate matter in Copenhagen. Atmospheric Environment40,2151Ͳ2159.

Wang,X.,Huang,J.,Ji,M.,Higuchi,K.,2008.VariabilityofEastAsiadust eventsandtheirlongͲtermtrend.AtmosphericEnvironment42,3156Ͳ 3165.

Wang,Y.Q.,Zhang,X.Y.,Arimoto,R.,2006.Thecontributionfromdistant dustsourcestotheatmosphericparticulatematterloadingsatXiAn, Chinaduringspring.ScienceoftheTotalEnvironment368,875Ͳ883. Watson,J.G.,Zhu,T.,Chow,J.C.,Engelbrecht,J.,Fujita,E.M.,Wilson,W.E.,

2002.Receptormodelingapplicationframeworkforparticlesource apportionment.Chemosphere49,1093Ͳ1136.

Xuan,J.,Sokolik,I.N.,Hao,J.,Guo,F.,Mao,H.,Yang,G.,2004.Identification andcharacterizationofsourcesofatmosphericmineraldustinEast Asia.AtmosphericEnvironment38,6239Ͳ6252.

Zhang,R.,Arimoto,R.,An,J.,Yabuki,S.,Sun,J.,2005.Groundobservations ofastrongduststorminBeijinginMarch2002.JournalofGeophysical ResearchͲAtmospheres110,art.no.D18S06.

Zhang,X.Y.,Gong,S.L.,Shen,Z.X.,Mei,F.M.,Xi,X.X.,Liu,L.C.,Zhou,Z.J., Wang,D.,Wang,Y.Q.,Cheng,Y.,2003.Characterizationofsoildust aerosolinChinaanditstransportanddistributionduring2001ACEͲ Asia:1.Networkobservations.JournalofGeophysicalResearchͲ Atmospheres108,art.no.4261.

Zheng,M.,Salmon,L.G.,Schauer,J.J.,Zeng,L.,Kiang,C.S.,ZhangY.,Cass, G.R.,2005.SeasonaltrendsinPM2.5sourcecontributionsinBeijing, China.AtmosphericEnvironment39,3967Ͳ3976.