Herefor A rticl e GoddardEarthSciencesandTechnologyCenterUniversityofMarylandBaltimoreCountyBaltimoreMarylandUSADepartmentofChemistryUniversityofMarylandCollegeParkMarylandUSAAlsoatDepartme ID: 609201
Download Pdf The PPT/PDF document "ValidationofSOretrievalsfromtheOzoneMoni..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
ValidationofSOretrievalsfromtheOzoneMonitoringInstrumentoverNEChinaNickolayA.Krotkov,BrittanyMcClure,RussellR.Dickerson,SimonA.Carn,CanLi,PawanK.Bhartia,KaiYang,ArlinJ.Krueger,ZhanqingLi,PieternelF.Levelt,HongbinChen,PucaiWang,andDarenLuReceived15April2007;revised14November2007;accepted16January2008;published18June2008.2008.1]TheDutch-FinnishOzoneMonitoringInstrument(OMI)launchedontheNASAAurasatelliteinJuly2004offersunprecedentedspatialresolution,coupledwithcontiguousdailyglobalcoverage,forspace-basedUVmeasurementsofsulfurdioxide(SO).WepresentafirstvalidationoftheOMISOdatawithinsituaircraftmeasurementsinNEChinainApril2005.ThestudydemonstratesthatOMIcandistinguishbetweenbackgroundSOconditionsandheavypollutiononadailybasis.Thenoise(expressedasthestandarddeviation,)is1.5DU(Dobsonunits;1DU=2.69forinstantaneousfieldofviewboundarylayer(PBL)SOdata.Temporalandspatialaveragingcanreducethenoiseto0.3DUoveraremoteregionoftheSouthPacific;thelong-termaverageoverthisremotelocationwaswithin0.1DUofzero.Underpollutedconditionscollection2OMIdataarehigherthanaircraftmeasurementsbyafactoroftwo.Improvedcalibrationsoftheradianceandirradiancedata(collection3)resultinbetteragreementwithaircraftmeasurementsonpolluteddays.Theairmass–correctedcollection3datastillshowpositivebiasandsensitivitytoUVabsorbingaerosols.ThedifferencebetweentheinsitudataandtheOMISOmeasurementswithin30kmoftheaircraftprofileswasabout1DU,equivalentto5ppbfrom0to3000maltitude.Quantifyingtheandaerosolprofilesandspectraldependenceofaerosolabsorptionbetween310and330nmarecriticalforanaccurateestimateofSOfromsatelliteUVmeasurements.Krotkov,N.A.,etal.(2008),ValidationofSOretrievalsfromtheOzoneMonitoringInstrumentoverNEChina,J.Geophys.Res.113,D16S40,doi:10.1029/2007JD008818.1.Introductionoduction2]Sulfurdioxide(SO)isashort-livedgas,producedprimarilybyvolcanoes,powerplants,refineries,metalsmeltingandburningoffossilandbiofuels.Itcanbeanoxiouspollutantoramajorplayeringlobalclimateforcing,dependingonaltitude.MostfossilfuelburningoccursnearthesurfacewhereSOisreleasedintotheplanetaryboundarylayer(PBL).WhenSOremainsneartheEarth’ssurface,ithasdetrimentalhealthandacidifyingeffectsbutexertslittleimpactonglobalclimateorradiativeforcing.EmittedSOissoonconvertedtosulfateaerosolbyreactionwithOHinairorbyreactionwithHinaqueoussolutions(clouds)[SeinfeldandPandis,1998;Chinetal.2000].Themeanlifetimevariesfrom1–2daysorlessnearthesurfacetomorethanamonthinthestratospherestratosphereKruegeretal.,2000;Benkovitzetal.,2004].Theresultingsulfateaerosol,whichcanbetransporteddistantlyinthefreetroposphere,canhaveclimateeffects,includingdirectradiativeforcingandaerosol-inducedchangesincloudmicrophysics.Theconcentrationandlifetimeof,themeteorologicalmechanismsthatloftitabovethePBL,andtheefficiencyofthosemechanismsremainmajorunansweredquestionsinglobalatmosphericchemistryandclimatescience[e.g.,Dickersonetal.,2007].2007].3]EmissioninventoriesindicatethatthelargestincreasesintroposphericSOemissionshaveoccurredinAsiaduringthelast20years[StreetsandWaldhoff,2000;Streetsetal.,2003;Larssenetal.,2006;Oharaetal.2007].Theseincreasedemissionsresultedinapositivewintertrend(17%perdecade)insulfateaerosolloadingoverAsiabetween1979and2000detectedbytheNASATotalOzoneMappingSpectrometer(TOMS)instrumentntMassieetal.,2004].TheTOMSaerosoltrendisconsistentwithapronouncedregionalincreaseinaerosolopticalJOURNALOFGEOPHYSICALRESEARCH,VOL.113,D16S40,doi:10.1029/2007JD008818,2008 Herefor A rticl e GoddardEarthSciencesandTechnologyCenter,UniversityofMaryland,BaltimoreCounty,Baltimore,Maryland,USA.DepartmentofChemistry,UniversityofMaryland,CollegePark,Maryland,USA.AlsoatDepartmentofAtmosphericandOceanicSciences,UniversityofMaryland,CollegePark,Maryland,USA.JointCenterforEarthSystemsTechnology,UniversityofMaryland,BaltimoreCounty,Baltimore,Maryland,USA.DepartmentofAtmosphericandOceanicSciences,UniversityofMaryland,CollegePark,Maryland,USA.LaboratoryforAtmospheres,NASAGoddardSpaceFlightCenter,Greenbelt,Maryland,USA.RoyalNetherlandsMeteorologicalInstitute,DeBilt,Netherlands.InstituteofAtmosphericPhysics,ChineseAcademyofSciences,Beijing,China.Copyright2008bytheAmericanGeophysicalUnion.0148-0227/08/2007JD008818$09.001of13 thicknessovercoastalareasofAsiadetectedusingAd-vancedVeryHighResolutionRadiometer(AVHRR)satel-litedata[MishchenkoandGeogdzhayev,2007].However,nohistoricsatellitedataareavailableforestimatingtrendsinanthropogenicSOemissionsoverthesameperiod.TheTOMSsensitivitytoSOwaslimitedbytheavailableinstrumentwavelengthsandlowspatialresolution(50kmatnadirand100kmaverage)tolargeSOamountsinvolcaniceruptions[Kruegeretal.,1995,2000;Carnetal.2003]andexceptionalSOpollutionevents[Carnetal.2004].GreatlyimprovedsensitivitywasdemonstratedthroughthedetectionofvolcanicandanthropogenicSOinfull-spectrumUVdataprovidedbytheGlobalOzoneMonitoringExperiment(GOME)[EisingerandBurrowsBurrowsetal.,1999;Thomasetal.,2005;Khokharet,2005]andScanningImagingAbsorptionSpectrometerforAtmosphericCartography(SCIAMACHY)[Bovensmannetal.,1999;Bramstedtetal.,2004;Richteretal.,2006].However,GOMEneeds3daysandSCIAMACHY6daystoacquireacontiguousglobalmapandhencecouldmissshort-livedpollutionevents.TheOzoneMonitoringInstru-ment(OMI)[Leveltetal.,2006]launchedonNASA’sAurasatellite[Schoeberletal.,2006]inJuly2004offersbetterspatialresolution(1324kmatnadir),coupledwithcontiguousglobaldailycoverage,forspace-basedUVmeasurementsofSO.OMISOdataarepubliclyavailablefromNASA’sGoddardSpaceFlightCenterEarthSciences(GES)DataandInformationServicesCenter(DISC)athttp://disc.sci.gsfc.nasa.gov/Aura/OMI/omso2_v003.shtml.Byusingoptimumwavelengths,theretrievalsensitivityisimprovedoverTOMSbyfactorsof10to20,dependingonlocation[Krotkovetal.,2006].ThegroundfootprintofOMIisoneeighththeareaofTOMS[Leveltetal.,2006].ThesefactorsproduceatwoordersofmagnitudeimprovementintheminimumdetectablemassofSO.TheimprovedOMIsensitivitypermitsdailyobservationsofstronganthropo-genicSOemissions(fromsmeltersandcoalburningpowerplants)[Krotkovetal.,2006,2007;Carnetal.,2007].Theseretrievalsrequirevalidationagainstindependentdent4]Insituprofilemeasurementsbylow-altitudeaircraftareimportantforvalidatingsatelliteSOretrievalsandforprovidingcriticalinformationaboutverticalprofilesofgases(SO,ozone)andaerosolsinthePBL.AlthoughsuchmeasurementsareavailableforseveralyearsovertheeasternU.S.[Taubmanetal.,2006],nodatawereprevi-ouslyavailableoverChina.HerewepresentthefirstcomparisonsoftheOMISOretrievalswithinsituaircraftmeasurementsnearShenyanginNEChina(Figure1)1)Dickersonetal.,2007].AircraftandOMIPBLSOsetsarebrieflydiscussedinsection2togetherwitherrorestimates.Parameterizationoftheairmassfactor(AMF)issuggestedtoimprovetheaccuracyoftheoperationalOMIdata.Section3describesaircraftcomparisonswithopera- Figure1.(left)Flightpathon5April2005.Aircraftprofilingflightpatterns(spirals)wereperformednearXiaoming(42.45N,123.7E)andLiaozhong(41.35N,122.648E).AllflightsdepartedfromTaoxianInternationalAirport(41.640N,123.488E)intheLiaoningregionofChina.Flightpathsweresimilaron5,7and10April2005.(http://www.atmos.umd.edu~zli/EAST-AIRE/air_camp/air_camp.htm).(right)A2-yearaverageOMISOmapovereasternChinainDobsonunits(1DU=2.69)showingpersistentareasofhighSOconcentrationsinatrianglebetweenBeijing,Shanghai,andtheSichuanbasininagreementwithemissioninventories[StreetsandWaldhoffStreetsetal.,Larssenetal.,2006;Oharaetal.,2007].SmallerSOenhancements0.5DU)overtheShenyangregioninnortheastChina(blacksquare)aresignificantascomparedtothebackgroundregions.KROTKOVETAL.:OMISOVALIDATIONOVERNECHINA2of13 tionalOMIPBLSOdatafor4daysinApril2005withdifferentmeteorologicalandairqualityconditions.SeveralimprovementstotheoperationalOMISOdataaredis-cussedincludingnewOMIradianceandirradiancecalibra-tions(collection3)andoff-lineAMFcorrectionsbasedonaircraftmeasurementsofaltitudeprofilesofSOand2.DataCollection2.1.AircraftDataData5]TheaircraftmeasurementswereperformedaspartoftheEastAsianStudyofTroposphericAerosols:anInterna-tionalRegionalExperiment(EAST-AIRE)[C.Lietal.Z.Lietal.,2007].Eightflightswerecompletedbetween1April2005and12April2005.Only4dayswithdifferentairqualityandmeteorologicalconditions(1,5,7and10April)wereselectedforcomparisonsbecauseoftheoptimalOMIobservationalconditions(noclouds,nearnadirviewingdirections).AllflightsdepartedfromTaoxianInternationalAirport(41.640N,123.488E)intheLiaon-ingprovinceofChina.Flightpathsweresimilaron5,7and10April(Figure1).Twoprofilingflightpatterns(spirals)wereperformeddailyoverfarmland,from300mupto4,000mabovesealevel(asl)tothesouth(Liaozhong)andnorth(Xiaoming)oftheairport(Figure1).Theflighton1Aprilwasinanorth-southdirection(toHarbin)withonlyonemeasuredSOprofileonthedescenttotheTaoxianInternationalAirport.TheflightsweremadeonaChineseY-12twin-engineturbopropplane.Twoinletswerelocatedontopofthecockpitinfrontoftheengines:aforwardfacingisokineticinlettocollectaerosolsandabackwardfacinginletfortracegasmeasurements.AllinstrumentsonboardhavebeenusedonpreviousflightsovertheeasternU.S.[Taubmanetal.,2006]andarewellcharacterizedcharacterizedDickersonetal.,2007].SOwasmeasuredusingacommerciallymodifiedpulse-florescencedetectorwithadetectionlimitofabout0.2ppbvandanestimatedabsoluteaccuracyof15%(95%confidencelevel)[,1997].Thecontributionstothisuncertaintyincludesamplinglineloss,instrumentnoise,andinterferencebyotherspecies.Relativehumidityandtemperatureweremeasuredwithasolid-stateprobe(EILInstrumentsInc.,RustrakRR2–252,HuntValley,MD)andpressurewasmonitoredwithaRosemountModel2005PressureTransducer.LocationwasmonitoredwithaGlobalPositioningSystemreceiver(GarminGPS-90).Thetemperatureandpressuremeasure-ments(accurateto1%)wereusedtoconverttheSOratiotoabsoluteconcentration.Theconcentrationsmea-suredduringeachspiralwereextrapolatedtothesurfaceandverticallyintegratedbelowthemaximumaircraftalti-tude(4km)toestimateSOcolumnamountswithanoveralluncertaintyof20%forcolumncontentsabove0.3DU.Otherinstrumentsonboardhavebeenpreviouslydescribedinmoredetail[Taubmanetal.,2006;etal.,2007].Aerosolscatteringcoefficientsatthreewave-lengths(450,550,and700nm)weremeasuredwithanintegratingnephelometer(TSIModel3563)[Andersonet,1996].ThenonidealforwardscatteringtruncationwascorrectedfollowingAndersonandOgren[1998].Theaerosolopticalthickness(AOT)andAngstromcoefficientwereestimatedfromthemeasuredscatteringcoefficientsassumingaconstantsinglescatteringalbedo(SSA)ofTheAngstromcoefficientwasbiasedhighandAOTwasbiasedlowcomparedtocoincidentground-basedAERO-NETobservationsbecauseofinsufficientsamplingoflargedustparticles.Theestimatedbiaswas20%foranthropo-genicpollution(whenfineparticlesdominate),butcouldincreaseto50–100%duringduststorms(whencoarseparticlesdominate).dominate).6]Figure2showstypicalaircraftSOdataasafunctionofflighttime(localtime(LT):LT=UT+8h)on5April. Figure2.Altitude(blue)andinsituSO(red)aircraftdataasfunctionofflighttime(localtime(LT):LT=UT+8h)on5April2005overnortheasternChina(Figure1).TheamountofSOincreasesasthealtitudedecreasesduringtheflight.TheserawdatawereconvertedintothreeSOverticalprofilesoverLiaozhong(1130LT)andXiaoming(1230LT)andonthedescenttoTaoxianInternationalairport(1330LT).TheOMIoverpasstime(0457UT)correspondsto1257LT.KROTKOVETAL.:OMISOVALIDATIONOVERNECHINA3of13 Onthatday,highconcentrationsofSO10ppbv)andaerosols(scatteringcoefficients)wereob-servedwithinthePBL,below1000masl.AbovethePBL,pollutantlevelsdroppedrapidlywithaltitude,butwerestillsubstantialcomparedtobackgroundlevels,withSOcloseto1ppbvandaerosolscatteringcoefficientsclosetoabove2000masl.2.2.OMISOSO7]InthepubliclyreleasedOMISOproduct(OMSO2)fourreportedtotalSOvaluescorrespondrespectivelytotheaprioriprofileassumptionsofSOinthePBL(below2km)andinthelowertropospherefromanthropogenicsources,SOdistributedbetween5and10kmemittedbypassivevolcanicdegassinginthefreetroposphere,andSOdistributedbetween15and20kmrepresentinginjectionfromexplosivevolcaniceruptions.ThePBLdataareprocessedwiththeBandResidualDifference(BRD)algo-rithm[Krotkovetal.,2006],whileallotherdataareprocessedwiththeLinearFit(LF)algorithm[Yangetal.2007].BothalgorithmsusetheTOMS–liketotalozoneretrieval(OMTO3)[BhartiaandWellemeyer,2002]asalinearizationsteptoderiveaninitialestimateoftotalozone(assumingzeroSO)andthewavelength-independentLambertianeffectivesurfacereflectivity(LER).TheOMTO3algorithmaccomplishesthisbymatchingcalcu-latedradiancestothemeasuredradiancesatapairofwavelengths(317.5nmand331.2nmundermostcon-ditions).Theresidualsatthe10otherwavelengthsarethencalculatedasthedifferencebetweenthemeasuredandcomputedNvalues(100*isEarthradianceandissolarirradiance),wherethelatteraccountfortheeffectsofmultipleRayleighscattering,ozoneabsorp-tion,Ringeffect,andsurfacereflectivity.Inthepresenceof,theresidualscontainspectralstructuresthatcorrelatewiththeSOabsorptioncrosssections.Theresidualsalsohavecontributionsfromothererrorsources.Toreducethisinterference,amedianresidualforaslidinggroupofSOandcloud-freepixelscovering±15latitudealongtheorbittrackissubtractedforeachspectralbandandcross-trackposition[Yangetal.,2007].This‘‘slidingmedian’’empiricalcorrectionessentiallyactsasahigh-passfilterforcingmedianresidualstoequalzeroforallthebands.Doingso,anycross-trackandlatitudinalbiasesarereduced.reduced.8]OnlyoperationalOMIPBLdata,processedwiththeBRDalgorithm,willbediscussedinthispaper.ThealgorithmusesdifferentialresidualsatthethreewavelengthpairswiththelargestdifferentialSOcrosssectionsintheOMIUV2spectralregion(310–380nm)tomaximizemeasurementsensitivitytoanthropogenicemissionsinthePBL.EachpairresidualisconvertedtoSOslantcolumndensity(SCD)usingdifferentialSOcross-sectiondataatconstanttemperature(275K)[Bogumiletal.,2003].TheSCDsofthethreepairsareaveragedandtheaverageSCDisconvertedtothetotalSOverticalcolumndensity(VCD)inDobsonunits,(1DU=2.69),usingaconstantairmassfactor(AMF)of0.36:totalSO 9]ThisoperationalAMFwasestimatedforcloud-andaerosol-freeconditions,usingasolarzenithangleof30nadirviewingdirection,asurfacealbedoof0.05,asurfacepressureof1013.3hPa,a325DUmidlatitudeozoneprofileprofileMcPetersetal.,2007]andatypicalsummerSOprofileovertheeasternU.S.[Taubmanetal.,2006].2.2.1.OMISOSO10]AnestimateoftheOMISOprecision(noise)anddetectionlimitcanbeobtainedbyexaminingtheretrievalstatisticsoverpristineoceanicregions,whereSO Table1a.OMIInstantaneousFieldofView(IFOV)SONoiseStatisticsOverBackgroundRegionsintheSouthPacificon6April2005(OMICollection2OperationalData,Orbit3858)SampleAreaNumberofIFOVDataAreaMeanSOColumnContent,DUStandardDeviationofIFOVData,DU41–43S;130–13240–44S;128–134Wholeorbitfrom70Sto30 Table1b.StatisticsfortheArea-AveragedSONoiseintheSouthPacificRegion(41–43S;130–132E)on6April2005(OMICollection2OperationalData,Orbit3858)AveragingRadius,kmAverageNumberofIFOV,nAverageoftheAreaMeans(DU)StandardDeviationoftheAreaMeans(SDM),n StdDevIFOVKROTKOVETAL.:OMISOVALIDATIONOVERNECHINA4of13 areindistinguishablefromzero(DU)[Seinfeldand,1998;Chinetal.,2000].BecauseofmeasurementandretrievalerrorstheOMIInstantaneousFieldofView(IFOV)SOdatahaveastatisticaldistributioninsuchbackgroundregions.Therandomnoiseplusspatialvari-abilityandtemporalvariabilityshouldexplainallthevarianceintheIFOVdata.IftheretrievedOMISOisduetorandomfluctuationssuchasdetectornoisethenthevariabilityshoulddecreaseasthetimeforwhichthesignalisaveraged(andtheareaoverwhichthesignalisaveraged)increases.TheuncertaintyintheaveragecolumncontentshoulddecreasewiththesquarerootofthenumberofIFOVobservationsaveraged(n).Forexample,consideraregionintheSouthPacificOceanboundedby40–44Sand128–Eon6April2005(Table1a).ThisregionwaschosenbecauseoftheabsenceofSOemissions,andsimilarobservationalconditionstotheEAST-AIREregion(sameorbit,sameviewinggeometry,absenceofclouds).Table1ashowsthatboththemeanandstandarddeviation(1.5DU)oftheIFOVdatadonotchangewiththesizeoftheaveragingarea,sotheregionisspatiallyhomogeneous.AssumingGaussiannoisestatistics,eachIFOVSOalisexpectedtobewithin+/3DU(2)in95%ofcases.ThisnoiselevelistoohightodetectbackgroundSOamountsormostanthropogenicSOSOTaubmanetal.2006].Therefore,spatialsmoothingisrecommendedtoreducethenoise(i.e.,averagingallIFOVdatawithincertainradiusforagivenlocation).Thestandarddeviationofthemeansoftheareas(SDM)canbetakenasanestimateofthesmoothednoise.Table1bshowsthesmoothednoisestatis-ticsoverthesameareaon6April.AsthenumberofaveragedIFOVdata(n)increaseswiththeaveragingradius,theSDMdecreases,butmoreslowlythanForexample,theOMInoisein30kmaverages(n=isestimatedtobe0.6DU,whiletaking100kmaverages(n=90)wouldreducethenoiseto0.3DU.Takingintoaccountlargeday-to-dayvariabilityinareameanvaluesevenoverpristinelocations(Table1c),wedoubletheOMInoiseestimatetoaccountforthetemporalnoisecomponent.TheOMIdetectionlimitistherefore3DUforIFOVdata(±twoStdDev(IFOV)),butreducesto1.2DU(±fourSDM(n),n=90,Table1b)fordaily100kmaverages.Time-averagingfurtherreducesthenoise,enablingdetec-tionofweakerstationarySOemissionsnotobviousindailydata(Figure1).Thelong-term(1year)meanovertheS.Pacificregioniszerowithastandarddeviationof Table1c.TemporalStatisticsoftheDailyAreaMeanOMIPBLSOValuesFrom2005OveraBackgroundRegionintheSouthPacificandOvertheEAST-AIREFlightRegioninNEChinaofDaysMean,DUDeviation,DUPercentOutliersOutside±2StDevSouthPacific41–43S,130–132NEChina41–43N,122–124110Onlydayswithatleast70cloud-freeIFOVretrievalswereused.Thelong-termaverageSOcolumncontentof0.65DUisequivalenttoabout3ppbinthelowest2kmoftheatmosphereoverNEChina. Figure3.(left)SOverticalprofiles:measuredduringEAST-AIREcampaignon5April2005overLiaozhong,NEChina(bluestars);mediansummerSOprofileovereasternU.S.[Taubmanetal.,2006](reddiamonds).BothSOprofilesaredimensionless,normalizedtoaunitcolumnSOamount.BlacklineshowsverticallyresolvedOMISOsensitivityorlocalAMF:)calculatedfornadirviewingdirection(),solarzenithangle=46,surfacereflectivity=0.05,=325DUmiddlelatitudeTOMSclimatologicalozoneprofile,noaerosolsandclouds.(right)LocalAMFdependenceonsurfacealbedo,(solidline,=0.05;dashedline,=0.1)andtotalozone,(redindicatesTOMS425DUmidlatitudeozoneprofileandblackindicates=325DUmidlatitudeozoneprofile).KROTKOVETAL.:OMISOVALIDATIONOVERNECHINA5of13 0.6DU,whileoverNEChinathelong-termmeanvalueis0.65DUwithastandarddeviation1.1DU(Table1c).TheseresultsillustratepersistentSOemissionsoverNEChina(Figure1).2.2.2.OMISOSO11]TheSOretrievalaccuracydependsontheuncer-taintyinbothSCD,,andinaveragephotonpath,characterizedbytheerrorinassumedAMF:(1).Thecombinederror,,canbeexpressedusingequation(1)andassumingthatarenotcorrelated: eSCSCD2\t isthetotalSO=0.36(1),andisestimatedfromthebackgroundnoisestatistics(0.2DUorfora2by2errorhasadditionalsystematiccomponentsduetouncertaintiesinlaboratorymeasuredSOcrosssectionssectionsBogumiletal.,2003]andthealgorithmicassumptionofaconstantPBLtemperature.AssumingthattheeffectivePBLtemperaturerangesfrom260Kto300K,theoperationalOMIPBLSOvalues,derivedbyassumingaconstanteffectivetemperatureof275K,areoverestimatedby4%orunderestimatedby8%atthesetwoextremes.Thiserrorshouldbecorrectedoff-lineifSOandtemperatureprofilesareknownfromancillarymeasurementsormodels(seesection3.2).TheAMFerror,,issystematicandincreaseswithdeviationoftheobservationalconditionsfromthoseassumedintheoperationalalgorithm.ForaquantitativeuseofPBLSOdata(i.e.,emissionsestimates),theoperationalAMFshouldbecorrectedtoaccountforactualobservationalconditions.GiventherelativelyhighnoiseinSCDvalues(30–50%forEAST-AIREregionalaverages(Table2)andupto200%fortheindividualpixels(Table3)),theAMFcorrectiondoesnotneedtobeexact.Here,weproposeasimpleparameterizationthatcanbeusedtoestimatetheAMFwith20%uncertaintyusingonlyOMImeasurementsoftotalozone.2.2.3.AMFParameterizationon12]TheAMFdependsonSOverticalprofile,surfacealbedo(),observationalgeometry(viewing(),solarzenith()andrelativesolarazimuth()angles),totalcolumnozone(),aerosolsandclouds: )istheverticallyresolvedOMIsensitivity(i.e.,localAMF),istheOMImeasuredreflectanceatthetopofatmosphere,)istheSOabsorptionopticalthicknessinaverticallayerbetweenz[km]andz+dz[km],and)isthedimensionlessnormalizedSOverticalprofileshape(Figure3).Sinceisweaklydependentonwavelength,weconsiderspectrallyaveragedvalueoveranarrowspectralwindow(i.e.,311–315nm)usedbytheOMIBRDalgorithm[Krotkovetal.2006].Figure3comparesthesummermedianSOmeasuredovertheeasternU.S.[Taubmanetal.,2006]with Figure4.AMFparameterizationasfunctionofslantcolumnozoneamount(SCO):(sec(sec())inDobsonunits.TheparameterizationwascalculatedusingthemediansummerSOmeasuredbyaircraftovertheeasternU.S.[Taubmanetal.,2006],TOMSclimatologicalmiddlelatitudeozoneprofiles,differentviewingangles(0–60),solarzenithangles(0–60),andrelativesolarazimuthangles(60–120)andasurfacealbedoof0.05withnoaerosolsorclouds.Dashedlineisthelinearregression(6),where=0.486and=0.000177[DU]KROTKOVETAL.:OMISOVALIDATIONOVERNECHINA6of13 aprofilemeasuredoverNEChinaduringtheEAST-AIREcampaigninApril2005.BothSOprofilesarenormalizedtoaunitcolumnSOamount.Thus,theAMFcanbeinterpretedasaprofile-weightedmeanvalueoftheverticallyresolvedOMISOsensitivity,)(4).AlthoughthetypicalSOprofileshapesarequitedifferentfortheeasternU.S.andChina,itturnsoutthattheprofile-integratedAMFiscloseto0.4inbothcases.Therefore,AMFcorrectionsduetothemeasuredSOprofileshapeforEAST-AIREconditionsareexpectedtobesmall.Increasingthesurfacealbedofrom0.05to0.1increasestheAMFby35%(Figure3,right).However,thelandalbedoistypicallysmallatUVBwavelengths(310–315nm),sousingthedefaultvalue(=0.05)willresultinlessthana15%AMFerror.Alargeincreaseintotalozonefrom325DUto425DUcausesonlya10%decreaseintheAMF,butthissystematiceffectcanbeeasilytakenintoaccountusingtheOMItotalozonemeasurements. AMFcorrected13]TheAMFcorrectionduetoobservationalgeometry)canbecombinedwiththetotalozonecorrectionusingasimplelinearregressionwithrespecttotheslantcolumnozone(SCO)(Figure4):AMFcorrected\tisthetotalcolumnozonemeasuredbyOMI.Theregressioncoefficients,dependontheshapeoftheverticalprofile,thesurfacealbedo,azimuthangleandthepresenceofaerosolsandclouds.Figure4comparesparameterizedAMFvaluescalculatedusingtheregression(5)–(7)withvaluescalculatedwiththeforwardradiativetransfermodel[,1964]fordifferentozoneprofilesandobservationalgeometries.TheoperationalAMFof0.36isunderestimatedby20%forSCO700DU(300DU,smallsolarzenithandviewingangles),butoverestimatedby30%forSCO=1500DU.ForlargerSCOvalues(highozoneand/orhighsolarzenithandviewingangles,mostlyathighlatitudes),theAMFbecomesverysmall,sovalid Figure5.AuraOMIPBLSO2maps(linearcolorscalefrom0to6DU)superimposedonAquaMODISRGBimages(imagescourtesyofMODISRapidResponseProjectatNASA/GSFC)acquired15minpriortotheOMIoverpassduringtheEAST-AIREfieldcampaignon(a)1April,(b)5April(pollutedairseenoverNEChinaaheadoftheacoldfront),(c)7April(pollutedairpushedoffthecoast),and(d)10April(behindcoldfront)2005.TheaircraftspirallocationsaredenotedbywhiteaircraftsymbolsinFigure5b.ApparentlyhighSOvalues(redcolors)indicateSOloftingabovethePBLand/orabovelow-levelclouds,whereOMIsensitivityisenhanced.Troposphericwindsareindicatedbyyellow(surface)andwhite(850mbar)arrows.BackgroundmapsarefromGoogleEarthscreencaptures(GoogleEarthimageryGoogleInc.,usedwithpermission)containingimageryfromTerraMetrics,Inc.(Copyright2008TerraMetrics,Inc.,http://www.truearth.com).KROTKOVETAL.:OMISOVALIDATIONOVERNECHINA7of13 PBLSOretrievalsarenotexpected.Theparameterization(6)doesnotaccountforAMFdependenceonrelativesatellite-solarazimuthalangle,.BecauseoftheOMIpolarsun-synchronousafternoonorbit(1345localequatorcross-ingtime),rangesbetween0and30forcross-trackpositions1–30,but150–180forpositions31–60inthetropics.Inthemiddletohighlatitudesapproaches60–.TheAMFtypicallydecreasesasincreasesfrom0 Table2.OMIandAircraftSOColumnDensityAveragedOvertheEAST-AIREFlightRegioninApril2005IFOV,nOMICollection2DataOMICollection3:SO,DU(SD)AircraftColumnOzone,DU(SD)LER(SD)AI(SD)SO,DU(SD)1Apr74399(4)0.06(0.01)1.8(0.2)2.1±0.7(1.4)1.4±0.75Apr113367(4)0.08(0.01)2.1(0.2)2.9±0.6(1.2)1.9±0.6(1.2)7Apr98342(3)0.05(0.02)2.8(0.6)0.9±0.6(1.4)0.9±0.6(1.3)10Apr90382(5)0.09(0.05)0.9(0.3)0.45±0.6(1.2)0.5±0.6(1.1)AlsoshownareOMIregionalaverageozone(DU),aerosolindex(AI)andreflectivity(LER)fromtheOMTO3algorithm.Thevariabilityovertheflightregion(onestandarddeviation)isshowninparentheses.ThehighAIrecordedon7April2005suggeststhatmineraldustataltitudesabovethoseflownbytheaircraftmaycreateapositiveinterference.Verticallyintegrated(fromthesurfaceto4km)insituSOprofileaveragedoverallspirals.StandarddeviationoftheareameanSOretrievalswithcomparableIFOVsamplesize(n)overabackgroundareaintheSouthPacific(Table1b)multiplyingbyfactorof2toaccountforday-to-dayvariability(Table1c)andnotincludingAMFerror.Collection3noisestatisticsarethesameascollection2. Figure6.Operational(collection2)OMIPBLSOdataovertheEAST-AIREflightregioninApril2005.BackgroundmapsarefromGoogleEarthscreencaptures(GoogleEarthimageryGoogleInc.,usedwithpermission)containingimageryfromTerraMetrics,Inc.(Copyright2008TerraMetrics,Inc.,http://www.truearth.com).MODISimagecourtesyofMODISRapidResponseProjectatNASA/GSFC.KROTKOVETAL.:OMISOVALIDATIONOVERNECHINA8of13 to180,soinextremecasesa20%errorispossible.However,averagingregressioncoefficientsforintherange60–120reducestheerrorto+/10%.Therefore,undercloud-andaerosol-freeconditionstheremainingAMFerrorshouldbewithin20%,lessthanthenoiseinSCDvalues.However,aerosolsandcloudscanaffecttheerrordifferentlydependingonmanyfactors.Theassump-tionisthatcloudsscreenPBLSO,butnoAMFcorrectionisattemptedtoaccountforthis‘‘hidden’’SO.Toavoidthisuncertainty,onlycloud-freedaysareconsideredforvalidation,sothatOMIarea-weightedreflectivity(LER)islessthan0.1.TheaerosolAMFcorrectionsarefurtherdiscussedinsection4.3.OMISOValidationsalidations14]Figure5showsoperationalOMIPBLSOcombinedwithhigh-resolutionMODISRGBcompositesonthe4daysinApril2005selectedforvalidation:1,5,7and10April.TheMODISinstrumentisaboardtheEOSAquapolarorbitingsatellite[Kingetal.,2003],whichorbitstheEarth15minaheadofAuraalongthesameorbit.Accordingtohigh-resolutionMODISimagery,theflightregion(i.e.,41–43Nand122–124E)wascloud-freeonalldaysasconfirmedbylowaverageOMIreflectivity(LER0.1;Table2).However,airqualitywasdramaticallydifferentonthesedaysbecauseofdifferentmeteorologicalregimes.Between5and7Apriland9and10AprilcoldfrontstraveledfromcontinentalChinaoverKoreatotheSeaofJapan.PollutedSO-richairwas‘‘pushed’’aheadofthecoldfrontsandloftedabovethePBL.Mostlycloud-freeand-freeairwasbehindthecoldfronts.Forexample,thelocationofthecoldfronton7AprilcouldbeclearlyidentifiedinbothMODISimageryandOMISO(Figure5c).ApparentlyhighOMIoperationalSO&#x-470;(5DU)overKoreaandSEChinaon1and7AprilprovideevidenceofpollutedairloftingabovethePBL,whereOMIsensitivityincreases(Figure3).SincethePBLOMIdataarenotcorrectedfortheSOloftingeffect,theoperationalvaluesareoverestimates.Whenanelevatedplumetravelsabovelow-levelmeteorologicalclouds,theOMIsensitivityisenhancedbecauseofcloudreflection(Figure3)andadditionaloverestimationresults.Therefore,forelevatedplumes,off-linecorrectionisneededtoaccountforbothSOplumeheightandunderlyingcloudreflectivity.EstimatingtheSOamountforelevatedplumesinthelowerfreetroposphere(below5km)needsspecialconsiderationandisthesubjectoffuturestudies.Inthispaper,weonlyconsiderOMIdataovertheEAST-AIREflightregion(41–43Nand122–124E)whereSOwaslocatedpredominantlyinthePBL,asconfirmedbyaircraftinsituprofilemeasurements(Figure2).3.1.AreaAverageComparisonsComparisons15]Figure6showssmoothedoperationalOMISOovertheEAST-AIREflightregiononthecomparisondays.Qualitatively,OMImeasurementsofSOagreewiththeaircraftinsituobservationsofhighconcentrationsofSO(approximately1–2DU)aheadofthecoldfront(1and5April)andlowerconcentrationsbehindit(7and10April).ThiscomparisondemonstratesthatOMIcandistinguishbetweenbackgroundSOconditionsandheavypollutiononadailybasis.Quantitatively,therearedefinitedifferencesbetweentheaircraftandOMImeasurementsaveragedover Table3.ComparisonsBetweenAircraftSpiralsandOMIIFOVSOMeasurementsAveragedWithin30kmofEachSpiralDay2005SpiralLocationandStartTime,UTCollection3SO,DUCollection2,DUAMFCorrected,DUnO,DUAILER1AprXiaoming,N/AN/AnodataN/A0.7±1.41.963951.60.071AprTaoxian07230.51.3±0.22.1±1.61.4±1.62.744031.70.061AprLiaozhong,N/AN/AnodataN/A1.1±1.41.564042.10.075AprXiaoming04291.191.3±0.21.6±1.31.5±1.32.673672.00.075AprTaoxian05170.641.5±0.32.7±1.22.8±1.23.893712.10.075AprLiaozhong03121.112.3±0.42.5±1.22.4±1.22.783672.40.097AprXiaoming07350.560.0±0.1N/A0.8±1.21.183423.60.047AprTaoxian09200.430.22±0.11.2±1.31.5±1.31.673422.50.047AprLiaozhong08570.410.0±0.1N/A1.5±1.41.363414.30.0210AprXiaoming02570.230.07±0.10.5±1.40.6±1.40.763851.00.1210AprTaoxian04530.170.21±0.10.9±1.30.7±1.30.573810.70.1410AprLiaozhong04280.150.04±0.10.6±1.30.6±1.30.573760.80.10nisthenumberofIFOVdataaveraged.Spirallocationswere42.450N,123.70E(Xiaoming,138masl)and41.35N,122.648E(Liaozhong,15masl).AllflightsdepartedfromTaoxianInternationalAirport(41.640N,123.488E,58masl)intheShenyangregion.AircraftinsitumeasuredSOconcentrationswereextrapolatedtothesurfaceandverticallyintegrateduptothemaximalaircraftaltitude4kmtoestimateSOcolumndensitieswithanuncertaintyofOMIIFOVdatawereaveragedwithin30kmofeachspirallocation.OMIoperationalcollection2data:ozone,aerosolindex(AI)andreflectivitycollection2datafromtheOMTO3algorithm[BhartiaandWellemeyerAerosolopticalthickness(AOT)at500nmwasestimatedfromaircraftinsitumeasurementsofaerosolscatteringcoefficientsat450nm,550nmand650nmandintegratedfromthesurfaceupto4kmassumingaSSAof0.9at500nm.Collection3SOdatawithtemperatureandAMFcorrectionsapplied.ReprocessedOMIdatausingtheoperationalSOalgorithmwithrecalibratedradianceandirradiance(collection3level1b)data[Dobberetal.,2008].On1Aprilaircraftmeasurementsweretakingduringaquickdescenttotheairport,sothataerosolAOTdataarelessreliable(nephelometer’saveragingtimeis5min).TheAOTwasobtainedfromaground-basedhandheldSunphotometer.TheobservationtimeoftheSunphotometerisabout0230–0500UTCandtheOMIoverpasswasat0522UTC.TheaircraftinsitumeasurementsoverXiaomingandLiaozhongon7AprilsuggestalmostnoSOthroughoutthewholecolumncoveredbytheaircraft.Theaircraftaerosoldatawereaffectedbyflyingthroughadeckoffairweathercumulusclouds.Ground-basedAERONETaerosolopticalthicknessdatafromneartheLiaozhongsite(41Nand122E)on10Aprilwereused.KROTKOVETAL.:OMISOVALIDATIONOVERNECHINA9of13 theflightregiononbothcleanandpolluteddaysasshowninTable2.Oncleandays7and10ApriltheaircraftverticallyintegratedinsituSOmeasurements(0.1DU)werebelowtheOMIdetectionlimit(1DU).Thedis-agreementbetweenaircraftandsatelliteobservationscouldoriginatefromalayerofSOintheuppertroposphereorlowerstratosphere.ThereisnowaytoruleouthighconcentrationsofSOabovethemaximumflightaltitude(4000m),althoughthisisunlikely.Thedifferencebetweenthesedaysisthatthe5–6AprilcoldfrontbroughtwithitlargeamountsofdesertdustasconfirmedbyahighOMIaerosolindex(AI)on7April(AI=3;Table2),whiletheAIvaluewaslowon10April,suggestingnodust.DustaerosolstypicallyhaveastrongabsorptionenhancementintheshorterUVBwavelengths[d’Almeidaetal.,1991;SokolikandToon,1999;Alfaroetal.,2004]andcouldinterferewithOMISOretrievals.Thedustinterferenceisasubjectoffuturestudy..16]Asopposedtocleandays,1and5Aprilrepresentpollutedairmassesaheadofcoldfronts(Figure5).Trajec-toryanalysis[DraxlerandRolph,2003]suggeststhattheairon1Aprilwasfromthenorthwestat500,1000,and2000mabovegroundlevel(Figure5a).ThelocalmeteorologicalrecordsatTaoxianairportindicateweakwindswithvariabledirectionsonthatday.Thetrajectorieson5Aprilaremainlyfromthesouthwest.BeforereachingtheShenyangarea,theairpassedovermanyemissionsourcesinmajorindustrialregions,includingtheBeijing[Anetal.,2007]andShang-haiareas.Aircraftmeasurementson5AprilshowhighSOconcentrationsthroughoutthePBL(upto19ppbv)andinthelowerfreetroposphere(Figure2).Thedailyaircraftspiralaverageswere1.3DUon1April(onlyoneprofilemeasured)and1.7DUon5April(averageof3spirals).TheoperationalareaaverageOMISOvalues(2DUon1Apriland3DUon5April)aresubstantiallyhigher(uptoafactorof2)thanaverageaircraftmeasurements(Table2).2).17]OneofthereasonsfortheOMISOnoiseandbiasisalgorithmicassumptions[Yangetal.,2007].TheotherreasonisimperfectcalibrationappliedtotheOMImeasuredradianceandirradiancedataavailableinthecollection2level1bdata.Recentlylevel1bdatahavebeenreprocessed(collection3data)withseveralimprovementsappliedtothecalibration:(1)dailydarkcurrentmapssubtractedfromthemeasuredradianceandirradiancedataand(2)improvedradiometriccalibrationincombinationwithamodifiedstraylightcorrection[Dobberetal.,2008].Preliminaryanalysishasshownthatimprovedtreatmentofthedarkcurrenthasreduced‘‘striping’’inresiduals.TheimprovedstraylightcorrectionisalsoexpectedtohaveapositiveeffectbyreducingthebiasinoperationalSOdata.Todemonstratethis,wererantheoperationalOMISOalgorithmusingcollection3OMTO3residualswithoutmakinganychangestothealgorithmandsoftware.TheseresultsareshowninTable2ascollection3OMISOdata.Ascanbeseen,thenoiseandpositivebiasoncleardaysremainthesame.However,collection3SOdataarelowerandinbetteragreementwithaircraftmeasurements(within15%)onbothpolluteddays.3.2.IndividualCaseComparisonssons18]Comparisonswithaircraftmeasurementsfortheindividualprofilingflightpatterns(spirals)aresummarizedinTable3.Toreducethenoise,theOMIIFOVdatawereaveragedwithin30kmofeachspiral(4–9)andcomparedwithverticallyintegrated(upto4km)insituaircraftdata.Forsuchsmallsamples,theOMInoiseisestimatedtobe1.2–1.6DU(doubleSDM(n)inTable1b)dependingonthenumberofIFOVaveraged(n),notincludingapossibleAMFerror.OMI(collection2)SOretrievalsarelargerthantheaircraftmeasurementsinallcases.Thelargestdifferenceof2.3DU(150%)isfoundoverTaoxianInternationalAirporton5Aprilandthesmallestdifference(0.4DUor17%)overLiaozhongonthesameday(Table3).InotherpollutedcasesOMIvaluesaredoubledcomparedtotheaircraftmeasurements.Onbothcleandays(7and10April)OMIvaluesarewithinthenoiselevel.ReprocessingOMISOdatawiththeoperationalalgorithm,butbettercalibratedlevel1bdata(collection3datainTable3)resultedinimprovedagreementwithaircraftmeasurementsforallretrievalsonpolluteddays.ThemaximaldifferenceoverTaoxianInternationalAirporton5Aprilisreducedto1.3DU(90%),whichiscomparabletothenoise.Thedifferencesaresmallerinotherpollutedpolluted19]UsingaircraftSOandaerosolprofiledataonecanpartitionthetotalOMIerrorbetweenAMFerrorandretrievalerrornotrelatedtotheoperationalAMFassump-tions(equation(2)).Toquantifythelatter,theoperationalAMFwascorrectedtoaccountforknownsourcesofsystematicerrors.First,theSOcrosssections[etal.,2003]werecorrectedusingaircraft-measuredSOandtemperatureprofiles.Forexample,theSO-weightedtemperatureoverLiaozhongon5Aprilwascloseto289K,higherthanthePBLtemperatureof275KassumedintheoperationalOMIalgorithm.Inthiscasetheassumeddiffer-entialSOcrosssectionsareoverestimatedby4.5%andoperationalSOvaluesareunderestimatedbythesamepercentage.Therefore,thetemperaturecorrectionresultedinanincreaseintheoperationalSOvalues.Next,theoperationalAMFwascorrectedtoaccountfortheaircraft-measuredSOandaerosolprofiles(Figure7),totalozone,andOMIviewinggeometry.Thecorrectionwasdoneusingalinearregression(equations(5)–(7)),whereregressioncoefficientswererecalculatedforeachspiralusingtheactualaircraftmeasuredaerosolopticalpropertiesandaerosolandSOverticalprofiles.Theaerosolscatteringcoefficientsweremeasuredinsituat3visiblewavelengths:450nm,550nm,and650nmandtheAOTat500nmwasestimatedassumingaconstantvalueofSSAat500nm(0.9).ThespectraldependenceoftheAOTandSSAintheUVwascalculatedusingMiecode(sphericalparticleassumption)withrefractiveindexandsizedistributionsfromOMIdustandindustrialaerosolmodels[Torreset,2007].Theseparameters,togetherwiththemeasuredandaerosolverticalprofilesandTOMSclimatologicalozoneverticalprofiles,wereinputtotheradiativetransfercodetocalculatetheAMFregressionasafunctionofSCO,foreachspirallocation(7).TheresultingAMFfortheLiaozhongspiralon5AprilareshowninFigure7for3scenarios:noaerosols,industrialaerosols,anddustaerosols.TheindustrialaerosolswithaflatSSAspectraldependence(mixturesofsulfateandblackcarbon)havelittleeffectontheAMF,whiledustaerosolswithenhancedUVabsorption(dustandsecondaryorganicaerosols(SOA))wouldreduceKROTKOVETAL.:OMISOVALIDATIONOVERNECHINA10of13 theAMFbyhalf(Figure7,left).Wenotethatthesetwotypesofaerosolshavethesameopticalpropertiesinthevisiblewavelengths(AOT1andSSA0.9at450nm),whereaerosolmeasurementsaretypicallyavailable,butquitedifferentSSAattheUVBwavelengths(310–315nm)whereOMISOretrievalsaredone.Assumingonlyasoot/sulfatemixturewouldresultinanOMIretrievedvalueof2DUSOinthisparticularcase.However,assumingadustaerosolregimewiththesamepropertiesinthevisiblewouldresultintwiceasmuchSO4–5DU)..20]AMFcorrectionsassuminganindustrialaerosolmodelaretypicallysmall()forallpollutedcases,exceptforonecaseoverTaoxianairporton1April(Table3).Inthiscasethecombinationofhightotalozone(400DU)andalargeOMIviewingangle(40)resultsinalargedownwardAMFcorrection,sothatthecorrectedSOisincreasedby50%from1.4DUto2.1DU.TheaircraftmeasuredSOwascloseto1.3DUinthiscase,sothattheAMFadjustmentusinganindustrialaerosolmodelincreasesthedifference.UsingadustaerosolmodelwouldresultinanevensmallerAMF,whichwouldfurtherincreasethedifference.WeconcludethattheOMIcollec-tion3SCDisstillhighinthiscase,whichisaccidentallycompensatedbytheoverestimatedvalueoftheoperationalAMF(0.36;equation(1)).Therefore,makingoff-lineAMFadjustmentsisimportantforestimatingunbiasederrorsintheOMISCDretrievals.Overall,AMF-correctedcollection3OMIdataarehigherthanaircraftmeasurements.Inpollutedcasesthebiasrangesfrom0.2DU(10%)to1.2DU(80%)andoncleardaysfrom0.4DUto1DU.Theaveragepositivebiasiscloseto+0.6DU.ThisbiasisinsignificantcomparedtothenoiseintheOMIdata(1.3DU).4.ConclusionsConclusions21]InthisfirstOMISOvalidationstudy,conductedoverShenyanginNEChinaaspartoftheEAST-AIREfieldcampaigninApril2005,SOobservationsfrominstrumentedaircraftflightsarecomparedwithOMIoper-ational(collection2)andreprocessed(collection3)OMIdata.Comparisonsaremadeforclearandpolluteddaysunderfavorableobservationalconditions(noclouds,nearnadirviewingdirections).Thefollowingconclusionscanbebe22]1.OperationalOMIPBLSOmeasurementsquali-tativelyagreewiththeaircraftinsituobservationsofhighcolumnamounts(approximately1–2.DU)aheadofacoldfront(1and5April)andlowerconcentrationsbehindit(7and10April).ThiscomparisondemonstratesthatOMIcandistinguishbetweenbackgroundSOconditionsandheavypollutiononadailybasis.ThissupportsthepotentialapplicationoftheOMIPBLSOproducttoregionalairpollutionstudies.studies.23]2.ToevaluatetheminimumdetectableamountofOMIPBLSOunderoptimalobservationalconditions(noclouds,solarzenithangleandnearnadirviewingangles)weexaminedthesignalanditsvariabilityoveraremoteregionoftheSouthPacificwherethecolumncontentshouldbeconsistentlywellbelow0.1DU.Forindividualinstantaneousfieldofview(IFOV)data,thestandarddeviation()is1.5DU.Thenoisecanbereducedbyaveragingoveralargerareaorforalongertime.Forasingleday,fallsto0.82DUforanaveragingareawitharadiusof20km(4IFOVdata)andto0.36DUforradiusof70km(44IFOVdata),butincreasingtheareafurtherhaslittleimpactonthenoise;fora100kmradiuswith89IFOVdata=0.28.Averagingover75days(foreachofwhichatleast70cloud-freeIFOVpointswereavailable)forthewhole2areagaveastandarddeviationof0.6DU.Thevariabilityinthemeasurementisgreaterthanexpectedfrompurelyrandomerrorssuchasinstrumentnoisedue,perhaps,tosystematicerrorssuchasvariabilityinsurfaceconditionsorstratosphericozone. Figure7.(left)SOAMFparameterizationasfunctionofSCOfora275DUozoneprofile,differentviewingandsolarzenithangles,solarazimuth60?120and0.05surfacealbedo.Bluediamonds,redstars,andpurplecrossesindicatetheaerosolmodelusedintheparameterization,respectively:noaerosolsorclouds,anOMIindustrialaerosolmodel(sulfate/sootmixture),andadustaerosolmodelwiththesameopticalparametersinthevisiblewavelengths.(right)Normalizedverticalprofileoftheaerosolscatteringcoefficientfromaircraftinsitumeasurements(bluelines)andnormalizedSOconcentrationprofilefromaircraftinsitumeasurementson5April(redlines).Symbolsindicatespirallocation:Xiaoming(triangle),Taoxian(diamond),andLiaozhong(square).KROTKOVETAL.:OMISOVALIDATIONOVERNECHINA11of13 24]3.Apositivebiasofupto150%isdemonstratedintheoperationalcollection2OMIdata(OMIbeinghigher)onpolluteddays.ReprocessedOMISOdatawithbettercalibratedradiance/irradiancedata(collection3level1bdata)reducesthedifferencewithaircraftmeasurementsonpolluteddays,buthaslittleeffectoncleandays.days.25]4.TheoperationalSOdatawerecorrectedoff-linetoaccountfortheAMFdependenceontotalozone,SOprofile,viewinggeometry,andaerosoleffects.Overall,AMF-correctedcollection3OMIdataarehigherthanaircraftmeasurements.Inpollutedcasestheoverestimationrangesfrom0.2DU(10%)to1.2DU(80%)andoncleardaysfrom0.4DUto1DU.ThecampaignaverageOMIbiasiscloseto+0.6DU.DU.26]5.AbsorbingaerosolsinterferewithOMISOretrievalsbyaffectingtheAMF.TheindustrialaerosolshavelittleeffectontheAMF,whiledustaerosols(largeparticles,withstrongabsorptioninUV)reducetheAMFbyhalfdoublingtheretrievedSO.ThereforequantifyingthespectraldependenceofaerosolabsorptionatSOlengths(310–330nm)iscriticalforaccurateretrievalofamountsusingsatelliteUVmeasurements.nts.27]Acknowledgments.TheDutch-FinnishbuiltOMIinstrumentispartoftheNASAEOSAurasatellitepayload.TheOMIprojectismanagedbyNIVRandKNMIintheNetherlands.TheauthorswouldliketothanktheKNMIOMIteamforproducingL1BradiancedataandtheU.S.OMIoperationalteamforcontinuingsupport.NickolayKrotkov,RussellDick-erson,ArlinKrueger,SimonCarnandBrittanyMcClureacknowledgeNASAfundingofOMISOresearchandvalidation(grantsNNG06GI00GandNNG06GJ02G);ArlinKrueger,SimonCarn,andKaiYangacknowl-edgeNASACAN(NNS06AA05G)funding.ZhanqingLiacknowledgessupportfromtheNASARadiationScienceProgram(NNG04GE79G)andsupportfortheEAST-AIREprojectfromUMDCMPS.TheaircraftcampaignwasalsosupportedbyNSF.TheauthorsaregratefultoJenniferHainsforhelpwithprocessingaircraftdataandtoColinSeftorforproducingOMI/MODIScompositefigures.TheauthorsaregratefultoJianYuan,FujiuGong,andX.BianfromLiaoningProvincialMeteorologicalBureau,ChinaforhelpduringtheEAST-AIREfieldcampaign.ReferencesAlfaro,S.C.,S.Lafton,J.L.Rajot,P.Formenti,A.Gaudichet,andM.Maille(2004),Ironoxidesandlightabsorptionbypuredesertdust:Anexperimentalstudy,J.Geophys.Res.109,D08208,doi:10.1029/2003JD004374.An,X.,T.Zhu,Z.Wang,C.Li,andY.Wang(2007),AmodelinganalysisofaheavyairpollutionepisodeoccurredinBeijing,Atmos.Chem.Phys.,3103–3114.Anderson,T.L.,andJ.A.Ogren(1998),DeterminingaerosolradiativepropertiesusingtheTSI3563IntegratingNephelometer,AerosolSci.Technol.,57–69,doi:10.1080/02786829808965551.Anderson,T.L.,etal.(1996),Performancecharacteristicsofahigh-sensitivity,three-wavelength,totalscatter/backscatternephelometer,J.Atmos.OceanicTechnol.,967–986,doi:10.1175/1520-0426(1996)013967:PCOAİHS2.0.CO;2.Benkovitz,C.M.,S.E.Schwartz,M.P.Jensen,M.A.Miller,R.C.Easter,andT.S.Bates(2004),ModelingatmosphericsulfurovertheNorthernHemisphereduringtheAerosolCharacterizationExperiment2experimen-talperiod,J.Geophys.Res.,D22207,doi:10.1029/2004JD004939.Bhartia,P.K.,andC.W.Wellemeyer(2002),OMITOMS-V8totalO3algorithm,AlgorithmTheoreticalBasisDocument,OMIOzoneProductsATBDversion2.0,vol.II,editedbyP.K.Bhartia,NASAGod-dardSpaceFlightCent.,Greenbelt,Md.(Availableathttp://eospso.gsfc.nasa.gov/eos_homepage/for_scientists/atbd/docs/OMI/ATBD-OMI-Bogumil,K.,etal.(2003),MeasurementsofmolecularabsorptionspectrawiththeSCIAMACHYpre-flightmodel:Instrumentcharacterizationandreferencedataforatmosphericremote-sensinginthe230–2380nmre-J.Photochem.Photobiol.Chem.,167–184,doi:10.1016/S1010-6030(03)00062-5.Bovensmann,H.,J.P.Burrows,M.Buchwitz,J.Frerick,S.Noe¨l,V.V.Rozanov,K.V.Chance,andA.P.H.Goede(1999),SCIAMACHY:Missionobjectivesandmeasurementmodes,J.Atmos.Sci.(2),127–150,doi:10.1175/1520-0469(1999)056127:SMOAMM�-60;2.0.Bramstedt,K.,A.Richter,M.VanRoozendael,andI.DeSmedt(2004),ComparisonsofSCIAMACHYsulfurdioxideobservations,inProceed-ingsoftheSecondWorkshopontheAtmosphericValidationofEnvisat2),3–7May2004,Frascati[CD-ROM],Eur.SpaceAgencySpec.Publ.ESASPBurrows,J.P.,etal.(1999),TheGlobalOzoneMonitoringExperiment(GOME):Missionconceptandfirstscientificresults,J.Atmos.Sci.151–175,doi:10.1175/1520-0469(1999)05651:-6;TGOMEG2.0.Carn,S.A.,A.J.Krueger,G.J.S.Bluth,S.J.Schaefer,N.A.Krotkov,I.M.Watson,andS.Datta(2003),VolcaniceruptiondetectionbytheTotalOzoneMappingSpectrometer(TOMS)instruments:A22-yearrecordofsulphurdioxideandashemissions,inVolcanicDegassingeditedbyC.Oppenheimer,D.M.Pyle,andJ.Barclay,Spec.Publ.Geol.Soc.London,177–202.Carn,S.A.,N.A.Krotkov,M.A.Gray,andA.J.Krueger(2004),FireatIraqisulfurplantemitsSOcloudsdetectedbyEarthProbeTOMS,Geophys.Res.Lett.,L19105,doi:10.1029/2004GL020719.Carn,S.A.,A.J.Krueger,N.A.Krotkov,K.Yang,andP.F.Levelt(2007),SulfurdioxideemissionsfromPeruviancoppersmeltersdetectedbytheOzoneMonitoringInstrument,Geophys.Res.Lett.,L09801,doi:10.1029/2006GL029020.Chin,M.,etal.(2000),AtmosphericsulfurcyclesimulatedintheglobalmodelGOCART:Comparisonwithfieldobservationsandregionalbud-gets,J.Geophys.Res.105(D20),24,689–24,712,doi:10.1029/2000JD900385.d’Almeida,G.A.,P.Koepke,andE.P.Shettle(1991),AtmosphericAero-sols:GlobalClimatologyandRadiativeCharacteristics,557pp.,A.DeepakPubl.,Hampton,Va.Dave,J.V.(1964),Meaningofsuccessiveiterationoftheauxiliaryequationofradiativetransfer,Astrophys.J.,1292–1303,doi:10.1086/148024.Dickerson,R.R.,etal.(2007),AircraftobservationsofdustandpollutantsovernortheastChina:InsightintothemeteorologicalmechanismsofJ.Geophys.Res.112,D24S90,doi:10.1029/2007JD008999.Dobber,M.R.,Q.Kleipool,R.Dirksen,P.F.Levelt,G.Jaross,S.Taylor,T.Kelly,andL.Flynn(2008),ValidationofOzoneMonitoringInstru-mentlevel1bdataproducts,J.Geophys.Res.113,D15S06,doi:10.1029/2007JD008665.Draxler,R.R.,andG.D.Rolph(2003),HYSPLIT(HYbridSingle-ParticleLagrangianIntegratedTrajectory)model,NOAAAirResour.Lab.,SilverSpring,Md.(Availableathttp://www.arl.noaa.gov/ready/hysplit4.html)Eisinger,M.,andJ.P.Burrows(1998),Troposphericsulfurdioxideob-servedbytheERS-2GOMEinstrument,Geophys.Res.Lett.,4177–4180,doi:10.1029/1998GL900128.Khokhar,M.F.,C.Frankenberg,M.VanRoozendael,S.Beirle,S.Ku¨hl,A.Richter,U.Platt,andT.Wagner(2005),Satelliteobservationsofatmo-sphericSOfromvolcaniceruptionsduringthetime-periodof1996–Adv.SpaceRes.(5),879–887,doi:10.1016/j.asr.2005.04.114.King,M.D.,W.P.Menzel,Y.J.Kaufman,D.Tanre,B.Gao,S.Platnick,S.A.Ackerman,L.A.Remer,R.Pincus,andP.A.Hubanks(2003),Cloudandaerosolproperties,precipitablewater,andprofilesoftempera-tureandwatervaporfromMODIS,IEEETrans.Geosci.RemoteSens.(2),442–458,doi:10.1109/TGRS.2002.808226.Krotkov,N.A.,S.A.Carn,A.J.Krueger,P.K.Bhartia,andK.Yang(2006),BandresidualdifferencealgorithmforretrievalofSOfromtheAuraOzoneMonitoringInstrument(OMI),IEEETrans.Geosci.RemoteSens.(5),1259–1266,doi:10.1109/TGRS.2005.861932.Krotkov,N.,A.Krueger,K.Yang,S.Carn,P.K.Bhartia,andP.Levelt(2007),SODatafromtheOzoneMonitoringInstrument(OMI),Pro-ceedingsoftheENVISATSymposium2007[CD-ROM],editedbyH.LacosteandL.Ouwehand,Eur.SpaceAgencySpec.Publ.ESASPKrueger,A.J.,L.S.Walter,P.K.Bhartia,C.C.Schnetzler,N.A.Krotkov,I.Sprod,andG.J.S.Bluth(1995),Volcanicsulfurdioxidemeasurementsfromthetotalozonemappingspectrometerinstruments,J.Geophys.Res.(D7),14,057–14,076,doi:10.1029/95JD01222.Krueger,A.J.,S.J.Schaefer,N.Krotkov,G.Bluth,andS.Barker(2000),Ultravioletremotesensingofvolcanicemissions,inRemoteSensingofActiveVolcanismGeophys.Monogr.Ser.,vol.116,editedbyP.J.Mougi-nis-Mark,J.A.Crisp,andJ.H.Fink,pp.25–43,AGU,Washington,D.C.Larssen,T.,etal.(2006),AcidraininChina,Environ.Sci.Technol.418–425.Levelt,P.F.,G.H.J.vandenOord,M.R.Dobber,A.Ma¨lkki,H.Visser,J.deVries,P.Stammes,J.Lundell,andH.Saari(2006),TheOzoneMonitoringInstrument,IEEETrans.Geosci.RemoteSens.1093–1101,doi:10.1109/TGRS.2006.872333.Li,C.,L.T.Marufu,R.R.Dickerson,Z.Li,T.Wen,Y.Wang,P.Wang,H.Chen,andJ.W.Stehr(2007),InsitumeasurementsoftracegasesKROTKOVETAL.:OMISOVALIDATIONOVERNECHINA12of13 andaerosolopticalpropertiesataruralsiteinnorthernChinaduringEastAsianStudyofTroposphericAerosols:AnInternationalRegionalExperiment2005,J.Geophys.Res.112,D22S04,doi:10.1029/2006JD007592.Li,Z.,etal.(2007),PrefacetospecialsectiononEastAsianStudiesofTroposphericAerosols:AnInternationalRegionalExperiment(EAST-J.Geophys.Res.112,D22S00,doi:10.1029/2007JD008853.Luke,W.T.(1997),Evaluationofacommercialpulsedfluorescencede-tectorforthemeasurementoflow-levelSOconcentrationsduringtheGas-PhaseSulfurIntercomparisonExperiment,J.Geophys.Res.16,255–16,265,doi:10.1029/96JD03347.Massie,S.T.,O.Torres,andS.J.Smith(2004),TotalOzoneMappingSpectrometer(TOMS)observationsincreasesofAsianaerosolinwinterfrom1979to2000,J.Geophys.Res.109,D18211,doi:10.1029/2004JD004620.McPeters,R.D.,G.J.Labow,andJ.A.Logan(2007),Ozoneclimatolo-gicalprofilesforsatelliteretrievalalgorithms,J.Geophys.Res.112D05308,doi:10.1029/2005JD006823.Mishchenko,M.,andI.Geogdzhayev(2007),Satelliteremotesensingre-vealsregionaltroposphericaerosoltrends,Opt.Express,7423–7438,doi:10.1364/OE.15.007423.Ohara,T.,H.Akimoto,J.Kurokawa,N.Horii,K.Yamaji,X.Yan,andT.Hayasaka(2007),AnAsianemissioninventoryofanthropogenicemissionsourcesfortheperiod1980–2020,Atmos.Chem.Phys.4419–4444.Richter,A.,F.Wittrock,andJ.P.Burrows(2006),SOmeasurementswithSCIAMACHY,inProceedingsoftheAtmosphericScienceConference,8–12May2006,ESRIN,Frascati[CD-ROM],Eur.SpaceAgencySpec.ESASPSchoeberl,M.R.,etal.(2006),OverviewoftheEOSAuraMission,Trans.Geosci.Rem.Sens.(5),1066–1074,doi:10.1109/TGRS.2005.861950.Seinfeld,J.H.,andS.N.Pandis(1998),AtmosphericChemistryandPhysics:FromAirPollutiontoClimateChange,1326pp.,JohnWiley,NewYork.Sokolik,I.N.,andO.B.Toon(1999),IncorporationofmineralogicalcompositionintomodelsoftheradiativepropertiesofmineralaerosolfromUVtoIRwavelengths,J.Geophys.Res.(D8),9423–9444,doi:10.1029/1998JD200048.Streets,D.G.,andS.T.Waldhoff(2000),PresentandfutureemissionsofairpollutantsinChina:SO,NO,andCO,Atmos.Environ.(3),363–374,doi:10.1016/S1352-2310(99)00167-3.Streets,D.G.,etal.(2003),AninventoryofgaseousandprimaryaerosolemissionsinAsiaintheyear2000,J.Geophys.Res.(D21),8809,doi:10.1029/2002JD003093.Taubman,B.F.,J.C.Hains,A.M.Thompson,L.T.Marufu,B.G.Doddridge,J.W.Stehr,C.A.Piety,andR.R.Dickerson(2006),Air-craftverticalprofilesoftracegasandaerosolpollutionoverthemid-AtlanticUnitedStates:Statisticsandmeteorologicalclusteranalysis,J.Geophys.Res.,D10S07,doi:10.1029/2005JD006196.Thomas,W.,T.Erbertseder,T.Ruppert,M.VanRoozendael,J.Verdebout,D.Balis,C.Meleti,andC.Zerefos(2005),OntheretrievalofvolcanicsulfurdioxideemissionsfromGOMEbackscattermeasurements,J.At-mos.Chem.,295–320,doi:10.1007/s10874-005-5544-1.Torres,O.,A.Tanskanen,B.Veihelmann,C.Ahn,R.Braak,P.K.Bhartia,P.Veefkind,andP.Levelt(2007),AerosolsandsurfaceUVproductsfromOzoneMonitoringInstrumentobservations:Anoverview,J.Geo-phys.Res.112,D24S47,doi:10.1029/2007JD008809.Yang,K.,N.A.Krotkov,A.J.Krueger,S.A.Carn,P.K.Bhartia,andP.F.Levelt(2007),RetrievaloflargevolcanicSOcolumnsfromtheAuraOzoneMonitoringInstrument:Comparisonandlimitations,J.Geophys.112,D24S43,doi:10.1029/2007JD008825.P.K.Bhartia,LaboratoryforAtmospheres,NASAGoddardSpaceFlightCenter,Greenbelt,MD20771,USA.S.A.CarnandA.J.Krueger,JointCenterforEarthSystemsTechnology,UniversityofMaryland,BaltimoreCounty,Baltimore,MD21250,USA.H.Chen,D.Lu,andP.Wang,InstituteofAtmosphericPhysics,ChineseAcademyofSciences,Beijing100021,China.R.R.DickersonandB.McClure,DepartmentofChemistry,UniversityofMaryland,CollegePark,MD20742,USA.N.A.KrotkovandK.Yang,GoddardEarthSciencesandTechnologyCenter,UniversityofMaryland,BaltimoreCounty,Baltimore,MD21228,USA.(krotkov@mhatter.gsfc.nasa.gov)P.F.Levelt,RoyalNetherlandsMeteorologicalInstitute,NL-3730AEDeBilt,Netherlands.C.LiandZ.Li,DepartmentofAtmosphericandOceanicSciences,UniversityofMaryland,CollegePark,MD20742,USA.KROTKOVETAL.:OMISOVALIDATIONOVERNECHINA13of13