Emissions of ozone precursors from stationary sour

AtmosphericEnvironment34(2000)2183}2204Emissionsofozoneprecursorsfromstationarysources:acriticalreviewM.Placet*,C.O.Mann1,R.O.Gilbert,M.J.NieferPacixcNorthwestNationalLaboratory,901DStreet,S.W.Suite900,Washington,DC20024,USAReceived6July1998;accepted11September1999AbstractThispaperdiscussesandcritiquesmethodsusedtoestimateemissionsof,andcreatebothaggregateanddetailedmodelinginventoriesfor,nitrogenoxides(NOx),volatileorganiccompounds(VOC)andcarbonmonoxide(CO),themainpollutantsinvolvedinozoneformation.Emissionsofsulfurdioxide(SO2)andmethodstoprojectemissionsintothefuturearealsobrie#ydiscussed.Manyimprovementshavebeenmadeinemissionsinventoriesoverthepastdecade.Forexample,therequireduseofcontinuousemissionmonitors(CEMS)hasproducedsite-specicemissionsestimatesfromalmostallUSelectricutilitypowerplants,whicharethemajorstationarysourceofNOx.However,manydataqualityissuesremain.Forexample,theoverallqualityofstandardizedemissionfactorsisverypoor.Inaddition,uncertaintieshavebeenintroducedbyuseofsimplisticassumptionsontheexistentlevelofemissioncontrol.EventheuseofCEMShasnoteliminateduncertaintyinemissionsfrompowerplants,becausemethodstodealwithmissingdatacanintroducebias.EmissionsdataforMexicoarenotcomprehensive,makingozonemodelinginUSborderregionsdi$cult.DataforVOCspeciationisoutdated,andcrudedataisoftenusedtodisaggregateemissionstothenelevelofspatialandtemporaldetailneededforatmosphericmodeling.Itisdi$culttomakegeneralstatementsabouttheimportanceofeachoftheseproblems,becausetherearenoreliableestimatesoftheoveralluncertaintyofemissionsvalues,andbecausetheimpactofemissioninventoryerrorsisverysitespecic.TheEmissionsInventoryImprovementProgram(EIIP)initiatedbytheUSEnvironmentalProtectionAgencypromisestoenhancethequalityoffutureinventories,mainlythroughcommunicationofbestpracticesamongstateagencies.Furtherinventoryimprovemente!ortsmustbefocusedonproblemsthatmoststronglyin#uencepoorpredictionofozoneconcentrations.Targetsforimprovementcouldbebasedoncomparisonofphotochemicalmodelingresultstoobservedconcentrations,coupledwithothertechniquesthatbetterexplainsource-receptorrelationships.(2000ElsevierScienceLtd.Allrightsreserved.Keywords:Emissionsinventories;Emissionsuncertainty;Emissionsprojections;Airpollution;Ozoneprecursors1.Introduction:theimportanceofemissionestimatesandtheiruncertaintiesTounderstandthecomplexprocessofozoneforma-tionandformulateozonemanagementapproaches,the*Correspondingauthor.Tel.:[phone];fax:[phone].E-mailaddress:[e-mail](M.Placet)1CharlesMannoftheUSEnvironmentalProtectionAgency,NationalRiskManagementResearchLaboratory,AirPollu-tionPreventionandControlDivision,ResearchTrianglePark,NorthCarolina,contributedthesectiononareasourcesandtotheplanningofthee!ort.scienticandpublicpolicycommunitiesemployvarioustypesofphotochemicalmodels.Thesemodelsbasetheircalculationsonemissionsinventories,alongwithmeteorologicaldataandtheoreticalchemicaltrans-formationalgorithms.Thesemodelsoftenpoorlypredictobservedozonelevels.Althoughpoorcharac-terizationofchemicalmechanismsorotherpotentialcausesalsomustbeexplored,thepredictionerrorisoftenattributedtodecienciesintheemissioninven-tories.Variousstudieshaveexaminedthesensitivityofozoneconcentrationstovariationsinemissionlevels(e.g.,Hannaetal.,1998;Jiangetal.,1997;Berginetal.,1999.).Suchstudiesindicatethatthesensitivityofozoneconcentrationstoerrorsintheemissioninput1352-2310/00/$-seefrontmatter(2000ElsevierScienceLtd.Allrightsreserved.PII:S1352-2310(99)00464-1dataisverydependentontheVOC/NOxratioandothersite-specicconditions.Importantemissionsourcecategoriesininventoriesincludenaturalsources(e.g.,vegetation),mobilesources(motorvehicles),andstationarysources(powerplants,industrialfacilities,residentialandcommercialestab-lishments,andamultitudeofmiscellaneoussources).Comparedtotheanalysisofmanmadesources,naturalemissionsestimationisarelativelynewlineofscienticinquiry,andconsiderableuncertaintyisassociatedwithnaturalemissionestimates.Becauseofthelargenumberofvehiclesontheroadandthevarietyofconditionsunderwhichtheyareoperated,transportation-relatedemissionsarealsosubjecttoconsiderableuncertainty.Althoughemissionsestimatesfromstationarysourcesaresometimesconsideredlesserror-pronethanthosefromnaturalandmobilesources,manyofthemethodsanddatausedtoestimatethemintroduceaworrisomedegreeoferror.2Itisworrisomenotonlybecausethedegreeoferrorcouldbehigh,butbecausestationarysourcescontributelargefractionsoftotalmassemissionsinsomelocations,andevenasmallpercentageerroronalargeemissionsourcecanintroducealargeerrorinabsoluteterms.Inaddition,thelevelofuncertaintyisnotwellquantied,anditse!ectsonatmosphericmodelingarenotfullyunderstood.Atmosphericmodelingstudieshelppolicymakersde-terminecause}e!ectrelationshipsandtargetsourcesforcontrol,resultinginexpendituresforpollutioncontrolbyUSindustryandconsumers.Todate,controlschemeshavenotbeensuccessfulineliminatingtheambientozoneproblem.Inadditiontovehicles,thesecontrolstrategiesincreasinglyfocusonsmallerstationarysour-cesthataredi$cultandexpen