Spatial and temporal variation of historical anthropogenic NMVOCs emission inventories in China

©Author(s)2008.ThisworkisdistributedundertheCreativeCommonsAttribution3.0License.

AtmosphericChemistryandPhysics

SpatialandtemporalvariationofhistoricalanthropogenicNMVOCsemissioninventoriesinChina

Y.Bo,H.Cai,andS.D.Xie

CollegeofEnvironmentalSciencesandEngineering,StateKeyJointLaboratoryofEnvironmentalSimulationandPollutionControl,PekingUniversity,Beijing,China

Received:7April2008–PublishedinAtmos.Chem.Phys.Discuss.:11June2008Revised:13October2008–Accepted:3November2008–Published:Abstract.MultiyearemissioninventoriesofanthropogenicNMVOCsinChinafor1980–2005wereestablishedbasedontime-varyingstatisticaldata,literaturesurveyedandmodelcalculatedemissionfactors,whichwerefurthergriddedatahighspatialresolutionof40km×40kmusingtheGISmethodology.ResultsshowacontinuousgrowthtrendofChina’shistoricalNMVOCsemissionsduringtheperiodof1980–2005,withtheemissionincreasingby4.2timesatanannualaveragerateof10.6%from3.91Tgin1980to16.49Tgin2005.Vehicles,biomassburning,industrialprocesses,fossilfuelcombustion,solventutilization,andstorageandtransportgenerated5.50Tg,3.84Tg,2.76Tg,1.98Tg,1.87Tg,and0.55TgofNMVOCs,respectively,in2005.Motorcycles,biofuelburning,heavydutyvans,syn-theticfibreproduction,biomassopenburning,andindus-trialandcommercialconsumptionwereprimaryemissionsources.Besides,sourcecontributionsofNMVOCsemis-sionsshowedremarkableannualvariation.However,emis-sionsofthesesourceshadbeencontinuouslyincreasing,whichcoincidedwellwithChina’seconomicgrowth.Spa-tialdistributionofNMVOCsemissionsillustratesthathighemissionsmainlyconcentratesindevelopedregionsofnorth-ern,easternandsoutherncoastalareas,whichproducedmoreemissionsthantherelativelyunderdevelopedwesternandin-landregions.Particularly,southeastern,northern,andcentralChinacovering35.2%ofChina’sterritory,generated59.4%ofthetotalemissions,whilethepopulouscapitalcitiescov-eringmerely4.5%ofChina’sterritory,accountedfor24.9%ofthenationalemissions.Annualvariationofregionalemis-sionintensityshowsthatemissionsconcentratinginurbanareastendedtotransfertoruralareasyearbyyear.More-Correspondenceto:S.D.Xie(sdxie@pku.edu.cn)

over,eastern,southern,central,andnortheasternChinaweretypicalareasofhighemissionintensityandhadatendencyofexpandingtothenorthwesternChina,whichrevealedthetransferofemission-intensiveplantstotheseareas,togetherwiththeincreaseofbiomassopenburning.

1Introduction

Nonmethanevolatileorganiccompounds(NMVOCs),whichcanbestoredandreactintheair,arealargenumberofdif-ferentspeciesoforganiccompoundswithhighvaporpres-suresandlowboilingpoints.NMVOCshaveanimpor-tantinfluenceontheformationofphotochemicalsmog,sec-ondaryorganicaerosols,andorganicacidsintheair,andtheyareseverelyharmfultohumanhealth(Boeglinetal.,2006).Therefore,researchonthesourceandemissioncharacteris-ticsofNMVOCshasbeenanimportantissueinatmosphericchemistrystudies,whichnotonlyprovidesscientificbasisforthecontroloflocalphotochemicalsmogpollutionandforthepreventionandcontrolofacidrain,butalsohasapro-foundimpactonsystematicstudyonatmosphericchemicalreactions,andontheassessmentandforecastofthepossiblechangeofatmosphericenvironmentontheglobalscale.NMVOCsareemittedfrombothbiogenicandanthro-pogenicsources(Altshuller,1991;Fieldetal.,1992).Biogenicsourcesmainlyemitisopreneandmonoterpene,whileanthropogenicsourcesexhaustalmostallspeciesofNMVOCs.Thoughgloballybiogenicsourcesgeneratedmuchmoreemissions(1150Tg;Guentheretal.,1995)thananthropogenicsources(142Tg;Seinfeldetal.,1998),an-thropogenicNMVOCsemissionsdominateinthetotalemis-sionsontheurbanandregionalscale(Atkinson,2003).

PublishedbyCopernicusPublicationsonbehalfoftheEuropeanGeosciencesUnion.

2Y.Boetal.:Tempo-spatialvariationofanthropogenicNMVOCsemissions

Since1989,Europe,theUS,Australiaandotherdevel-opedcountrieshavebeguntocompilethenativeNMVOCsemissioninventories(LubkertandDetilly,1989;Loibletal.,1993),andhaveestimatedglobalNMVOCsemissioninventoryincludingChina(Piccotetal.,1992).Besides,somereseacherspaidspecialattentiontoestimatingChina’sNMVOCsemissioninventory.Klimontetal.(2002)devel-opedChina’sNMVOCsemissioninventoryin1990,1995,and2000ontheprovinciallevelforthefirsttime,basedonforeignemissionfactorsandestimatedactivitydata.Dur-ingtheTRACE-PandtheACE-Asiaexperiments,Streetsetal.(2003)estimatedChina’santhropogenicNMVOCsemis-sionsin2000,whichwerebasedontheresultsofKlimontandfurtherconsideredthesourceofsavannaandforestburn-ing.RegardingtheREASinventoryestablishedbyOharaetal.(2007),NMVOCsemissionsinChinafor1995and2000camefromtheresultsofKlimontetal.(2002)andStreetsetal.(2003).Inaddition,theyestimatedtheNMVOCsemis-sionsfortheperiod1980–2003byanextrapolationoftheemissionsin1995and2000,usingaproxyindicatorpersec-tor.Chinesescholarsalsocarriedoutsomeresearch.Zhangetal.(2000)demonstratedthespatialdistributionofhydro-carbonemissionfromstationarysourcesintheYangtzeRiverDelta.Lietal.(2003)usedMOBILE5modeltoestimatetheNMVOCsemissionsfromroadtrafficinChinain1995.Shenetal.(2006)estimatedtheVOCsemissionsfromChina’sgasstationsin2002.CaiandXie(2007)calculatedtheNMVOCsemissionfactorsofvariousvehiclecategoriesunderdrivingconditionsofurban,ruralandfreewayforeachprovinceofChina,usingCOPERTIIImodel,andestablishedChina’svehicularemissioninventoryincludingNMVOCsemissionsfortheperiodof1980–2005,whichwasfurthergriddedatahighspatialresolutionof40km×40kmusingtheGeo-graphicInformationSystem(GIS)methodology.

Asdescribedabove,internationalscientiststooktheleadindevelopingChina’sNMVOCsemissioninventories,andChinesescholarsalsocontributedtosomeanthropogenicemissionresults.However,thereweresignificantdifferencesbetweentheresults,andthoseinventoriesmostlyfocusedonChina’semissionsforasingleorfewyears.Thus,theaimofthisstudyistodevelopChina’shistoricalNMVOCsemis-sioninventoriesatahighspatialresolution.IntheprocessofestablishingChina’sNMVOCsemissioninventory,lo-callymeasuredemissionfactorsforanthropogenicNMVOCsemissionsourceswereselected,andtheemissionfactorsadoptedbyEurope,Americaandotherdevelopedcountrieswerechosenbasedonspecifichypotheses.Furthermore,mostactivitydatawereobtainedbyreferringtofirsthandsta-tisticaldatainChina.Thus,anthropogenicNMVOCsemis-sioninventoriesinChinaweredevelopedonthecountylevelfortheperiodof1980–2005,whichwerefurthergriddedatahighspatialresolutionof40km×40kmbasedontheGISmethodology.

Atmos.Chem.Phys.,8,1–20,2008

2Methodology

2.1Estimationandallocationofemissions2.1.1

Emissionestimation

Theemissionsourceswereclassifiedintoatotalofsixsourcesofvehicles,fossilfuelcombustion,biomassburn-ing,industrialprocesses,storageandtransport,andsolventutilization,accordingtotheactualsituationofanthropogenicNMVOCsemissionsinChina.Vehicularemissionswerecal-culatedbasedontheemissionfactors,thepopulationofve-hicles,andthecorrespondingmileagetraveledforeachvehi-clecategory.Theemissionestimationofothersourceswasbasedonannualaverageratesofemissionrelatedactivities,detailedemissionfactors,andremovalefficiencyforeachsource.Thus,thetotalNMVOCsemissionswereestimatedbyEq.(1).

Qm=󰀂󰀃

Pm,n󰀁+󰀂

×Mn,j×EFn,j

Am,i×EFi×(1−Ri)(1)where:QmistheNMVOCsemission(Mg)inprovincem;nisacertainvehiclecategory;jisacertaindrivingcycle;iistheothersourcesexcludingvehicles,i.e.fossilfuelcom-bustion,biomassburning,industrialprocesses,storageandtransport,andsolventutilization.Forvehicularemission,Pm,nisthepopulationofvehiclesincategoryninprovincem;Mn,jistheannualmileage(kilometers)forvehiclesincategorynunderdrivingcyclej;andEFn,jistheemissionfactor(g/km)ofvehiclesforcategorynunderdrivingcyclej;Am,iistheactivityrateofsourceiinprovincem;EFiistheemissionfactorofsourcei;andRiistheremovaleffi-ciencyofsourcei.

Upontheconsiderationofthefactthattheeconomicde-velopmentplansinChinahavebeencompiledineveryfiveyears,withacompleteandsystematicsummaryconductedinthelastyearofeachfive-year-planperiod,say,1980and1985.Therefore,theanthropogenicNMVOCsemissionin-ventoriesinChinawereestimatedbyanintervaloffiveyearsfortheperiodof1980–2005inthisstudy,sincethestatisticaldatawererelativelycompleteandaccurateinthelastyearofeachfive-year-planperiod.Specifically,therelatedactivitydatarequiredbythecompilationoftheemissioninventorywereobtainedfromtheprovincialstatisticaldatacovering31provinces,excludingTaiwan,HongKong,andMacao.TheNMVOCsemissionfactorsoftheanthropogenicsourcesexcludingvehiclesweremainlyobtainedfromtherecentlymeasureddata,literaturesandtheCompilationofAirPollutantEmissionFactors,whichwasusuallycalledtheAP42Report(EPA,1995).ThoughtheindustrializationandurbanizationinChinahavebeenspeedingup,thefactthatthereisagapofthetechnicallevelsandresourceutilizationefficiencybetweenChinaanddevelopedcountriesmustberecognized.Besides,thereisalsoatremendousgapbetweendifferentregionsofChina.Forexample,therewerenotable

www.atmos-chem-phys.net/8/1/2008/

Y.Boetal.:Tempo-spatialvariationofanthropogenicNMVOCsemissions3

differencesinthetechnicalinnovationandresourceutiliza-tionefficiencybetweenthedevelopedeasternChinaandtheunderdevelopedwesternChina,andbetweenurbanareasandruralregions.Thus,theemissionfactorsofthesamesourcesindifferentpartsofChinadifferred.Takingintoaccountofsuchaspects,thetechnicallevelsinChinawereabouttenyearsbehindthoseinthedevelopedcountries.Therefore,theemissionfactorsinagivenyearinChinawereselectedfromtheemissionfactordatatenyearsbeforeforthesamesourcesintheAP42Report,insteadofselectingtherenewedemissionfactordatathatrepresentthecurrentemissionlevelsinthedevelopedcountries.

Moreover,theannualvariationofemissionfactorsforeachsourcebefore2005isconsidered.Forthesourcesoffos-silfuelcombustionandindustrialprocesses,itwasassumedthattheNMVOCsemissionsperunitoffuelconsumedwerestable,andtheemissionfactorofeachsectorwasdirectlyproportionaltotheannualenergyconsumptionperproduc-tion.Thus,emissionfactorsforindustrialprocessesandfos-silfuelcombustionbefore2005wereadjustedbasedonthedeterminedemissionfactorsfor2005,usingEq.(2).EF󰀄

=Ca

C×EF0

0

(2)

where:EF󰀄isthemodifiedemissionfactorofeachsubsectorinyeara;Caistheannualenergyconsumptionperproduc-tionofeachsubsectorinyeara;C0isenergyconsumptionperproductionofeachsubsectorin2005;andEF0istheemissionfactorofeachsubsectorin2005.

Regardingthesourceofsolventutilization,itwasassumedthattheemissionfactorswerepositivelycorrelatedwithan-nualincomepercapital.Thus,theemissionfactorsforthissourcebefore2005wereestimatedbasedonthedeterminedemissionfactorsfor2005,usingEq.(3).EF󰀄󰀄=

Ib

×EF󰀄I󰀄0(3)

0where:EF󰀄󰀄

isthemodifiedemissionfactorofeachsubsectorofsolventutilizationinyearb;Ibistheannualincomepercapitalinyearb;I0󰀄istheannualincomepercapitalin2005;andEF0󰀄istheemissionfactorofeachsubsectorin2005.Emissionfactorsofbiomassopenburning,andstorageandtransportwereassumedtoremainconstantovertheyears,becausethewayofbiomassopenburningandthemeansofthestorageandtransportoforganicsolventinChinawereneverremarkablyimproved.2.1.2Emissionallocation

TheNMVOCsemissionsofeachanthropogenicsourceineachprovinceofChinawerefirstlycalculated,followedbyfurtherallocationtothecountylevelbasedonatop-downapproach.Uponthecompilationofthenationalemissioninventoriesbasedonprovincialcalculations,countylevelwww.atmos-chem-phys.net/8/1/2008/

Table1.Proxyvariablesusedforallocatingprovincialemissionsfromeachsourcetothecountylevel.

ProxyvariableEmissionsourcecategoryGDPa

vehicles,industrialprocesses,storageandtransport,fossilfuelcombustionPopulationb

Cropsseededareaa

solventutilizationbiomassburning

adatafromChinaCounty(city)SocialEconomicStatisticalYear-

books(2001,2006),ChinaAgricultureYearbooks(1996,2001,2006),bandprovincialstatisticalyearbooksin1996.

datafromChinaSstatisticalYearbooks(1996,2001,2006).

emissionswerecalculatedbyallocatingprovincialinvento-riesbasedonproxyvariablesofGDP,population,andcropseededarea,usingEq.(4).ETi,m,n=󰀂n

T×Ei,m

(4)

n=1

n

where:Ei,mistheemission(Mg)ofsourceiinprovince

m;Ei,m,nistheemission(Mg)ofsourceiincountynofprovincem;andTnistheGDP,population,orcropseededareaincountyn.TherelationshipbetweenemissionsourcesandproxyvariablesisshowninTable1.

Upontheallocationofprovincialemissionstocounties,thelatitudeandlongitudeprojectedmapofChinawasgrid-dedataresolutionof40km×40kmusingMapinfo,aGISsoftware,whichfurthercalculatedtheratiosoftheareasofeachcountythatfellovervariousgridcellstotheareaofthatcounty.Then,theemissionofthatcountywasallocatedtoeachgridcellbasedonthecalculatedratios.Finally,theemissionsfromvariouscountieswithinagridcellwereag-gregated,whichwasthetotalemissionofthatgridcell.2.2Compilationofemissionfactorsandactivitydata2.2.1

Vehicles

EmissionfactorsofeachvehiclecategorywerecalculatedbyusingCOPERTIIIprogramme,awidelyusedtoolfores-timationofemissionfactorsatEuropeanlevelforon-roadtransportation(EEA,2000).ItwassuitableforestimatingChina’svehicularemissionsbecausethedominantvehiclemanufacturingtechnologiesinChinaoriginatedfromEuro-peantechnologiesandvehicleemissionregulationsimple-mentedinChinahadbeenalmostthesameasthoseper-formedinEurope(Xieetal.,2006).Inaddition,comparedwithMOBILEmodels,theresultscalculatedbyCOPERTfitbetterwiththemeasuredemissionfactorsinChina(SongandXie,2006),andCOPERTwasmoreapplicableincoun-trieswithdifferentemissionregulationsanddeficienttrans-portationdata,anditwaswellusedtodevelophighspatial

Atmos.Chem.Phys.,8,1–20,2008

4Y.Boetal.:Tempo-spatialvariationofanthropogenicNMVOCsemissions

Table2.China’svehiclemileagetravelled(km)ofeachvehiclecategorybasedonliteratures.

PC

31510±8942

MB20002

Taxi153623

LDV129623±8343

LDV232983±6643

HDV55680

Coaches51040

Buses104507

MC11554±2314

PC:Passengercars;MB:Microbuses;LDV1:GasolineLightDutyVans;LDV2:DieselLightDutyVans;HDV:HeavyDutyVans;MC:Motorcycles.

Table3.Vehiclemileagetravelled(km)ofeachvehiclecategoryfrom1980to2005.

Vehiclecategories198019851990199520002005Passengercars250002500030000350003750040000Light-dutyvans200002200025000275003000035000Heavydutyvans400004000050000525005500057500Busesandcoaches600006000065000650007000070000Motorcycles

10000

10000

11000

12000

13000

13500

andtemporalresolutionemissiondatabase(ZachariadisandSamaras,1999).Therefore,COPERTIIIprogrammewasusedtocalculatetheemissionfactorsofNMVOCsemissionsineachprovinceofChinafortheperiodof1980–2005underthedrivingcyclesofurban,ruralandfreeway.Thedetailedmethodologywasreferredtotheestablishedmethodology(CaiandXie,2007).

AccordingtotheactualsituationinChina,vehicleswereclassifiedintoninecategoriesofsmall-sizegasolinecars,small-sizepetrolcars,middleandlarge-sizepassengercars,gasolinelightdutyvans,diesellightdutyvans,gasolinemid-dledutyvans,dieselmiddledutyvans,dieselheavydutyvansandmotorcycles,thepopulationdataofwhichwereob-tainedfromtheChineseannualstatisticalyearbooks.

Anotherimportantparameterrequireddeterminationwastheannualvehiclemilestravelled(VMT).TheVMTdatawereestimatedbasedontheannualaveragedatareportedinliteratures(Zhangetal.,2004;He,1999;Huetal.,2002;Yuetal.,2008;Yeetal.,2007;Fuetal.,2008),whichweresummarizedinTable2.Upontheconsiderationthattheroadlengthhadincreasedfrom890000kmin1978to3580000kilometersin2007,andthefreewaylengthhadbrokenthrough50000kmoverthepasttwentyyears,itisthereforebelievedthattheannualVMTforeachvehiclecat-egoryincreasedaswellinChina.Thus,theannualaverageVMTforeachvehiclecategoryinChinawasestimatedforperiodof1980–2005basedonTable2,asshowninTable3.2.2.2

Fossilfuelcombustion

Thesourceoffossilfuelcombustioncontainsvarioussta-tionarycombustionfacilitiesthatprovideelectricityandheat.TheNMVOCsemissionsfromthissourcearedeterminedbyfueltypesandcombustionmodes.Normally,burninggasandliquidfuelgeneratelessNMVOCsemissionsthanburningAtmos.Chem.Phys.,8,1–20,2008

Table4.NMVOCsemissionfactorsforspecificcategoriesofsta-tionaryfossilfuelcombustionin2005.

Sector

FueltypeEmissionfactorValueUnitThermalpowergeneration

Coal0.15akg/tonFueloil0.13g/m3LPG

66g/m3Naturalgas0.18g/m3Heating

Coal0.18ag/m3Fueloil0.19g/m3LPG

66g/m3Naturalgas0.18g/m3Industrialandcommercialconsumption

Coal0.18ag/m3Fueloil0.15g/m3Coalgas0.00044ag/m3LPG

66g/m3Naturalgas0.18g/m3Urbanresidentconsumption

Coal0.6a

g/m3Coalgas0.00044ag/m3LPG

66g/m3Naturalgas0.18g/m3Ruralresidentconsumption

Coal0.6ag/m3LPG

66

g/m3

aexperimentaldataobtainedfromScienceResearchInstituteof

EnvironmentalProtection,GuangZhou,1998.OthersreferredtoAP42report(1999).

solidfuellikecoalbecausethecombustionefficiencyofsolidfuelislowerandlargecombustionplantsemitlessNMVOCsthanhouseholdboilerswithmuchsmallertonnagewhencon-sumingthesameamountofenergy.Fossilfuelcombus-tionsourcewasdividedintofivesubsourcesbasedonthedifferentfueltypesconsumed:thermalpowergeneration,heating,industrialandcommercialconsumption,urbanres-identconsumption,andruralresidentconsumption.Table4presentstheNMVOCsemissionfactorsforspecificfossilfuelcategoriesin2005,andtheemissionfactorsofthesecategoriesinotheryearswerecalculatedbyEq.(2).Ac-tivitydataofthissourcewerecollectedfromChinaEnergyStatisticalYearbooks(1986–2006),ChinaRuralStatisticalYearbooks(1985–2006)andprovincialstatisticalyearbooks(1981–1984).

www.atmos-chem-phys.net/8/1/2008/

Y.Boetal.:Tempo-spatialvariationofanthropogenicNMVOCsemissions

5

Table5.NMVOCsemissionfactorsforspecificcategoriesofbiomassburning.SectorSubsourceEmissionfactorUnitcornstalk5.3ag/kgbiofuel

firewood5.3ag/kgbiogas0.18ag/m3openburning

straw7.5bg/kgcornstalk

10b

g/kg

abvaluesbasedonAP42(1999).

dataarefromnativeresearch(Lietal.,2007).

2.2.3Biomassburning

Thesourceofbiomassburningcontainsbiomassfuelinclud-ingbiogas,cornstalkandfirewood,andopenburningofagri-culturalwaste.Twonotablecomponentsofbiomassburningaretheincinerationofwood,charcoalandagriculturalwasteashouseholdfuel,andthecombustionofcropresiduesinopenfields(YevichandLogan,2003).Particularly,biofuelhasbeenoneofthemajorenergyconsumptiontypesinChinaforalongtime,especiallyinruralareas,wherebiofuelac-countedforover70%ofthetotalenergyconsumptionbefore1979(Caoetal.,2005).

EmissionfactorsofChina’sbiofuelburningwerereferredtoAP42duetoinformationlimitation,andtheemissionfac-torsofbiomassopenburningwerethelocallymeasureddatainChina,asarelistedinTable5.Activitydataofbiofuel,includingtheconsumptionofbiogas,cornstalk,andfire-woodashouseholdfuel,wereobtainedfromChinaEnergyStatisticalYearbooks(1986–2006),ChinaRuralStatisticalYearbooks(1985–2006),andprovincialstatisticalyearbooks(1981–1984).

Asforbiomassopenburning,theamountofcropresiduesineachprovincewascalculatedbyEq.(5).Qm=Yk,m×Rk×Pm

(5)

where:Qmistheopenburningamountofcropresidues;Yk,mistheyieldofcropkinprovincem;Rkistheproduction-to-residueratioofcropk;andPmistheopenburningproportioninprovincem.

Production-to-residueratiosandcropyieldswereshowninTable6.Inaddition,provincialproportionofopenburningofcropresidueswascalculatedbycombininginformationonregionalclimate,rurallivingstandard,energyconsump-tionmodes,andcontrolpoliciesforopenburningineachprovince.

2.2.4Storageandtransport

Volatilizationandleakageoforganicsolventsduringthestorageandtransportprocess,includingdistributionandwww.atmos-chem-phys.net/8/1/2008/

Table6.Production-to-residueratiosandcropyieldsinChina.

CroptypeProduction-to-residueratioaYield(Tg/yr)b199520002005paddy0.623185.23187.91180.59wheat1.366102.2199.6497.45corn

2.0111.99106.00139.37othergrains1.016.7011.6810.37legume1.517.8820.1021.58potato0.532.6336.8534.69oilcrops2.022.5029.5530.77cotton3.04.774.425.71hemp1.70.900.531.10sugarcane

0.1

79.40

76.35

94.52

abdatafromChinaAssociationofRuralEnergyIndustry,2000.datafromChinaRuralStatisticalYearbook,2006.

marketingofpetroleum-derivedproductssuchasgasoline,crudeoilandothervolatileorganicsolventsusedindailylife,isapotentialsourceforNMVOCsemissions.Therefore,gasolineservicestationsandstorageandtransportofgaso-lineandcrudeoilwereidentifiedassourcesforNMVOCsemissionsinthisstudy.

NMVOCsemissionsfromgasstationsmainlyinvolvewiththreeaspects:liquidloadinglosses,tankbreathlosses,andvehiclerefuelingoperationlosses.InChina,mostgassta-tionsadoptedthewayofimmergenceoildischargingandun-dergroundtanks,andhadnocontrolintheprocessofvehiclerefueling.Therefore,theemissionfactorsmeasuredunderthesamesituationinAP42wereselected,asshowninTa-ble7.

Emissionfactorsforgasolineandcrudeoilcouldbees-timatedbycalculatingtheirhydrocarbondissipationcoeffi-cients,basedontheapproachintroducedbyEnvironmentalPollutionControlduringOilStorageandTransport(XiaandSun,1992).ThefollowinghypotheseswereassumedbasedonChina’sactualsituationtoestimatereasonableNMVOCsemissionfactorsofoilandliquefiedpetroleumproductsdur-ingtheirstorage:(1)rawmaterialsandproductswerestoreduponlyforonemonth;(2)onlytheevaporativeemissionsduringthestorageofcrudeoilandgasolinewereconsidered,andevaporativeemissionsoflightfueloilwerenegligible;(3)evaporativeemissionsfromthestorageofheavyfueloilandpressurestorageforhighvolatileoilwerenegligible;(4)gasoline,crudeoilandfueloilwerestoredinfloatingrooftanks;(5)anddensitiesofcrudeoilandgasolinefromotheroilfieldswereconsistentwiththeoilstandardimple-mentedinDaQingoilfield.EstimatedNMVOCsemissionfactorsfromstorageevaporationforcrudeoilandgasolinewereshowninTable8.

Thetransportofpetroleumliquidsinvolvedmanydistinctoperations,eachofwhichwasapotentialsourceofevapora-tionloss.Forexample,crudeoilwastransportedfrompro-ductionplantstoarefinerybytankers,barges,railtankcars,tanktrucks,orpipelinesandrefinedpetroleumproductswere

Atmos.Chem.Phys.,8,1–20,2008

6Y.Boetal.:Tempo-spatialvariationofanthropogenicNMVOCsemissions

Table7.SuitableNMVOCsemissionfactorsforChina’sgasstationsselectedfromAP42.

EmissionRatemg/Llb/103galThroughputThroughput88012013207.3111EmissionSource

FillingundergroundtanksubmergedfillingUndergroundtankbreathingandemptying

Vehiclerefuelingoperationsdisplacementlosses(uncontrolled)

Spillage

Table8.NMVOCsemissionfactorsforstorageofcrudeoilandgasoline.

TypeDensityDissipationcoefficientEmissionfactors(ton/m3)(kg/dm3)kg/m3kg/toncrudeoil0.8510.003480.10440.123gasoline

0.760

0.00395

0.1185

0.156

conveyedbytanktruckstoservicestations,commercialsup-plystations,andlocalbulkstorageplants.Besides,fueloilsandotherpetroleumproductshadsimilartransportpaths.ByreferringtothenationalstandardofGB11085-89,thevalueof1.6036g/kg,themeanvalueoftransport,loadinganddis-charginglosses,wasadoptedastheemissionfactorforoiltransport.

Quantitiesofpetroleumfromtransportandstoragewereunavailablefromstatisticaldata.Therefore,theaggregatedquantitiesofpetroleumfromvariousprocesses,includingimportation,exportation,refinery,andvehiclerefuelingwereestimatedastheactivitydata.Particularly,provincialdatafrom1985to2005wereobtainedfromChinaEnergyStatis-ticalYearbooks(1986–2006),whilethosein1980wereesti-matedbyallocatingthenationaldatain1980totheprovin-ciallevelbasedontheprovincialproportionin1985.2.2.5

Solventutilization

Solventutilization,whichhadoccupiedapproximately25%ofNMVOCsemissionsperyearinUSfrom1991to2006(EPA,2007),wasoneimportantsourceofanthropogenicNMVOCsemissions.Emissionfactorsofthissourcein2005wereselectedfromthe1995datainAP42,duetothelackofappropriateemissionfactorsinChina,andtheemissionfac-torsofthissourceinotheryearswereestimatedbyEq.(3).Majorsolventcategoriesandthecorrespondingemissionfac-torsin2005wereshowninTable9.

Activitydataofthissource,whichmainlyincludepop-ulation,productionoutputs,numberofplantsandproduc-tionlinesofsolvent,wereobtainedfromChinaMarketYear-books(1999–2006),ChinaLightIndustryYearbooks1949–1983(1985),ChinaLightIndustryYearbooks(1985–2006),ChinaIndustrialEconomicStatisticalYearbooks(1988–Atmos.Chem.Phys.,8,1–20,2008

80

0.7

Table9.NMVOCsemissionfactorsforsolventutilization.

Sector

EmissionUnitsfactorCancoating

100Mg/yr/wireMagnetwirecoating

84Mg/yr/wireAgriculturemachinessurfacecoating236Mg/yr/plantSurfacecoatingofplasticparts236Mg/yr/plantforbusinessmachines

Metalfurnituresurfacecoating218Mg/yr/plantMucilagegluefiber50Mg/yr/plantTypewriter

60Mg/yr/plantOtherofficialres25Mg/yr/plantWoodfurniture

0.4kg/pieceMachinetoolequipment0.4kg/pieceTextilefabricprinting

81.4kg/MgfabricAutomobile&lightdutytruck21.2kg/vehiclesurfacecoating

Largeappliancesurfacecoating0.2kg/productionBicyclesurfacecoating0.3kg/bikeArchitecturesurfacecoating0.051∗kg/yr/capitaAutomobilerecoating0.021∗kg/yr/capitaPainting0.01∗Drycleaning

0.02∗kg/yr/capitakg/yr/capitaSolventdegreasing

0.044∗kg/yr/capitaCommercial/Consumersolventuse

0.1∗

kg/yr/capita

EmissionfactorswereselectedfromAP42inChapter4(EPA,1995),and∗emissionfactorspercapitawerecalculatedbymul-tiplyingconversioncoefficientof0.0246,usingper-capitaincomeasproxyvariable,basedonthedifferencesinlivingstandardsbe-tweenChinaandtheUS.

2006),andChinaStatisticalYearbooks(1981–2006)fortheperiodof1990–2005.Datain1980and1985werecollectedfromprovincialyearbooksofthoseyears.2.2.6

Industrialprocesses

Productionprocessesoforganicchemistry,inorganicchem-istry,food,woodproductionandotherindustriesarepoten-tialsourcesforNMVOCsemissions.Duetolittleinfor-mationofNMVOCsemissionfactorsoftheseindustriesinChina,emissionfactorsforthesesourcesin2005werese-lectedfromAP42,basedontheconsiderationuponthetech-nicaldifferencesinthesesourcesbetweenChinaandUS,as

www.atmos-chem-phys.net/8/1/2008/

Y.Boetal.:Tempo-spatialvariationofanthropogenicNMVOCsemissions

7

Table10.NMVOCsemissionfactorsforindustrialprocesses.SectorEmissionSectorEmissionfactorsfactors(g/kg)(g/k0g)Syntheticfibre73.4Ceramicclay0.003manufacturingSyntheticrubber7.17Portlandcement0.177manufacturingPlastic2.2Bricksandrelated0.033clayproductsPaintproduction15Glassfiber3.15manufacturingSyntheticammonia4.72Coalcleaning0.2Vegetableoil2.45

Plywood

Fig. 1.0.5 Source contribution to the total emission in 2005. processingmanufacturingMaltbeverages0.74Pulp

0.25manufacturingSugarprocessing0.6Printingink0.003Cokeproduction

1.25

Charcoal

101

manufacturing

EmissionfactorswereselectedfromAP42inChapter6–11(EPA,

1995).

showninTable10.Activitydataforthesesourceswerere-ferredtothesamedatasourcesassolventutilization.3Resultsanddiscussion3.1

Emissionsin2005

TotalNMVOCsemissionsinChinawas16.5Tgin2005,in-cludingvehicles5.5Tg,biomassburning3.8Tg,industrialprocesses2.8Tg,fossilfuelcombustion2.0Tg,solventuti-lization1.9Tg,andstorageandtransport0.5Tg.Thecon-tributionofeachsourcetothetotalemissionsisshowninFig.1,whichrevealsthatvehiclesandbiomassburningwerethedominantsourcesonthenationalscaleandcontributed33.3%and23.3%,respectively.

DetailedcontributionofsubsectorsorsubsourceswithinthesixmainsourceswasshowninFig.2.Regardingve-hicularemissions,motorcycleswerethedominantsubsourcewhichoccupied49.8%ofvehicularemissionsin2005,duetotheirmuchlargerpopulationandhighVMT.Whyweremo-torcyclesstillthemostinfluentialsubsourceovertheyears,sincemanycitieshadrestrictedthedevelopmentofmotorcy-clesinurbanareasince1990s?Therearetwomajorreasons:oneisthatmotorcyclepopulationinruralareasincreasedfastwww.atmos-chem-phys.net/8/1/2008/

Vehicles33.3%Storage and transport3.3%Biomass burningSolvent 23.3%utilization11.4%Fossil fuel Industrial combustionprocesses12.0%16.8%Fig.1.Sourcecontributiontothetotalemissionin2005.

withtherapidgrowthofeconomicsandimprovementofru-ralroads;theotheristhetremendoustransferofmotorcyclesfromcitiestothevastcountryside,whichwasstimulatedbytheabsenceofmotorcyclerestrictionpoliciesinthecoun-tryside,thelow-graderuralroadsandlowpriceofmotorcy-clesthatforcedchoosingmotorcyclestosatisfytheneedofgoodsandmanpowertransportationinruralareas.There-fore,themotorcyclepollutiononthenationalscalewasnotwellcontrolledbutinsteadcontinuedworsening.Heavydutyvans,whosepopulationandVMThadrapidlyincreasedwiththefasteconomicgrowth,contributed27.0%ofthetotalve-hicularNMVOCsemissions.Inaddition,passengercarsoc-cupied15.6%ofvehicularemissions,followedbylightdutyvansandbuses,whichoccupied4.0%and3.0%,respectively.Withrespectofbiomassburning,biofuel(cornstalk,fire-wood,andbiogasashouseholdfuels)hadoccupied76.3%ofthetotalNMVOCsemissionsofthissource,andemis-sionfromcornstalkburningwastwicemorethanthatfromfirewoodburning,duetothedecreasedamountofavailablefirewoodowingtothebanningofhewingwoods.Moreover,openburningcontributedtheremaining23.7%ofemissions,ofwhichtheburningofstrawandcornstalkaccountedfor13.2%and10.5%,respectively.Theemissionfrombiogaswasonly0.0001Tg,duetothelowemissionfactorandthequitesmallconsumption.Industrialandcommercialcon-sumptionwasthedominantsubsourceofNMVOCsemis-sionswithinthesourceoffossilfuelcombustion,accountingfor43.3%,duetothegrowingdemandforfossilfuelstimu-latedbytherapiddevelopmentofindustrialisation.Anotherimportantcontributorwasurbanconsumption.Urbanresi-dents,occupying43%oftotalChinesepopulation,werere-sponsibleforthreetimesmoreemissionsthanruralresidentsin2005,duetothehugeconsumptiondiscrepancybetweenurbanandruralareas,thoughemissionfactorsforthefueltypesadoptedincitieswerelowerthanthoseusedinthecountryside.

Atmos.Chem.Phys.,8,1–20,2008

8

Light Duty Vans4%Passenger Cars15.6%

Y.Boetal.:Tempo-spatialvariationofanthropogenicNMVOCsemissions

Open burning Straw13.2%Biofuel Firewood24.9%

Open burning Cornstalk10.5%

Heavy Duty Vans27%

Buses3.5%

Biofuel Cornstalk51.4%

SourceofvehiclesPetroleum Refining10.1%Coke Production11.5%

Synthetic fibre39.9%

Motorcycles49.8%Synthetic ammonia7.9%Cement6.9%Ceramic Clay 5.7%Plywood 5.3%Others12.8%

Source of biomass burning Rural resident consumption11.4%

Thermal power generation7.6%

Heating1.5%Industry and commerce consumption43.3%

Urban resident consumption36.3%

Source of industrial processes Magnet Wire Coating13.5%Surface Coating for Machines29.8%

Others19.8%

Source of fossil fuel combustion gasoline service station20.2%Storage Crude Oil0.9%

Transport Crude Oil53.0%

Storage Gasoline0.6%

Large Furniture ApplianceSurface 10.3%Coating

13.0%

Commercial/Consumer Solvent Use7.0%

Automobile Surface Coating6.6%

Transport Gasoline25.2%

Source of solvent utilization Source of storage and transport Fig.2.SubsectorcontributionswithinsixmajorsourcesofNMVOCsemissions.

Fig. 2. Subsector contributions within six major sources of NMVOCs emissions. Asforindustrialprocesses,thesubsectorsofsyntheticfi-bre,cokeproduction,petroleumrefining,andsyntheticam-moniawerethemajorcontributorstotheNMVOCsemis-sions,accountingfor39.9%,11.5%,10.1%,and7.9%,re-spectively.

Regardingsolventutilizationemissions,surfacecoating,especiallythatforagriculturalmachinesandplasticpartsofbusinessmachines,wasthedominantsubsource,duetoitshighemissionfactorandwideuse.

Withrespectofsolventstorageandtransport,transportcontributed78.2%ofthetotalemissions,ofwhichthetans-portofcrudeoilandgasolinewerethedominantsubsectors,responsiblefor53.0%and25.2%,respectively,muchmorethanthosefromgasolineservicestationsandoilstorage.

Insum,themajorsubsourcesforNMVOCsemissionsinChinain2005weremotorcycles,biofuelburning,heavydutyvans,syntheticfibreproduction,biomassopenburning,andindustrialandcommercialconsumption,whichaccountedforAtmos.Chem.Phys.,8,1–20,2008

18.4%,17.8%,7.7%,6.6%,5.5%,5.4%ofthetotalemis-sions,respectively.

Provincialemissionsandsourcecontributionwerede-pictedinFig.3,whichshowsthatGuangdongwastheprovinceproducingmostNMVOCsemissionsthroughoutChina,withatotalNMVOCsemissionof1.83Tgin2005,whichoccupied11.1%ofthetotalemissioninChina,duetoexplosiveincreaseofmotorcycles,prosperouselectricpowergeneration,andboomingindustrialandcommercialmarket.Simultaneously,provincesofJiangsu,Zhejiang,andShan-donggenerated1.46Tg,1.27Tg,and1.12Tgoftheemis-sions,respectively.Thesefourprovinceslocatedineast-erncoastalareaswerealldevelopedregionswithhighGDP,highpopulationdensityandmodernfarming.Onthecon-1trary,Xizang,Qinghai,andNingxia,threeprovinceslocatedinthevastwesterninlandareaswithasparsepopulationandsloweconomicgrowth,generatedonly0.03Tg,0.05Tg,and0.09Tg,respectively,in2005.

www.atmos-chem-phys.net/8/1/2008/

Y.Boetal.:Tempo-spatialvariationofanthropogenicNMVOCsemissions

2.0 1.8 1.6 NMVOCs emissions (Tg)1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 

GuangdongJiangsuZhejiangShandongHenanHebeiSichuanGuangxiFujianShanghaiLiaoningHeilongjiangAnhuiHunanHubeiShanxiJiangxiNeimengguJilinYunnanShaanxiGuizhouBeijingXinjiangChongqingGansuTianjinHainanNingxiaQinghaiXizang9

Industrial ProcessesSolvent utilizationStorage and TransportionFossil Fuel CombustionBiomass BurningVehicles

Province

Fig.3.AnthropogenicNMVOCsemissioninventoryforeachprovinceinChina(2005).

Fig. 3. Anthropogenic NMVOCs emission inventory for each province in China 100%(2005). 90%

percentage80%70%60%50%40%30%20%10%0%

BeijingShandongYunnanHainanHebeiGuangdongGansuHenanJiangsuShaanxiFujianXizangXinjiangHunanTianjinQinghaiNingxiaLiaoningHeilongjiangNeimengguJiangxiHubeiJilinGuangxiAnhuiChongqingSichuanGuizhouZhejiangShanxiShanghaiGroup 1

Vehicles

Storage and Transportion

Province

Fossil Fuel CombustionSolvent utilization

Group 2Group 3

Biomass BurningIndustrial Processes

Fig.4.Sourcecontributions(vehicles,fossilfuelcombustion,biomassburning,storageandtransport,solventutilization,andindustrialprocesses)ofNMVOCsemissionsontheprovincialscalein2005.

Fig. 4. Source contributions (vehicles, fossil fuel combustion, biomass burning, storage and transport, solvent utilization, and industrial processes) of NMVOCs emissions on the provincial scale in 2005. www.atmos-chem-phys.net/8/1/2008/Atmos.Chem.Phys.,8,1–20,2008

10Y.Boetal.:Tempo-spatialvariationofanthropogenicNMVOCsemissions

Table11.EmissionsofNMVOCsbysourceinChinafrom1980to2005(Units:Gg).

Source198019851990199520002005vehicle203421794192330635497fossilfuel340

362

614

841

1497

1979

combustionbiomass158914831502381533543837

burningstorageand237278299319389546

transportsolvent109313691307145813621871

utilizationindustrial445551695104913272761

processestotal

39074464521194051099216491

SourcecontributionsofNMVOCsemissionsattheprovin-ciallevelweredepictedinFig.4,todistinguishthelo-caldiscrepancyofsourcecontributiontoNMVOCsemis-sionsandfindoutthedominantsourcesresponsibleforlocalemissions.Figure4showsthattheNMVOCsemissionsinShanghaiweredominatedbythesourceoffossilfuelcom-bustion,whichcontributed28.1%,withthefossilfuelcon-sumedbyindustryandcommerceaccountingfor61.2%ofthetotalfossilfuelcombustion.Theotherthirtyprovincescouldbeclassifiedintothreegroupsaccordingtothecorre-spondingdominantsourceineachgroup.Group1consist-ingofseventeenprovinceswasdominatedbyvehiclesastheprimarycontributingsource;Group2consistingofelevenprovinceswasdominatedbybiomassburningastheprimarycontributingsource;andGroup3consistingoftheprovincesofZhejiangandShanxiwasdominatedbyindustrialpro-cessesastheprimarycontributingsource,whichaccountedfor41.2%and49.2%oftheNMVOCsemissionsinZhejiangandShanxi,respectively.Particularly,syntheticfibreandcokeproductionwerethemajorsubsectorsfortheNMVOCsemissionsinZhejiangandShanxiandwereresponsiblefor92.8%and69.8%,respectively,ofthetotalNMVOCsemis-sionsfromindustrialprocesses.NMVOCsemissionsinBei-jingofGroup1andGuizhouofGroup2werecloseandwere0.27Tgand0.30Tg,respectively.However,thesourcecon-tributionsinthesetwoprovincesvaried:vehiclesaccountedfor53.1%inBeijing,duetothelargepopulationofvehiclesandfrequentidledrivingconditioncausedbytrafficjams;Ontheotherhand,biomassburningcontributed57.8%ofemis-sionsinGuizhou,anunderdevelopedagriculturalprovinceinsouthwesternChina,wherecropresiduesandfirewoodsAtmos.Chem.Phys.,8,1–20,2008

100%90%80%70%ega60%tne50%crep40%30%20%10%0%

1980

19851990

1995

20002005

Ve

BioFoSto

So

Ind

Fig.5.Contribution of NMVOCs emissions from six sources during the period ContributionofNMVOCsemissionsfromsixsourcesduring

theperiod1980–2005.

werewidelyusedfordomesticcookingbyruralhouseholdsandbiomassopenburninginthefieldoftenhappened.3.2Historicalemissionsfrom1980to20053.2.1

HistoricalNMVOCsemissions

Historicalemissionswereestimatedfortheperiodof1980to2005,asshowninTable11,whichrevealsthatChina’sNMVOCsemissionshadincreasedby4.2timesatanan-nualaveragerateof10.6%from3.91Tgin1980to16.5Tgin2005.Emissionsfromvehicles,industrialprocesses,fossilfuelcombustion,biomassburning,storageandtransport,andsolventutilizationhadgrownby27,6.2,6.0,2.4,2.3,and1.7times,respectively.Asharpincreaseofemissionsoccurredduringtheperiodof1990–1995,withtheemissionalmostdoublingfrom5.2Tgin1990to9.4Tgin1995.Meanwhile,emissionsofvehiclesandbiomassburninghadincreasedby2.4and2.5times,respectively,themajorreasonsofwhichweretherapidgrowthofvehiclepopulationwhichhaddou-bledduringtheperiod,andthepopularpracticeofopenburn-ingofcropresiduesinthefieldsincetheearly1990swhichaccountedfor25%ofNMVOCsemissionsin1995.

Table11alsoshowsthattheNMVOCsemissionsfromthesixsourcesinChinaduringtheperiodof1980–2005increasedatdifferentdegree,andthecontributionofthesesourcesineachyearwasdepictedbyFig.5,whichshowsthatthecontributionofvehicleshadincreasedfrom5.2%in1980to33.3%in2005atasteadyandrapidrate,andhadbecomethedominantcontributortoNMVOCsemissions,duetothesubstantialgrowthofvehiclepopulationinChinawhichhadincreasedby60timesfrom1.8millionto107millionduringtheperiod(ChinaAutomobileIndustryAs-sociation,2006).Asthebiggestdevelopingcountry,Chinahad57%ruralpopulationin2005(NBS,2006),andithadalonghistoryofusingagriculturalwasteandfirewoodtosat-isfyenergydemand.Thus,biomassburninghadbeenthe

www.atmos-chem-phys.net/8/1/2008/

Fig. 5. 1980-2005. Y.Boetal.:Tempo-spatialvariationofanthropogenicNMVOCsemissions

11

30 )1‐ry25 US  gT( n20 China

oissim15 e sC10 OVM5 N0 1980

1985

1990

1995

2000Fig. 7. 2005

Year

Fig.6.ComparisonofannualNMVOCsemissionsintheUSandChina.

largestemissionsourcebefore2005,especiallyintheyear

1980and1995,whenitoccupied40.7%and40.6%ofthe

totalemission,respectively.Itaccountedfor23.3%ofthetotalemissionin2005,showingadecreasingtrendofcon-tribution,thoughTable11revealsthattheemissionofthissourcecontinuedtoincrease.Meanwhile,theutilizationof

differenttypesofbiomasschangedovertheyears.Firewoodandagriculturalwaste,whichwerewidelyusedashouseholdfuelinlow-incomehouseholdsthatlackedconsumptionca-pacityforcommercialfuel,werethelargestNMVOCsemis-Fig. 8. sionsubsourcesin1980.Biomasswasopenburnedintheregions from 1995 to 2005. fieldasfarmingresiduesintheearly1990swhentheeco-nomicconditioninruralChinaimproved.Thus,openburn-ingofbiomassbecamethemajorsubsourceofNMVOCs

emissionsthereafter.TheNMVOCsemissionsfromfossil

fuelcombustionincreasedyearbyyear,andthecontribution

ofthissourcerosefrom8.7%in1980to12.0%in2005.The

sourceofindustrialprocesseshadthesameannualvariation

trendwiththesourceoffossilfuelcombustion,anditscon-tributiontothetotalemissionincreasedfrom11.3%in1980

to16.7%in2005.Thesephenomenarevealedthatbothen-ergyconsumptionandindustrializationinChinainthepast

twenty-fiveyearswereexperiencingrapidgrowth.Though

theemissionsofthesourcesofstorageandtransportandsol-ventutilizationalsohadanincreasingtrend,theirincreasing

rateswererelativelyslowerthanothersources.Therefore,

theircontributionstothetotalemissionhadbeendecreasing

yearbyyearfrom6.6%and28.0%,respectively,in1980to

3.3%and11.4%,respectively,in2005.

Figure6comparesthehistoricalNMVOCsemissions

bewteenChinaandtheUS.ItisshownthattheNMVOCs

emissionsintheUSwasovertentimesmorethanthatinChinain1980.However,NMVOCsemissionsgrewrapidlyinChinaandhadexceededthethoseintheUSin2005.TransportationhadbeenthedominantsourceofNMVOCswww.atmos-chem-phys.net/8/1/2008/)17.5 20 

1‐ rNMVOCs emissions

 y) 1g15.5 16 ‐TGDP

r(y  snn13.5 ao12 uis11.5 Y s2i1m9.5 8 01( e sC7.5 PODV4 GM5.5 N3.5 

0 

1980

19851990199520002005

Year

Fig.7.DevelopingtrendsofChinesehistoricalNMVOCsemissionsandGDP.

7

12

)gT6NMGDP

(10

) 1‐nro5y is8nsaium4

Y 21 es36

0C1( OPV24

DMGN1

200

199520002005199520002005199520002005CountyCounty‐level cityCity

Fig.Variations of NMVOCs emissions and GDP for different administration 8.VariationsofNMVOCsemissionsandGDPfordifferentadministrationregionsfrom1995to2005.emissionsintheUS,butthevehicularNMVOCsemissionstherehaddecreasedfrom12.6Tgin1980to3.68Tgin2005,duetothedecreasedevaporationofmotorcyclefuel,whichbroughtabouttherapiddecreaseofthetotalNMVOCsemis-sionssince1989(CilekandKohout,1992).Hence,thesuc-cessfulexperienceintheUSisdefinitelyvaluableforcon-trollingthevehicularNMVOCsemissionsinChina.3.2.2RelationshipbetweenNMVOCsemissionsandtheeconomyinChinaTofindoutthepivotalfactorscausingcontinuousgrowthofNMVOCsemissions,demographic,socio-economic,andgeoponicstatisticaldatawerewidelyinvestigated.StatisticalresultsshowedthattheNMVOCsemissionsandGDPwerewellpositivelycorrelated,withacorrelationcoefficientof0.99,asillustratedinFig.7,whichdemonstratedthateco-nomicgrowthhadpromotedtherapidincreaseofanthro-pogenicNMVOCsemissionsinChina.TheNMVOCsemissionsandGDPofcounties,countylevelcities,andcities,threetypesofadministrationregionswithdiverselevelsofpopulationandeconomicdevelopment,werepresentedinFig.8tofurtherinvestigatetheinfluenceof

Atmos.Chem.Phys.,8,1–20,2008

Developing trends of Chinese historical NMVOCs emissions and GDP.

Comparison of annual NMVOCs emissions in the US and China. 12Y.Boetal.:Tempo-spatialvariationofanthropogenicNMVOCsemissions

2000

2005

Fig.9.ThespatialdistributionofNMVOCsemissionsin1995,

2000,and200540 Km. atahighresolutionof40km×40km.differenteconomicdevelopmentlevelsonthecorrepondingNMVOCsemissions.Figure8revealsthatNMVOCsemis-sionsincounties,countylevelcities,andcitieswentupwiththecorrespondinggrowthofGDPthere,andtheNMVOCsemissionsperunitofGDPincountyareasweresignificantlyhigherthanthoseincountylevelcitiesandcities.In2005,countiesproducedabout50.9%oftheGDPincities,butNMVOCsemissionsincountieswas6.2Tg,closetotheAtmos.Chem.Phys.,8,1–20,2008

6.6Tggeneratedbycities.Thiswasmainlybecausemanysmall-scaleandpollution-heavyplantsweretransferredfromcitieswheremoreandmorestringentenvironmentalman-agementpolicieswereimplementedtotheunderdevelopedruralareaswithlooseenvironmentalmanagement.Further-more,theloweconomicdevelopmentlevelinmostruralar-easledtothewidespreadoflow-efficientutilizationofbio-fuelinthoseareas,likebiomassburning,whichalsocausedthehighNMVOCsemissionsperunitofGDPinruralar-eas.Thus,thewayofpollution-heavyandenergy-intensivedevelopmentincountiesshouldbeimprovedtoreducetheincreasingNMVOCsemissionsinChina’scountyareas.

3.3

SpatialdistributionofanthropogenicNMVOCsemis-sionsinChina

Torecognizethetemporalandspatialvariationcharacter-isticsoftheanthropogenicNMVOCsemissionsinChina,theNMVOCsemissioninventoriesatahighresolutionof40km×40kmwereestablishedusingtheGISmethodology,asdepictedinFig.9,whichshowsthespatialdistributionofNMVOCsemissionsin1995,2000,2005.Figure9revealsthatonenotablecharacteristicofthespatialdistributionofChina’sNMVOCsemissionsisthattheNMVOCsemissionsweremuchhigherineasternChinathaninwesternChina,andthecoastalregionshadbeenzonedasmostpollutedareasofNMVOCs.Moreover,thenortherncoastalregion,easterncoastalregion,southerncoastalregion,thethreeprovincesinnortheasternChinaandtheSichuanBasinweretypicalhigh-emissionregions,andparticularly,theYangtzeRiverdelta,thePearlRiverDelta,andthecircumjacentareasofBei-jingstoodoutwithhighestemissions.ThischaracteristicofspatialdistributionwascloselyrelatedwiththeunbalancedeconomicdevelopmentinChina,anditisobviousthatvari-ousareasofChinahadtremendousdifferencesinNMVOCsemissions,withtheannualNMVOCsemissionsatcountylevelrangingfrom462GgintheurbandistrictofShanghaito0.025GginCuonacountyofXizangAutonomousDis-trict.AnotherdistinguishedemissioncharacteristicwasthattheNMVOCsemissionsmainlypeakedoverthelargeur-banareas,andthesehighemissionregionsbegantoscatteraroundthedomain,resultingintheexpansionofhighemis-sionregionsduringtheperiodof1995–2005.Inaddition,extraordinarylyhighemissionsmainlyconcentratedinsouth-eastern,northern,andcentralregions,whichaltogethercover

35.2%ofChina’sterritory,butgenerated59.4%ofthetotalemissions.Onthecontrary,thewesternprovincesofXizang,

Shaanxi,Gansu,Qinghai,Ningxia,andXinjiang,whichal-togethercover44.6%oftheterritory,contributedonly5.8%ofthetotalemissions.Besides,thecapitalcitiesofChina,whichcovermerely4.5%ofChina’sterritory,wereresponsi-blefor24.9%ofthenationalemissionsofNMVOCsin2005.Figure9alsoillustratesthatthespatialdistributionin1995and2000wassimilarwiththatin2005,buttheemissionintensityofhighemissionregionshadbeenincreasingand

www.atmos-chem-phys.net/8/1/2008/

9. The spatial distribution of NMVOCs emissions in 1995, 2000, and 2005 at a resolution of 40 Km×Y.Boetal.:Tempo-spatialvariationofanthropogenicNMVOCsemissions13

theareaoftheseregionshadbeenexpandingovertheyears,showinganexpandingtrendofemissionsfromeasternandsouthernChinatonorthwesternChina.

Todiscernthecharacteristicsofthetemporalandspatialdistributionofemissionsfromeachsource,thespatialdis-tributionsoftheemissionsfromthesixsourceswereillus-tratedatahighresolutionof40km×40kmfor1995,2000and2005,respectively,asshowninFig.10.

ThespatialdistributionoftheNMVOCsemissionsfromvehiclesin1995,2000and2005wasshownbyFig.10a,whichrevealsthatvehicularNMVOCsemissionsinChinapresentedaclearcharacteristicofregionaldistribution,withtheemissionintensityincreasingfromnorthwesterninlandregionstosoutheasterncoastalareas.Regionswithhighemissionintensityshowedanexpandingtrendfromsouth-easterncoastalareastonorthwesterninlandregions.Be-sides,regionswithhighemissionsmainlyconcentratedincities,whichwasinaccordancewiththefactthatvehiclesinChinaconcentratedincitiesanddovetailedwiththedis-proportionaldevelopmentoffreewayconstruction:till1998,59.7%offreewaywasbuiltintheeast,and26.9%offreewaywasconstructedinthemiddle,withtheremaining13.4%ly-inginthewest(Lietal.,2000).Moreover,theexpansionofNMVOCsemissionsfromeasttowestaccompaniedChina’sWestRegionDevelopmentcommencingin2000.ThecitiesofGuangzhou,Shanghai,Beijing,andShenzhenwerethedominantemissioncontributors.AsforGuangdong,theprovincegeneratingmostNMVOCsemissionsthroughoutthecountry,thepopularuseofmotorcyclestherewasthefun-damentalcause,whichsuggestedthattheemissionstandardsofmotorcyclesshouldbeimprovedandtheuseofmotorcy-clesshouldbeputunderstringentcontrol,intheperspectiveofcontrollingNMVOCsemissions.

ThespatialdistributionoftheNMVOCsemissionsfrombiomassburningin1995,2000and2005wasshownbyFig.10b,whichillustratesthatemissionfrombiomassburn-ingconcentratedinHeilongjiang,LiaoningandJilin,thethreenortheasternprovinces;inJiangsu,ZhejiangandFu-jian,threesoutheasternprovinces;inthesouthernprovincesofGuangdongandGuangxi;intheeasternandcentralprovincesofAnhui,Jiangxi,Hebei,Henan,Shandong,HubeiandHunan;andinsouthwesternprovincesofSichuan,Yun-nan,GuizhouandChongqing.Theseprovincesgeneratedasmuchas87.2%ofthetotalNMVOCsemissionsinChina,andtheemissionintensityintheseareasshowedanincreas-ingtrend.Inthenortheast,wherecropresidueswereusedforcookingandheating,agreatdealofcornstalkandstrawcouldbeobtainedandburnedasfreefueltosubstituteforpartofthecommercialfueltogetthroughthecoldwin-ter.AnotherregionresponsibleforenormousemissionfrombiomassopenburningwassouthernChina,especiallyinJiangsu,Hubei,andthetypicalagriculturalprovinceofAn-hui,whereopenfirespotsinharvestseasonhadbeendiscov-eredseveraltimesinrecentyears(NationalSatelliteMeteo-rologicalCenter,2005).

www.atmos-chem-phys.net/8/1/2008/

ThespatialdistributionoftheNMVOCsemissionsfromindustrialprocessesin1995,2000and2005wasshownbyFig.10c,whichshowsthattheareaswithhighNMVOCsemissionsmainlyconcentratedinShandong,Beijing,theYangtzeRiverDelta,andthePearlRiverDeltain1995,whereplentyoflarge-scaleindustrialplantswerebuilt.Af-tertenyears’industrialization,theregionaldistributionofNMVOCsemissionsfromthissourcechangedsubstantially,withtheformerhighemissionareascontinuingexpandingandtheemissionintensityincentralandnortheastregionsincreasingsignificantly.Particularly,52.3%oftheNMVOCsemissionsfromindustrialsourceswereproducedineasternChina,wherethegrossindustrialoutputincreasedfrom63%in1995to69%in2005.

ThespatialdistributionoftheNMVOCsemissionsfromfossilfuelcombustionin1995,2000and2005wasshownbyFig.10d,whichshowsaclearcharacteristicofregionaldistribution:in1995,highemissionareasmainlyconcen-tratedintheYangtzeRiverDelta,thePearlRiverDelta,andtheCircum-Bohai-SeaRegionandtheemissionintensityofthesehighemissionareascontinuedtoincreasethereafter,showinganexpandingtrendofemissionfromsoutheasterncoastalandsouthernareastocentral,northwesternandnorth-easternregions.

ThespatialdistributionoftheNMVOCsemissionsfromsolventutilizationin1995,2000and2005wasshownbyFig.10e,whichillustratesthattheemissionin1995fromsol-ventutilizationmainlyconcentratedinthePearlRiverDelta,theYangtzeRiverDelta,andthecircumjacentareasofBei-jing,wheretheproductionofcoatingmaterialaccountedfor60%ofthetotal.However,thousandsofsmall-scaleplantsoutspreadtosmalltownslackofstrictenvironmentalman-agement,whichcausedthetransferofpollutantfromcitiestothecountryside,asshowninFig.10e.Thus,thehighemissionareascontinuedtoexpandthereafter,showinganexpandingtrendofemissionsinnortheastern,southeastern,andcentralChina.

ThespatialdistributionoftheNMVOCsemissionsfromstorageandtransportin1995,2000and2005wasshownbyFig.10f,whichillustratesthatareaswithhighNMVOCsemissionsfromstorageandtransportmainlydistributingaroundmegalopolisandseveralhugeoilfieldsofChina,suchasDaqing,Shengli,Liaohe,Zhongyuan,andKelamayi,duetothehugeamountoftransportofcrudeoilaroundthefields.Besides,theNMVOCsemissionsinthePearlRiverDelta,theYangtzeRiverDeltaandthecircumjacentareasofBeijingweremainlyascribedtothelargequantitiesoftheconsump-tionofoilandpetroleumproducts.Therefore,itisforesee-ablethattheemissionintensityofNMVOCsinnorthwesternChinawouldcontinuetoincrease,withthefurtherexploitofhugeoilfieldslikeKelamayi.

Atmos.Chem.Phys.,8,1–20,2008

14Y.Boetal.:Tempo-spatialvariationofanthropogenicNMVOCsemissions

1995 2000 2005

(a)

1995 2000 2005

(b)

1995 2000 2005

(c)

1995 2000

2005

(d)

1995 2000

2005

(e)

1995

2000 2005

(f)

Fig.10.AnnualchangesinspatialdistributionofNMVOCsemissionsatahighresolutionof40km×40kmfor(a)vehicles,(b)biomassburning,(c)industrialprocesses,(d)fossilfuelcombustion,(e)solventutilization,and(f)storageandtransport,in1995,2000,and2005,basedonGISmethodology.

Atmos.Chem.Phys.,8,1–20,2008www.atmos-chem-phys.net/8/1/2008/

Y.Boetal.:Tempo-spatialvariationofanthropogenicNMVOCsemissions

Table12.ComparisonofChina’sNMVOCsemissioninventories(Unit:Tgyr−1).

NMVOCsemissions

Piccotetal.(1992)

Olivieretal.(1996,2001/2002)Tonookaetal.(2001)Klimontetal.(2002)Streetsetal.(2003)Oharaetal.(2007)Inthisstudy

1980

19855.3

18.2a11.1

6.83.9

8.24.5

9.75.2

13.7b13.813.112.29.6

1990

1995

2000

2005

2010

2020

15

15.617.414.711.0

17.2

17.2c16.5

18.2

astudieddomaincontainedtheregionsofChina,HongKong,Cambodia,NorthKorea,Laos,Macau,Mongolia,andVietnamintotal.

bstudieddomainincludedtheregionsofChinamainland,HongKong,Macao,Taiwan,Mongolia,NorthKorea,SouthKorea,andGuamin

total.

cemissionestimatedin2003.

3.4

ComparisonofChina’sNMVOCsemissioninvento-riesestimatedbydifferentscholarsandtheuncertaintyanalysis

ResultsaboutChina’sNMVOCsemissionsfirstlyappearedintheglobalNMVOCsemissionsestimatedbyPiccotetal.(1992),followedbytheresearchrelevantwithChina’sNMVOCsemissionsconductedbyOlivieretal.(1996,2001/2002),Tonookaetal.(2001),andKlimontetal.(2002).Recently,Streetsetal.(2003)andOharaetal.(2007)haveestimatedtheNMVOCsemissionsinChina.Therefore,wecomparedtheresultsofabovescholarswithourestimation,asshowninTable12.

Table12showsthatPiccot’sresultin1985was5.3Tg,whichwasalittlemorethanourestimationof4.5Tginthesameyear.China’semissionof18.2Tgin1990estimatedbyOlivieretal.(1996)coveredtheterritoriesofChina,HongKong,Cambodia,NorthKorea,Laos,Macau,Mongolia,andVietnam,andtheEastAsianemissionof13.7Tgin1995estimatedbyOlivieretal.(2001/2002)coveredtheterrito-riesoftheChinesemainland,HongKong,Macao,Taiwan,Mongolia,NorthKorea,SouthKorea,andGuam.Theywerebothhigherthantheemissionsestimatedinthisstudyinthecorrespondingyears,whichwere5.2Tgin1990and9.4Tgin1995,respectively.InTonooka’semissioninventoryfor1995,coalcombustionwasthelargestsourceoccupyingap-proximately40%ofthetotalNMVOCsemissionsbasedontheemissionfactorsforboilersofdiversesizes,whileourresultsshowedthatbiomassburningwasstillthedominantemissionsourcein1995.Klimontetal.(2002)estimatedtheemissionsinChinafortheyears1990,1995and2000,basedonwhichStreetsetal.(2003)obtainedanestima-tionof17.4TgforChinain2000byfurthercalculatingtheNMVOCsemissionsfromforestandpraipiefireswhichac-countedfor1.8Tgin2000.Inaddition,theNMVOCsemis-sionsfor1995and2000intheREASinventoryestablishedbyOharaetal.(2007)werenotestimatedindependently,butinsteadcamefromtheresultsofKlimontetal.(2002)andwww.atmos-chem-phys.net/8/1/2008/

Streetsetal.(2003).Moreover,theirNMVOCemissionsfortheperiod1980–2003wereestimatedsimplybyanex-trapolationoftheNMVOCemissionsfor1995and2000,usingaproxyindicatorpersector.Therefore,wefocusedoncomparingourestimationwiththeresultsbyKlimontetal.(2002).Table12showsthattherewasnotablediscrep-ancyintheamountofNMVOCsemissions,especiallyfortheyear1990,whentheestimationbyKlimontetal.(2002)wasabout2.1timesmorethanours.Thisemissiondiscrep-ancywasmainlyascribedtothedifferencesintheselectionofemissionfactorsandactivitydataforthecorrespondinganthropogenicsourcesofNMVOCsemissions.

UponcomparisonoftheactivitydataadoptedbyKlimontetal.(2002)andinthisstudy,asshowninTable13,wefoundthatactivitydatausedbyKlimontetal.(2002)wereallhigherthanthoseadoptedinthisstudy,exceptthoseforproductionandconsumptionofoilandgas.OuractivitydataweremainlyfromChinaEnergyStatisticalYearbook(2005);ChinaAutomotiveIndustryYearbook(2002);ChinaMarket-ingYearbook(2001);andChinaForestryStatisticalYear-book(2006).Theiractivitydatain2000weremainlyes-timatedbasedonChina’senergystatisticaldatainthefor-meryearsandrelateddataofenergyconsumptionintheUS,becausetherelevantdatain2000hadnotbeenreleasedyetwhenKlimontetal.(2002)establishedtheNMVOCsemissioninventoryofChinain2000.Therefore,ouractiv-itydatabasedonChina’sstatisticalyearbookdifferredfromthoseadoptedbyKlimontetal.(2002).Forexample,coalconsumption,biofuelandotherenergyconsumption,paintproduction,paintuse,andindustrialwoodproductionwere1.22,4.19,6.77,6.22and5.05timesmorethanourrespec-tivesources.Hence,ourestimationofChina’sNMVOCsemissioninventoryin2000waslessthanthatestimatedbyKlimontetal.(2002).

Anotherimportantfactorinfluencingtheresultistheemis-sionfactor.EmissionfactorsadoptedinthisstudywereslightlylowerthanthoseusedbyKlimontetal.(2002).

Atmos.Chem.Phys.,8,1–20,2008

16Y.Boetal.:Tempo-spatialvariationofanthropogenicNMVOCsemissions

Table13.ComparisonofactivitydataforeachsectorbetweenKlimont’sresearchandthisstudy.

Sector

UnitKlimontThisetal.(2002)studyDomesticoilproduction106tons150180aDomesticgasproduction106m31840878aEnergyconsumptionCoalPJ3353227532aOilPJ86559333aGasPJ19592839aHydroPJ1742801aNuclear

PJ14760aOther(includesbiofuel)PJ

85502725aVehiclemanufacturing106vehicles5.52.1bPaintproduction103tonspaint2417357cPaintuse

103tonspaint2735440cIndustrialwood

106m3

101

20d

a:ChinaEnergyStatisticalYearbook,2005;b:ChinaAutomotiveIndustryYearbook,2002;c:ChinaMarketYearbook,2001;andd:ChinaForestryStatisticalYearbook,2006.

Forexample,thevehicularemissionin2000calculatedbyKlimontetal.(2002)was5071Gg,whichwasmuchhigherthanourestimationof3063Gg.Thiswasnotonlybecausetheiractivitydatawerelargerthanours,asshowninTable13,butalsobecausetheiradoptedemissionfactorsforvehicleswereslightlyhigherthanours.TheemissionfactorsadoptedbyKlimontetal.(2002)toestimateChina’stransportsourceemissionsin2000werebasedontheEuropeanexperienceofthemid-1980s.However,emissionstandardsequivalenttoEuropeIhavecomeintoeffectinChinasince2000.There-fore,itwasnotquitesuitabletocalculateChina’semissionin2000byusingemissionfactorsbasedontheEuropeanex-perienceofthemid-1980s.Consideringtheintroductionandexecutionofstrictervehicleemissionstandardsanddistinctpoliciesandstatutesindifferenttimes,wecalculatedChina’sregion-specificandyear-specificNMVOCsemissionfactorsforvehicles,usingCOPERTIIImodelwhichwasreliablyapplicableinChina,becausetheCOPERTIIIresultswereclosertothemeasuredemissionfactorsbyplatformexperi-mentsthantheMOBILEresults,especiallyforthegasolinevehicles(Xieetal.,2006).Therefore,itaccordsbetterwithChina’sactualsituationtocalculatetheemissionfactorsus-ingCOPERTIII.Moreover,forsomevehicletypes,emissionfactorscalculatedbyMOBILEarehigherthanthosecalcu-latedbyCOPERTIII,andthisiswhysomeChinesevehicu-larNMVOCsemissionscalculatedbyMOBILEmodelswerehigherthanourresults.

Asdescribedabove,activitydataofeachsourceobtainedfromavailableandcrediblepublishedstatisticalyearbooksandtheemissionfactorsadoptedinthisstudywerebothlowerthanthosecompiledbyKlimontetal.(2002),Streetsetal.(2003),Oharaetal.(2007),Olivieretal.(2001,2002)andAtmos.Chem.Phys.,8,1–20,2008

Tonookaetal.(2001),whichresultedinalowerestimationinthisstudythanabovestudies.

Emissioninventoriesestimatedbydifferentscholarsdif-ferredsignificantly,andtheuncertaintyoftheinventoriesde-rivesfromtheuseofdifferentemissionfactorsandactivitydata.MostoftheemissionfactorsinthisstudywereselectedfromAP42andliteratures,duetothelackoflocallymea-sureddata,whichtosomeextentcausedtheuncertaintyoftheestimatedNMVOCsinventoryinChina.Therefore,weconductedtheuncertaintyanalysisfortheemissionfactorsandactivitydata,accordingtotheUSEPA’srecommendedmethodologyforuncertaintyanalysis(EPA,2007).

Whenconductingtheuncertaintyanalysis,theemissionfactorsandactivitydatatobeusedassampledatafortheuncertaintyanalysiswerefirstlyselected,andtheprobabil-itydensityfunction(PDF)thesampledatafitbestwasde-terminedbasedontheMoran’sstatistic,followedbythecalculationofthestatisticsofthemeanvalueandthestan-darddeviationofthedeterminedPDF,usingmaximumlike-lihoodestimation.Basedontheaboveprocedures,thesam-pledataoftheemissionfactorsfromeachsubsourceofthesixsources,aswellastheactivitydatawereselected,andthePDFtheyfitbestwasdetermined.Then,theMonteCarlosimulationwasrepeatedlyimplementedwithnewinputval-uesrandomlyselectedfromwithintherespectiveprobabilitydistributionoftheemissionfactorsandactivitydataofeachsource.Foreachsource,eachemissionfactorandeachactiv-itydatarandomlyselectedwasmultipliedtoobtainasampleoftheemissionofthatsource.Whenaddinguptheemis-sionsamplesofallthesixsources,onesamplerepresentingthetotalemissionwasobtained.Thisprocesswasrepeatedlyrunfor10000times,resultinginadatasetincluding10000totalemissions,whosestatisticsofthemedian,themean,the2.5thpercentileand97.5thpercentilevalueswerecalculatedandthereforethepropagationofuncertaintyofthetotalemis-sionatthe95%confidenceintervalwasobtained.

Theabovemethodologywasfollowedtoconducttheun-certaintyanalysisoftheemissionfactorsandactivitydataofeachsource.Takingthevehiclesourceforexample,datarangeofemissionfactorsofmotorcyclesusedforuncer-taintyanalysiswas[4.38,7.30]g/km(Fuetal.,2001;HaoandWang,2007;Lietal.,2003;Wangetal.,2001a,2001b;Haoetal.,2002;Xieetal.,2006),whichbestfitthelog-normaldistribution,withameanvalueof4.79g/kmandthestandarddeviationof0.45g/km.Theprobabilitydensitydis-tributionoftheemissionfactorsselectedforothervehiclecategorieswerealsolog-normaldistributed.Specifically,thedatarangeoftheemissionfactorsofpassengercarswas[0.54,6.08]g/km(Fuetal.,2001;Lietal.,2003;Wangetal.,2001a,2001b;Haoetal.,2002;HeandWang,2006;Xieetal.,2006),withameanvalueof1.23g/kmandthestan-darddeviationof2.4g/km;thedatarangeoftheemissionfactorsoflightdutyvanswas[0.09,6.47]g/km(Lietal.,2003;Wangetal.,2001a,2001b;Haoetal.,2002;Dengetal.,2000;Fuetal.,2005;HeandWang,2006;Xieetal.,

www.atmos-chem-phys.net/8/1/2008/

Y.Boetal.:Tempo-spatialvariationofanthropogenicNMVOCsemissions17

1 Fig.11.2 ProbabilitydistributionofChina’snationalNMVOCsemissionsin2005,basedonthe10000MonteCarlosimulations.Fig. 11. Probability distribution of China’s national NMVOCs emissions in 2005, based on 2006;Weietal.,2008;YuandYu,2008),withameanvalueof1.40g/kmandthestandarddeviationof2.1g/km;thedata

4 rangeoftheemissionfactorsofheavydutyvanswas[2.26,23.0]g/km(Lietal.,2003;Wangetal.,2001a,2001b;Haoetal.,2002;Xieetal.,2006;Weietal.,2008),withameanvalueof2.98g/kmandthestandarddeviationof1.65g/km;andthedatarangeoftheemissionfactorsofbuseswas[0.78,11.2]g/km(Lietal.,2003;Wangetal.,2001a,2001b;Haoetal.,2002;Xieetal.,2006;YuandYu,2008),withameanvalueof1.21g/kmandthestandarddeviationof0.94g/km.Withrespectofthosesourceswithonlyonesampledataofemissionfactor,weassumedthattheemissionfactorsofthissourcewerenormallydistributed.SincetheemissionfactorswerefromAP42,wereferredtotheEPA’sevaluationontheemissionfactorsinAP42,whenwetriedtodetermineaconvincingstandarddeviationofthelog-normaldistribu-tionoftheemissionfactors.BasedontheEPA’sevaluation,theuncertaintyoftheemissionfactorswasratedfromAtoEwiththeuncertaintyincreasingaccordingly.Wethereforeconvertedtheletter-gradeemissionfactorratingsofAtoEtocoefficientsofvariation(CV)of±5%–100%,±100%–200%,±200%–300%,±300%–400%,and±400%–500%,respectively,fortheemissionfactorrelatednormaldistribu-tions.Forexample,theemissionfactorofsurfacecoatingformachinetoolsisratedE,forwhichweassumedaCVvalueof±500%.Thus,themeanvalueof0.4kgperpieceandthestandarddeviationof2.0kgperpiecewereused.

Regardingthosesourceswhichhadlocallymeasuredemissionfactors,liketheopenburningofcornstalk,thelo-callymeasuredemissionfactorsandthosehadbeenusedtoestimateChina’semissioninventorywereselected,whichin-cluded10(Lietal.,2007),15.7(Streetsetal.,2003),8.7(Klimontetal.,2002),and7.0g/kg(AndreaeandMerlet,2001).Sinceitisbelievedthatthelocallymeasuredemissionwww.atmos-chem-phys.net/8/1/2008/

3 the 10,000 Monte Carlo simulations. factorscouldbetterreflecttheactualsituationofopenburn-ingandthusaremorecredible,theCVvalueof±25%wasassumed.Therefore,thelog-normaldistributionwiththemeanvalueof10g/kgandthestandarddeviationof2.5g/kgwasdetermined.

Regardingtheactivitydata,thebiomassamountofopenburningwastakenasanexample:theestimatedbiomassamountofopenburninginthisstudywas107.7Tg,whiletheestimationbyCaoetal.(2006)was131.78Tg.Thedatarangewas[75,150]Tg,basedonthemaximumandmin-imumratiosofopenburningineachprovince.Thus,thelog-normaldistributionwithameanvalueof107.7Tgandastandarddeviationof19Tgwasdetermined.

Basedontheabovedeterminedstatisticsofmeanval-uesandstandarddeviationsoftheprobabilitydistributionfunctionsoftheemissionfactorsandactivitydataofeachNMVOCssource,theMonteCarlosimulationwasrepeat-edlyimplementedwithnewinputvaluesrandomlyselectedfromwithintherespectiveprobabilitydistributionoftheemissionfactorsandactivitydataofeachsource.Foreachsource,eachemissionfactorandeachactivitydatarandomlyselectedwasmultipliedtoobtainasampleoftheemissionofthatsource.Whenaddinguptheemissionsamplesofallthesixsources,onesamplerepresentingthetotalemissionwasobtained.Thisprocesswasrepeatedlyrunfor10000times,resultinginasamplesetof10000ofthetotalemission,whosestatisticsofthemedian,mean,the2.5thpercentileand97.5thpercentilevalueswerecalculatedandthereforethepropagationofuncertaintyofthetotalemissionatthe95%confidenceintervalwasobtained.Figure11illustratestheprobabilitydistributionofChina’sNMVOCsemissioninven-toryin2005,themeanvalue,the2.5thpercentilevalueandthe97.5thpercentilevalueofwhichwere16.46Tg,10.56Tgand32.04Tg,respectively.Therefore,thepropagationof

Atmos.Chem.Phys.,8,1–20,2008

118Y.Boetal.:Tempo-spatialvariationofanthropogenicNMVOCsemissions

uncertaintyofthetotalemissionin2005atthe95%confi-denceintervalwasabout[−36%,94%].Resultsshowedthattheemissionfactorsofburningcornstalkandfirewood,oftheproductionofsyntheticfibre,andofthesurfacecoatingofbusinessmachinesweremostsensitivetotheuncertaintyresult.

4Summaryandconclusions

NMVOCsemissioninventoriesincludingsixanthropogenicsourceswereestimatedfortheperiodof1980–2005,basedonstatisticaldataatcountyandprovinciallevelfromyear-booksandrelevantresearchandliteratures.Theemissionin-ventorieswerefurtherallocatedtothecountylevelandgrid-dedatahighresolutionof40km×40kmbymeansoftheGISmethodology.

ResultsshowedthatChina’sNMVOCsemissionshadin-creasedby4.2timesatayearlyaveragerateof10.6%from3.91Tgin1980to16.49Tgin2005,whichhadexceededthatintheUSin2005.ThemajorsourcesofNMVOCsemis-sionsinChinain2005weremotorcycles,biofuelburning,heavydutyvans,syntheticfibreproduction,biomassopenburning,andindustrialandcommercialconsumptionoffos-silfuel.Therefore,thesesourcesshouldbeemphasizedinprioritywhendesigningcontrolandreductionregulationsforNMVOCsemissionsinChina.

Therewasnotablediscrepancyofthesourcecontributionstothetotalemissionsduringtheperiodof1980–2005:ve-hicularemissionhadatremendousincreasefrom5.2%to33.3%;emissionoffossilfuelcombustionexperiencedanincreasefrom8.7%to12.0%;emissionofindustrialpro-cessesincreasedfrom11.3%to16.7%;whileemissionsofbiomassburning,storageandtransport,andsolventutiliza-tionhadadecreasefrom41.0%to23.0%,from6.6%to3.3%,andfrom28.0%to11.4%,respectively.However,theNMVOCsemissionsofthesesourcescontinuedgrowingduringtheperiod,whichcoincidedwellwiththeeconomicgrowthinChina,asrevealedbythecorrelationanalysisofChina’sNMVOCsemissionsandtheGDP.ThismeantthatrapideconomicdevelopmentinChinahadstimulatedthein-creaseofNMVOCsemissions.Despitetheintroductionandimplementationofstricterfuelandemissionregulationsforvehicles,thegrowthofNMVOCsemissionsinChinahadyettobebroughtundercontrol,duetothesubstantialincreaseofvehiclepopulationandroadlengthinChina.Moreover,thesestrictfuelandemissionregulationsweremainlyimple-mentedincities,whichcausedthetransferofhighemissionvehiclesfromcitiestounderdevelopedregionsandthevastruralareas.Therefore,thetotalNMVOCsemissionsdidnotdecrease,butmerelytransferredfromoneregiontoanother.ThespatialdistributionofgriddedemissionsbasedontheGISmethodologyrevealedthatremarkablyhighemissionar-easmainlyconcentratedinthemiddlereachesoftheYangtzeRiver,andthenortheastern,northerncoastal,easterncoastalAtmos.Chem.Phys.,8,1–20,2008

andsoutherncoastalareas,whichaltogethercover35.2%ofChina’sterritory,butgenerated59.4%ofthetotalemissions.Onthecontrary,thewesternprovincesofXizang,Shaanxi,Gansu,Qinghai,Ningxia,andXinjiangcovering44.6%oftheterritory,contributedonly5.8%ofthetotalemissions.Besides,thecapitalcitiesinChinacoveringmerely4.5%ofChina’sterritory,wereresponsiblefor24.9%ofthenationalemissionsofNMVOCsin2005.Moreover,annualvariationofthespatialdistributionofemissionsrevealedthatthere-gionsofhighemissionintensityhadbeenexpandingfromeastern,southernandnortheasternareastothenorthwesternareas.Therefore,NMVOCsemissionsinChinacontinuedgrowingyearbyyear,andsomehighemissionintensityre-gionsappearedandbecameconcentrated.Emissionintensityinsomeindividualregionswasespeciallyhigh,andshowedanincreasingtrendofemissionintensity,alongwiththeex-pandingtrendoftheareaofthesehighemissionintensityregions.

Acknowledgements.ThisworkwassupportedbyagrantfromtheMajorStateBasicResearchDevelopmentProgramofChina(No.2002CB211600),andthePh.D.ProgramFoundationofMinistryofEducationofChina(No.20060001057).WewouldliketothankYuh-ShanHo,J.Zhang,andZ.Liufortheirhelpfulassistanceandsuggestions,andJiangZhihua(NationalBureauofStatistics)forprovidingmethodsonmarketsurveyandforecast.Editedby:R.Cohen

References

AllChinaStrategicResearchLtd.:ChinaMarketYearbook(1999–2006editions),ForeignLanguagePress,Beijing,1999–2006.Altshuller,A.P.:Theproductionofcarbon-monoxidebytheho-mogeneousNOx-inducedphotooxidationofvolatileorganic-compoundsinthetroposphere,J.Atmos.Chem.,13,155–182,1991.

Andreae,M.O.andMerlet,P.:Emissionoftracegasesandaerosolsfrombiomassburning,GlobalBiogeochem.Cy.,15(4),955–966,2001.

Atkinson,R.andArey,J.:Atmosphericdegradationofvolatileor-ganiccompounds,Chem.Rev.,108,4605–4638,2003.

Boeglin,M.L.,Wessels,D.,andHenshel,D.:Aninvestigationoftherelationshipbetweenairemissionsofvolatileorganiccom-poundsandtheincidenceofcancerinIndianacounties,Environ.Res.,100,242–254,2006.

Cai,H.andXie,S.D.:EstimationofvehicularemissioninventoriesinChinafrom1980to2005,Atmos.Environ.,41,8963–8979,2007.

Cao,G.L.,Zhang,X.Y.,Wang,D.,andZheng,F.C.:InventoryofatmosphericpollutantsdischargedfrombiomassburninginChinacontinent,ChinaEnviron.Sci.,25(4),389–393,2005.Cao,G.L.,Zhang,X.Y.,Zheng,F.C.,andWang,Y.Q.:EstimatingthequantityofcropresiduesburntinopenfieldinChina,Resour.Sci.,28(1),9–13,2006.

ChinaAutomobileIndustryAssociation:ChinaAutomobileIndus-tryYearbook,ChinaMachinePress,Beijing,2006.

www.atmos-chem-phys.net/8/1/2008/

Y.Boetal.:Tempo-spatialvariationofanthropogenicNMVOCsemissions

19

Cilek,C.M.andKohout,E.:Estimatingmonthlyandstate-levelNOx,SO2,VOCandCO2emissionsusingtheMSCETdatabase,annualPittsburghconferenceonmodelingandsimulation;Pitts-burgh,PA(UnitedStates);DOEProject,30April–1May1992,1992.

Deng,S.X.,Chen,J.,andLi,B.C.:EmissionfactorsofCO,HCandNOxfrommotorvehiclesonurbanmainroadinChina,ChinaEnviron.Sci.,20(1),82–85,2000.

EuropeanEnvironmentAgency(EEA):COPERTIII-computerpro-grammetocalculateemissionsfromroadtransport.Methodol-ogyandemissionfactors(version2.1),TechnicalReportNo.49,2000.

EditorialdepartmentofChinaagricultureyearbook:ChinaAgricul-tureYearbook(1996,2001,2006Editions),ChinaAgriculturePress,Beijing,1996,2001,2006.

EnvironmentalProtectionAgency(EPA):Compilationofairpollu-tantemissionfactors(AP42),FifthEdition,Chapter1–13,1995.EnvironmentalProtectionAgency(EPA):Emissionsfactoruncertaintyassessment,http://www.epa.gov/ttn/chief/efpac/uncertainty.html,2007.

EnvironmentalProtectionAgency(EPA):Nationalemis-sioninventory(NEI)airpollutantemissionstrends

data,

http://www.epa.gov/ttn/chief/trends/trends06/nationaltier1upto2006basedon2002finalv2.1.xls,2007.

Field,R.A.,Goldstone,M.E.,Lester,J.N.,andPerry,R.:Thesourcesandbehavioroftroposphericanthropogenicvolatilehy-drocarbons,Atmos.Environ.,26A,2983–2996,1992.

Fu,X.Q.,Weng,Y.B.,Qian,F.Z.,Ye,W.B.,Jin,Y.X.,andChen,Y.Y.:StudyoftheVOCsourceprofileandbenzenecompoundsemissionofvariousmotorvehicles,ActaScientiaeCircumstan-tiae,6,1056–1062,2008.

Fu,L.L.,Shao,M.,Liu,Y.,Liu,Y.,Lu,S.H.,andTang,D.G.:Tunnelexperimentalstudyontheemissionfactorsofvolatileor-ganiccompounds(VOCs)fromvehicles,ActaScientiaeCircum-stantiae,25(7),879–885,2005.

Fu,L.X.,Hao,J.M.,He,D.Q.,andHe,K.B.:AssessmentofvehicularpollutioninChina,J.AirWasteManage.,51,656–668,2001.

Guenther,A.,Hewitt,C.N.,Erickson,D.,Fall,R.,Geron,C.,Graedel,T.,Harley,P.,Klinger,L.,Lerdau,M.,Mckay,W.A.,Pierce,T.,Scholes,B.,Steinbrecher,R.,Tallamraju,R.,Tay-lor,J.,andZimmerman,P.:Aglobal-modelofnaturalvolatileorganiccompoundemissions,J.Geophys.Res.,100(D5),8873–8892,1995.

Hao,J.M.,Wu,Y.,Fu,L.X.,He,D.Q.,andHe,K.B.:StudyoftheurbanvehicularemissioncontrolprogramsysteminChina,J.Appl.Meteorol.,13,195–203,2002.

Hao,M.andWang,D.D.:StudyoftheurbanvehicularpollutioncontrolstrategyinChina,Sci.Technol.Inf.,25,143–144,2007.He,C.Y.andWang,Q.D.:Vehicleemissionfactorsdetermina-tionusingCMEMinBeijing,Res.Environ.Sci.,19(1),109–112,2006.

He,D.Q.:Studyonurbanvehicularpollutionassessmentsystemandemissioncontroltarget,TsinghuaUniversityDoctorThesis,1999.

Hu,B.X.,Zhu,H.,andTan,W.X.:AdetailedlistforexhaustindevelopmentmodelsofautomobiletechnologyinChina,J.WuhanAutomotivePolytechnicUniv.,2,99–102,2002.

www.atmos-chem-phys.net/8/1/2008/Klimont,Z.,Streets,D.G.,Gupta,S.,Cofala,J.,Fu,L.X.,andIchikawa,Y.:Anthropogenicemissionsofnon-methanevolatileorganiccompoundsinChina,Atmos.Environ.,36,1309–1322,2002.

Li,M.X.,Han,Z.H.,andWu,J.:Analysisandproposalondevel-opmenttrendsofairportconstructioninChina,Rev.EconomicRes.,43,13–27,2000.

Li,X.H.,Wang,S.X.,Duan,L.,Hao,J.,Li,C.,Chen,Y.S.,andYang,L.:ParticulateandtracegasemissionsfromopenburningofwheatstrawandcornstoverinChina,Environ.Sci.Technol.,41,6052–6058,2007.

Li,W.,Fu,L.,X.,Hao,J.,M.,Ma,H.,Li,S.Q.,andHu,W.:Emissioninventoryof10kindsofairpollutantsforroadtrafficvehiclesinChina,UrbanEnv.UrbanEcol.,16,36–38,2003.Loibl,W.,Orthofer,R.,andWiniwarter,W.:Spatiallydisaggre-gatedemissioninventoryforanthropogenicNMVOCinAus-tralia,Atmos.Environ.,27A,2575–2590,1993.

Lubkert,B.andDetilly,S.:TheOECDmapemissioninventoryforSO2,NOxandVOCinwestern-Europe,Atmos.Environ.,23,3–15,1989.

NationalBureauofStatistics(NBS):ChinaEnergyStatisticalYear-book(1986,1987,1991–1996,1997–1999,2000–2002,2006editions),ChinaStatisticsPress,Beijing,1986–2006.

NationalBureauofStatistics:ChinaStatisticalYearbook(1981–2006editions),ChinaStatisticsPress,Beijing,1981–2006.

NationalBureauofStatisticsteamofruralsocio-economicinves-tigation:ChinaRuralStatisticalYearbook(1985-2006editions),ChinaStatisticsPress,Beijing,1985–2006.

NationalSatelliteMeteorologicalCenter:Reportonmeteorologicalsatellitefiremonitoring,4,2005.

Ohara,T.,Akimoto,H.,Kurokawa,J.,Horii,N.,Yamaji,K.,Yan,X.,andHayasaka,T.:AnAsianemissioninventoryofanthro-pogenicemissionsourcesfortheperiod1980–2020,Atmos.Chem.Phys.,7,4419–4444,2007.

Olivier,J.G.J.,Bouwman,A.F.,Maas,C.W.M.vander,Berdowski,J.J.M.,Veldt,C.,Bloos,J.P.J.,Visschedijk,A.J.H.,Zandveld,P.Y.J.,andHaverlag,J.L.:DescriptionofEDGARVersion2.0:Asetofglobalemissioninventoriesofgreenhousegasesandozone-depletingsubstancesforallanthro-pogenicandmostnaturalsourcesonapercountrybasisandon1degreex1degreegrid,RIVMreport771060002,1996.

Olivier,J.G.J.,Berdowski,J.J.M.,Peters,J.A.H.W.,Bakker,J.,Visschedijk,A.,J.H.,andBloos,J.P.J.:ApplicationsofEDGARincludingadescriptionofEDGAR3.2:refer-encedatabasewithtrenddatafor1970–1995,RIVMreport773301001/NRPreport410200051,2001/2002.

Piccot,S.D.,Watson,J.J.,andJones,J.W.:Aglobalinven-toryofvolatileorganiccompoundemissionsfromanthropogenicsources,J.Geophys.Res.,97(D9),9897–9912,1992.

Ren,C.F.:ChinaIndustrialEconomicStatisticYearbook(1988–2006editions),ChinaStatisticsPress,Beijing,1988-2006.

Seinfeld,J.H.,Pandis,S.N.,andNoone,K.:Atmosphericchem-istryandphysics:fromairpollutiontoclimatechange,Phys.Today,51(10),88,doi:10.1063/1.882420,1998.

Shen,M.J.,Hao,J.M.,andWang,L.T.:VOCemissionsitua-tionandcontrolmeasuresofgasstationinChina,Environ.Sci.,27(8),1473–1478,2006.

Song,X.Y.andXie,S.D.:DevelopmentofvehicularemissioninventoryinChina,Environ.Sci.,27,1041-1045,2006.

Atmos.Chem.Phys.,8,1–20,2008

20Y.Boetal.:Tempo-spatialvariationofanthropogenicNMVOCsemissions

StateForestryAdministration:ChinaForestryStatisticalYearbook2006,ChinaForestryPublishingHouse,2007.

Tonooka,Y.,Kannari,A.,Higashino,H.,andMurano,K.:NMVOCsandCOemissioninventoryinEastAsia,WaterAirSoilPoll.,130,199–204,2001.

Wang,B.G.,Zhang,Y.H.,Wu,Z.Q.,andChen,L.Y.:TunneltestformotorvehicleemissionfactorsinGuangzhou,Res.Environ.Sci.,14(4),13–16,2001a.

Wang,B.G.,Zhang,Y.H.,Zhu,C.J.,Yu,K.H.,Chen,L.Y.,andChen,Z.Y.:Astudyoncitymotorvehicleemissionfactorsbytunneltest,Environ.Sci.,22(2),55-59,2001b.

Xie,S.D.,Song,X.Y.,andShen,X.H.:CalculatingvehicularemissionfactorswithCOPERTIIImodelinChina,Environ.Sci.,27,415–419,2006.

Wei,W.,Wang,S.X.,Chatani,S.,Klimont,Z.,Cofala,J.,andHao,J.M.:Emissionandspeciationofnon-methanevolatileorganiccompoundsfromanthropogenicsourcesinChina,Atmos.Envi-ron.,42,4976–4988,2008.

Xia,Y.M.andSun,L.K.:EnvironmentalPollutionControlduringOilStorageandTransport,ChinaPetrochemicalPress,Beijing,1992.

Atmos.Chem.Phys.,8,1–20,2008Yang,B.:ChinaLightIndustryYearbook1949-1983,EncyclopediaofChinaPublishingHouse,Beijing,1985.

Yang,Z.P.:ChinaLightIndustryYearbook(1985–2006editions),ChinaLightIndustryYearbookPress,1985–2006.

Ye,S.B.,Wang,Q.D.,andHe,X.:Tianjinon-roadvehicleactivitystudy,JournalofBeijingTechnologyandBusinessUniversity(NaturalScienceEdition),2,28–31,2007.

Yevich,R.andLogan,J.A.:Anassessmentofbiofueluseandburn-ingofagriculturalwasteinthedevelopingworld,GlobalBio-geochem.Cy.,17(4),1095,doi:10.1029/2002GB001952,2003.Yu,Y.C.andYu,M.Y.:Estimationandforecastoffor-hiremotorvehicleemissionvolume,J.HighwayandTransportationRes.Dev.,6,154–158,2008.

Zachariadis,T.andSamaras,Z.:Anintegratedmodelingsystemfortheestimationofmotorvehicleemissions,J.AirWasteManage.,49,1010–1026,1999.

Zhang,S.Y.,Bai,Y.H.,andLi,J.L.:DistributionofSOfromfixedcombustionsourcesin2,NOx,COandHCemissionsYangtzeRiverDelta,YunnanEnviron.Sci.,19,195–200,2000.

Zhang,Y.H.,Shao,M.,andYu,K.H.:Vehicleemissions,envi-romentalimpactandcontrol,ChemicalIndustryPress,Beijing,2004.

www.atmos-chem-phys.net/8/1/2008/