Content uploaded by Xiao-Ying Yu
Author content
All content in this area was uploaded by Xiao-Ying Yu
Content may be subject to copyright.
AtmosphericPollutionResearch2(2011)89‐98
Atmospheric Pollution Research
www.atmospolres.com
Insitu
monitoringoftracegasesinanon–urbanenvironment
JohnMioduszewski1,2,Xiao‐YingYu1,VictorMorris1,CarlBerkowitz1,JuliaFlaherty1
1PacificNorthwestNationalLaboratory,AtmosphericSciences andGlobalChangeDivision,Richland,WA99354,USA
2NowatRutgersUniversity,DepartmentofGeography,Piscataway,NJ08854,USA
ABSTRACT
Asetofcommercialinstrumentsmeasuringcarbonmonoxide(CO),ozone(O3),sulfurdioxide(SO2),andnitrogen
oxides[nitricoxide(NO),nitrogendioxide(NO2),andoddnitrogens(NOX)]wasintegratedanddeployedinanon–
urbanenvironment.ThedeploymentoccurredbetweenJuly2,2007andAugust7,2007inRichland,WA.Themixing
ratiosofallspecieswerelowerthaninmostrural–suburbanenvironments,andstrongdiurnalpatternswere
observed.NO2wasdepletedbyphotochemicallyformedozoneduringthedayandreplenishedatnightasozonewas
destroyed.Thehighestozoneconcentrationduringtheseepisodeswas45ppb.Theoverallaveragewas15ppbwith
readingsapproachingnearzeroattimes.Thisobservationislowcomparedtoaveragedaytimesummerreadingsof
60–80ppbinhighlypopulatedandindustrializedurbanareasinthePacificNorthwestregion.Back‐trajectoryanalysis
andprevailingweatherconditionsbothindicatedthatmuchoftheozonewastransportedlocallyorwasproducedin–
situ.AnalysisofSO2asatracerforO3advectionfurtherindicatedlackoflong–rangeregionaltransportofpollutants
toRichland.Wealsopresentresultsofanalysisofhighozoneepisodesandcomparisonsrelativetootherareasinthe
PacificNorthwestregion.Theseresultsprovideausefulsampledatasettostudythehistoricalrecordofairqualityin
ruralEasternWashington.
Keywords:
Tracegas
HYSPLIT
Non‐urbanenvironment
Ozone
ArticleHistory:
Received:28May2010
Revised:20September2010
Accepted:06October2010
CorrespondingAuthor:
Xiao‐YingYu
Tel:+1‐509‐3724524
Fax:+1‐509‐3726168
E‐mail:xiaoying.yu@pnl.gov
©Author(s)2011.ThisworkisdistributedundertheCreativeCommonsAttribution3.0License. doi:10.5094/APR.2011.011
1.Introduction
Atmospherictracegasesarechemicalcompoundsfoundin
verylowconcentrationsintheatmosphere;despitelowconcen‐
trations,however,theycanexertconsiderableinfluenceona
rangeofenvironmentalprocessesandhealthproblems(Seinfeld,
2004).Knowledgeoftheinteractionsamongthesegasesiscrucial
tounderstandtheiratmosphericconcentrationsandlifetimesand
theenvironmentalimpactsthatcanbeexpectedwithmodifica‐
tionstotheirsourcesandsinks.Ozone(O3),nitrogenoxides
(includingNO,NO2,andNOX),carbonmonoxide(CO),andsulfur
dioxide(SO2)arenotonlypollutantsthemselvesbutalsoreactwith
manyothercompoundssuchasvolatileorganiccompounds(VOCs)
leadingtochangesinatmosphericcomposition(Atkinson,2000).
Accuratein–situmeasurementsarecrucialtoprovidethe
foundationforinvestigatingcomplexphoto–oxidationprocesses.
Withpopulationgrowthbeingacommonphenomenoninmany
ruralareas,historicalrecordsarebecomingincreasinglyimportant
toidentifyairqualitytrendsassociatedwithlocalsourcesrelative
toincreasesresultingfromlongrangetransport.TheTri–Cities,
includingRichland,Pasco,andKennewickinthestateof
Washington,isoneofthestatisticallyfastestgrowingmetropolitan
areasinthecountry,definedasacoreurbanareawithatleast
50000people(USCBa,2008).Itaddedalmosttwentypercenttoits
populationinthelastsixyears(Cohen,2007),indicatingthe
importanceofdocumentingairqualitybeforetheareabecomes
farmoredeveloped.
Thestudypresentedherehastwopurposes.Froman
engineeringstandpoint,themotivationwastointegratemultiple
tracegasanalyzersintoonesystemandidentifyissuesrelatedto
thisintegrationprocess.Combiningbasictracegasanalyzersto
onesystemhasseveraladvantages.First,itsavesspace.Second,it
isconvenienttomoveortransfertheseanalyzers.Third,itprovides
integrateddataacquisitionandconsistenttimestampsforeaseof
datacollection,display,andanalysis.Inaddition,theThermo
Electron,Inc.,tracegasanalyzersusedinthisstudywerebeing
updatedfromC–Seriestoi–Series.Theneweri–Seriesinstruments
offerimprovedmeasuringcapabilities,useroptions,andstorage
space.Thisworkgivesapracticalsolutiontoissuesassociatedwith
acquiringdatafrominstrumentshavingdifferentmanufacturer
configurations.Forinstance,comparedwithstandardserial
connections,ourapproachprovidesfasttimeresolutionaseach
instrumentallows.Italsoprovidesflexibilityforuserstosettheir
owninstrumentconnectionscheme.
Thesecondpurposeofthisstudywastoprovideabasecase
ofpollutantconcentrations,andinparticular,O3concentrations,in
Richland,WA.RichlandisaruralyetgrowingareainthePacific
NorthwestregionoftheUnitedStates,whichatpresenthasozone
concentrationsfarlessthanthosefoundinurbanpartsofthe
UnitedStates.Ruralareashavehistoricallybeenunderrepresented
whenstudyingairpollutionandadatasetforRichlandpriorto
majorgrowthmayproveveryvaluableinitsfuture.Sourceregions
associatedwith“high”ozoneepisodeswerestudiedusingthe
NationalOceanicandAtmosphericAdministration’s(NOAA)
HYSPLITmodeltoidentifythebacktrajectoriesofairparcels
associatedwithelevatedpollutants.Thedeploymenttookplacein
summer,becausephotochemistryismostvigorous.Asaresult,
ozoneoftenpeaks.Thispotentiallycanprovideagoodopportunity
tostudyozoneatthisrurallocation.
90Mioduszewskietal.–AtmosphericPollutionResearch2(2011)89‐98
2.MaterialsandMethods
2.1.Instrumentationandexperimentalsetup
Measurementsofthetracegasesweremadeusingindividual
tracegasanalyzers(ThermoElectron)forO3,CO,SO2,and
NO/NO2/NOX.Theinstrumentswereinstalledinaninstrument
rack,andtheresultingunitdeployedjustsouthofthemain
campusofthePacificNorthwestNationalLaboratory(PNNL)in
Richland,WA(lat46°21’N,long119°17’W,elevation120mMSL)
(Figure1).MapsweremadeusingtheESRIArcMapsoftware
(version9.3).Thegeographicinformationwasobtainedfromthe
WashingtonStateDepartmentofTransportation(WSDOT,2010).
Theinstrumentswerearrangedintherackaccordingtohowdata
wereloggedinsequence.Aswillbedescribedinmoredetail,the
O3analyzerwasusedtocommunicatebetweenthedatalogger
andtherestoftheinstruments.Samplecollectionwasperformed
betweenJuly2andAugust7,2007.Localstandardtime(Pacific
DaylightSavingTime,PDT)wasusedindatarecording.The
conversionbetweenuniversaltime(UTC)andPDTisUTC–7hours
=PDT.
Figure1.TopisamapofthePacificNorthwestwithselectedozone
monitoringstationsidentified(NCDC,2007).Bottomisamapof
Washington’sTri‐citiesarea,includingtheobservationsiteatPNNLdenoted
bytheblackdot.
AlloftheinstrumentswereconnectedtoaCampbell
Scientific,Inc.,(C.S.)CR23XMicrologger.Thedatalogger
communicatedwithaDellLaptopforon–linedisplayviaaserial
cable.ThedataacquisitionprogramwaswritteninEdlog,a
programmingtoolwithinC.S.’sLoggerNet(C.S.,2002),and
downloadedtothedatalogger.Thedataloggerwasusedtostore
selectedinstrumentoutputsandcontrolinstrumentzeroandspan
checkstwiceaday.Instrumentreadingsweremeasuredas
voltagesfromtheanalogvoltageoutputsoftheinstrumentsand
transmittedtothedifferentialvoltageanaloginputsofthedata
logger(seetheSupportingMaterial,SM,FiguresS1andS2).
Ambientairwassampledthrough¼–inTeflonsamplingtubing,
whoseinletwasapproximatelyfourmetersabovethegroundand
passedthroughasamplingportinthewall.Figure2illustrateshow
airpassedthroughTeflonfilters(2µmpore,47mmZefluor,
GelmanP5PJ047)beforeenteringintotheinstruments’sample
bulkhead.Theexhaustlinesfromeachinstrumentwerecollected
intoacommonmanifoldthatreturnedtheairtotheatmosphere
afterscrubbingtheexhaustinapackfilledwithactivatedcarbon.
ADynamicGasCalibrator(ThermoElectron,Model146i)was
usedtocalibratetheinstrumentsautomaticallyonaregularbasis
usingzeroandstandardgases(ScottSpecialtyGases).Thedynamic
calibratorusestwomassflowcontrollerstoregulatetheflowof
zeroandspangas,andaTeflonmixingchambertoachieve
completemixingofthetwocomponentsatthedesired
concentrationlevel.Highpurityzeroaircontaininglessthan
0.5ppmtotalhydrocarbonswasused,whilethestandardgas
mixtureconsistedof10.1ppmNO,10.1ppmCO,and10.0ppm
SO2,balancedwithhighpurityN2.Theozoneanalyzerwas
calibratedusingaprimarystandardultraviolet(UV)photometric
ozonecalibrator(ThermoElectron,Model49C),followingtheEPA
transferstandardsforcalibrationofairmonitoringofanalyzersfor
ozone(EPA,2009).
AUVPhotometricO3Analyzer(ThermoElectron,Model49i)
wasusedtomeasureO3.AtmosphericSO2wasdeterminedwitha
TraceLevel–EnhancedPulsedFluorescenceSO2Analyzer(Thermo
Electron,Model43i).COwasdeterminedbyaTraceLevelCO
Analyzer(ThermoElectron,Model48C).Nitrogenoxides(NO,NO2,
andNOX)weremeasuredwithaTraceLevelChemiluminescence
NO–NO2–NOXanalyzer(ThermoElectron,Model42C).The
ozonatorusedtoconvertNOtoNO2requiresasupplyofdryair,
andDrierite(Cole–Parmer)wasusedtoremovewatervaporfrom
ambientairforthispurpose.Amoredetaileddescriptionofthe
principlesofoperationoftheseinstrumentsisreportedelsewhere
(Platt,1999)andtheoperationprinciplesofeachanalyzerare
brieflysummarizedinTable1.
APrecisionSpectralPyranometer(PSP)(EppleyLaboratory)
wasusedtomeasurethesolarradiationreachingEarth'ssurface.
ThePSPusesathermoelectricdevicethatproducesanelectric
currentproportionaltothebroadbandshortwavesolarradiation
reachingthedetector.Thedetectorispaintedblackandmounted
inanopticalglasssphereforprotectionfromtheelements.It
samplesatwavelengthsbetween0.3µmand3µmatoneminute
intervals.Ahumiditymeasurementprobe(Vaisala,Model
HMP45C)wasusedtomeasurethetemperature(indegrees
Celsius)andrelativehumidityofthesampledair.Windspeedand
directionweremeasuredbyapropelleranemometer(R.M.Young,
Model05103WindMonitor)locatedaboutfifteenmeterstothe
northofthetracegassamplingarea.Datawerecollectedevery
minuteandaveragedtofiveminutesforfinalreporting.
2.2.Instrumentationerror
Backgroundconcentrationchecksusingzeroairwere
conductedtocorrectinstrumentdrift.High–purityairwassentinto
theinstrumentstoobtainadailytruezero.ItiscriticalthattheCO
instrumentbecontinuouslypurgedwithalowpurgeflow(140
cc/minuterecommended)ofzeroair,otherwisesignificantdrift
mayinterferewithdataquality.Purgingisusedtoprevent
interferencefromambientlevelsofCOasairflowsthroughthe
filterwheelassembly,whichcontainsCOononesideandN2onthe
other.Allanalyzersunderwentbackgroundcorrectionsasoftenas
timepermitted,andtheCOanalyzerwascontinuouslypurgedwith
highpurityzeroairwhilesamplingwasoccurring.
Mioduszewskietal.–AtmosphericPollutionResearch2(2011)89‐9891
Figure2.AschematicdiagramoftheThermoElectronTraceGasAnalyzersuite.Thedarkestlinesindicatetubingthroughwhichambientairflows.
Table1.Basicpropertiesofthetracegasanalyzersuite(allareThermoElectroninstruments)
InstrumentPrimaryOperationPrincipleResponseTime
(seconds)
DetectionLimit
(ppb)ZeroDrift
O3(49i)O3absorptionofultravioletradiation(240nm)200.5<1ppb
NO/NO2/NOX(42C,TraceLevel)ChemiluminescencefromtheO3‐NOreaction600.05Negligible
CO(48C,TraceLevel)COabsorptionofinfraredradiation(4.6µm)6040<0.1ppm
SO2(43i)Pulsedultravioletfluorescence602<1ppb
3.ResultsandDiscussions
3.1.Overallobservations
Approximatelyfiveweeksofdatawerecollectedandthetime
seriesispresentedinFigure3.Theforemostobservationisthatall
concentrations,exceptforthatofCO,werequitelowrelativeto
valuesreportedelsewhere.SO2measurementswereonthelow
endofthetypicalmixingratiosof1–20ppbinrural–suburban
environments,agreeingwithpastfindings(Finlayson–Pittsand
Pitts,2000).NO,NO2,andNOXaveraged0.6,12.2,and12.7ppb,
respectively,whileCOaveraged155ppb.TypicalO3mixingratios
insimilarenvironmentspeakat80–150ppb,butthehighest
mixingratioofO3observedinRichlandwas∼45ppb.
Table2givesthestatisticalsummaryofthecollecteddata.The
maximummixingratioofanygivenspecieswaslowerthantypical
mixingratiosinmanyotherpartsofthecountry,withminima
sometimesapproachingthedetectionlimitoftheinstruments
(Table1).O3,aswellasNOX,hadthehigherstandarddeviations
relativetotheirrespectivemeanvaluesascomparedtoSO2and
CO.TheminimumCOvalueislowerthanliteraturevaluesin
midlatitudesinthenorthernhemisphere.Forinstance,the
minimumCOobservedbyMOPITTis50ppbinsummer,although
thiswasobtainedatmuchhigheraltitudepressureof280hPa(El
Amraouietal.,2010).Ingeneral,thepressureatourobservation
siteinsummeris990hPa.However,theaverageCOvalueissimilar
tootherlocationsbetween40–60°N,i.e.,∼105–155ppb(Wang
etal.,2003).
Therewerefourepisodesofelevatedtracegasconcen‐
trations:July7,2007(Julianday188),July13,2007(Julianday
194.6–194.9),July26,2007(Julianday207.4–207.6),andAugust2,
2007(Julianday214.4–214.7).Thesewereperiodswhenmostof
thespeciesexhibitedarapidincreaseinmixingratioforatleasta
fewhours.O3reacheditshighestconcentrationof44.7ppbonJuly
13,2007,SO2reachedahighof14.3onAugust2,2007,andNO,
NO2,andNOXof21.2,34.6,and54.6ppb,respectively,onAugust
2,2007.Diurnalminimummixingratioswerealsohigherthan
normalduringthesetimes,including10ppbofO3onJuly6,2007
andnear5ppbofSO2onJuly7,2007.Duetolackofwindspeed
andwinddirectiondataduringJulianday188,onlythreeoutofthe
fourepisodeswerestudiedinmoredetail.Thesearehighlightedin
thetimeseries(Figure3),i.e.EpisodesI,II,andIII.
Figure3depictsthetemporalvariationsofthetracegas
speciesandmainmeteorologicalparametersobservedduringthe
deployment.Changesinthemixingratioswerefrequently
associatedwithchangesinphotochemistry.Therelationship
betweentemperatureandO3hasbeenwellestablished,namely,
heatacceleratingthechemicalreactionsintheatmosphereresults
inhigherozoneconcentrations(ClarkandKarl,1982;Jacob,1993).
Threeepisodeswithelevatedconcentrationswerenotedin
chronologicalorderoverperiodsofapproximately1day(identified
inFigure3).Inmostepisodes,therewasaslowincreasein
concentrationofallspeciestoapeakonthedayonwhichthe
episodeisdefined,whereupontheconcentrationsdecreasedand
anothercyclestarted.ThescatterplotbetweenhourlyaverageO3
vs.GSWisillustratedinFigure4.Thesolidlineisthelinearleast–
squaresfit.
92Mioduszewskietal.–AtmosphericPollutionResearch2(2011)89‐98
Figure3.Timeseriesofalltracegases,solarradiation(GSW),relativehumidity(RH),andtemperature(°C)overtheentiresamplingperiod(timeisinJulian
Days).Threeepisodesofelevatedtracegasconcentrationsaredenoted(I,II,andIII)intheO3plot.Superimposedaretheaveragehourly
valuesofeachtracegasspecies.Relativehumiditymeasurementswereinterruptedperiodicallyduetoinstrumentproblems.
Table2.Statisticsencompassingtheentireperiodforallspecies,allinppb.
Maximaandminimastatisticsrepresentone‐minutevalues,whilethe
averageandstandarddeviationstatisticswerederivedfromvalues
averagedeveryhour
SpeciesMin.(ppb)Max.(ppb)Avg.(ppb)Std.Dev.(ppb)
O30.144.715.45.2
NO021.20.60.5
NO2036.712.24.2
NOx0.254.612.74.7
CO35.7327.915517.8
SO20.714.44.90.4
3.2.Temporalvariation
Diurnalpatternswereobservedinthetimeseriesalthough
therewerenotableexceptions.DiurnalvariationsinO3,NO,NO2,
andNOXweredistinct(Figure3),withminimumconcentrations
oftenapproachingzero.Photo–oxidationofprecursorgases,like
carbonmonoxide,isprimarilyresponsibleforthedailyriseinO3
(Nairetal.,2002).Providedthereisasufficientamountof
NO/NOX,theO3willpeakeachdayintheearlyafternoonwitha
correspondingdipinnitrogenoxides(Ghudeetal.,2006).The
diurnalvariationofO3andNOXisshowninFigure5,withNOX
reachingitspeakaround6am(PDT)andO3reachingabroader
peakbetween1pmand3pm.O3productionduringthedayis
drivenbythephotochemicalreactionbetweenhydroxylradicals,
organicperoxyradicals,andNO,whileitisremovedatnightby
depositionanddestructionbyalkenesandNO(Gerasopoulos,
2006).TheconversionofNOtoNO2byO3duringthenightisthe
primaryreactionthatincreasesNO2atnight,withthereverse
occurringduringthedaytoincreaseO3anddecreaseNO2(Mazzeo
etal.,2005).
Figure4.AscatterplotofhourlyaverageO3vs.hourlyaveragesolar
radiation(GSW).Thesolidlineisthelinearleast‐squaresfit.
Althoughthedetailsofthesynopticpatternvariedfrom
episodetoepisode,thereweredistinctsimilarities.Strongridging
waspresentat500mbduringEpisodesIandIIthatgavewayto
relativelystrongcoldfronts,astheridgeaxisprogressedeastward
andlowpressureapproachedfromthenorthwest.Theupperlevel
featuresforthelattertwoepisodeswerelessamplifiedwitha
morezonalflow,butacoldfronttrailingasurfacelowpressure
systemtothenorthstillmovedthroughthearea(HPC,2007).
Winddirectionsvariedandhadminimalinfluenceonpollutant
Mioduszewskietal.–AtmosphericPollutionResearch2(2011)89‐9893
concentrations.Instead,thepresenceofastagnantairmass
appearedtocontributethemosttoepisodesofhightracegas
concentrations,agreeingwithpastobservationsthathighO3is
bestcorrelatedwithlightwindsunderhighpressure(Vukovich,
1995).
Figure5.ThehourlyaverageofNOX,CO,O3,andSO2mixingratios,andsolar
radiationduringtheentiresamplingtime.Thestandarddeviationsare
plottedaserrorbars.
3.3.Pollutanttransport
NOAA’sHYSPLITmodelisoftenusedtocomputeair
trajectoriesatdifferentlevelsoftheatmosphere.Ifairpollutants
canbeassumedtotravelwiththemeanairflow,thentheir
trajectoriescanbecalculatedusingameteorologicalmodelthat
describesairmotionatdifferentlevelsoftheatmosphere.Figure6
showsbackwardairtrajectoriesproducedbyHYSPLITduringthree
episodesofelevatedtracegasconcentrations.Thetrajectories
usedGDAS1modeldataona1degreelatitudelongitudespatial
resolution,andusedGDAS1verticalvelocitytomodelvertical
motion(DraxlerandRolph,2003).
ThesecondarywindroseofO3asafunctionofwinddirection
isshowninFigure7.PanelIdepictsEpisodeIwhenthetrace
gases,particularlyO3,registeredhigh.Windwasclearlyfromthe
southeastandsouthduringthistime,whichisalsothelocationof
themostlikelylocalpollutionsources.Thelargerandmore
industrialcitiesofPascoandKennewicklietentofifteenmilesto
theeastandsoutheast(seeFigure1),whilenearbytothesouth
aremajorroads,suchasHighway240,thataccommodate
considerabletrafficattimes.Thesearealllikelylocalsourcesfor
thisparticularepisodeofelevatedtracegases.Windwas
predominantlyfromthenorthandnorthwestduringEpisodesII
andIIIwheretherearemainlyopenlandsandlackofpollution
pointsources(Figure1).Thewindspeedwasmuchweakerduring
theseepisodesowingtotheanticyclonenearby,andmuchofthe
pollutionwaslikelyduetothestagnantairmasspresentinboth
episodes,asdiscussedpreviously(Vukovich,1995).
Therearetwotrajectoriesforeachepisode,whichare
centeredthreehoursaroundthehourofthehighestmixingratios.
Windroseswerecalculatedforthesametimetodetermineif
therewasaclearsourcefortheincreaseintheobservedtracegas
concentration(seeFigure7).Thewinddataandbacktrajectories
agreereasonablywellwitheachotherconsideringHYSPLITdoes
notmodeltherelativelyturbulentboundarylayer(Draxlerand
Hess,1998).Thedominantwinddirectionwassomewhatdifferent
ineachepisode,rangingfromsouthtonorthwestandnorth.The
southerlywindontheeveningofJuly13duringEpisodeIcould
havebroughtinpollutantsprimarilyfromtrafficfromtheTri–cities
(particularlyGeorgeWashingtonWay;seeFigure1)while
pollutantsfromthenorthandnorthwestonJuly26andAugust2
(Julianday207and214)duringEpisodesIIandIIIrespectively
couldhavebeennearbyroadssuchasHighway240andStevens
Drive,whicharetheothertwomainnorth–southroadsin
Richland.
TimeseriesofO3andSO2duringthethreeO3episodeswere
plottedinFigure8.OnlyinEpisodeIdidO3andSO2peakatthe
sametime,whichindicatedthataplumewastransportedfromthe
southandcausedelevatedO3attheobservationsite.During
EpisodesIIandIII,therisingofO3concentrationdidnotcoincide
withthatofSO2.ThisindicatedthatO3wasfromadifferentsource
andpossiblymainlycausedbylocaltraffic.
3.4.Ozoneobservationsandcomparison
ProductionofO3canbeaffectedbyNOX,VOCreactivity,and
freeradicalproduction(WalcekandYuan,1995).NOXplaysa
criticalroleinthephotochemicalformationofO3,andhasbeen
foundtobealimitingfactorintheatmosphereatruralandremote
locations(Finlayson‐PittsandPitts,2000).Theprincipleformation
ofozoneisbythereactionofatomicoxygen(O)anddiatomic
oxygen(O2).Inthetroposphere,themajorsourceofatomic
oxygenisfromthephotochemicalcycleinvolvingNO,NO2,and
photolysis.ToolittleNOX,forexample,resultsin“NOXsensitive”O3
chemistry,wheretheamountofO3thatcanbeproducedislimited
bytheamountofNOXavailabletoreact(Kleinmanetal.,2005).
SinceVOCmeasurementswerenotavailableonsite,itwasnot
possibletoinvestigatethechemistrybetweenVOCsandO3/NOX.
However,theanalysisofNOXandO3wasconductedwiththe
understandingthatO3productionwasinfluencedbyNOX
concentration.
94Mioduszewskietal.–AtmosphericPollutionResearch2(2011)89‐98
Figure6.12hourbackwardtrajectoriesofairovertheregionusingHYSPLIT,startingatIa)15:00h13July2007,Ib)21:00h14July2007,IIa)06:00h26July
2007,IIb)10:00h26July2007,IIIa)04:00h1August2007,andIIIb)10:00h1August2007(alltimesinPDT).TheGDAS1modelisusedforextrapolation,
beginningatheightsof500m,1000m,and2000mabovetheground.
Figure7.ToparewindrosesdisplayingsurfacewindspeedanddirectionforEpisodesI,II,andIII.Bottomarewindroses
displayingsurfacewinddirectionandozoneconcentrationforthesameepisodes.
Mioduszewskietal.–AtmosphericPollutionResearch2(2011)89‐9895
Figure8.TimeseriesofO3andSO2duringthethreeepisodes.Bluediamond
isSO2,reddiamondisO3.
ThescatterplotsofO3vs.NOXandOxvs.NOXareshownin
Figure9.ThetoptwopanelsdepictthehourlyaverageofO3vs.
NOXandOxvs.NOXfortheentiresamplingperiod.O3increased
withdecreasingNOXconcentration.Theoxidantconcentration
(OX),where[OX]=[O3]+[NO2],isoftenusedindeterminingthe
dominantchannelforNO2formationinaparticularenvironment
(Itanoetal.,2007).AnincreaseintheOXconcentrationswas
observedwhichimpliestheoccurrenceoftheradicalchannelfor
theformationofNO2.Itisworthmentioningthatintheenviron‐
mentofhighNOXemissions,positivebiascanbeintroducedonthe
OXconcentration.TheNO2toNOX(NOX=NO2+NO)ratioconverges
at∼0.9duringbothepisodes,whichindicates90%ambientNOXas
theprimaryNO2concentration.
Theobservedoxidantconcentrations(OX)canbeinterpreted
intermsofthesumofaNOX–independent“regional”contribution
(i.e.,theO3background)andalinearlyNOX–dependent“local”
contribution(ClappandJenkin,2001;Jenkin,2004a;Jenkin,
2004b).Thesolidlinesarelinearregressionfitsinthetoptwo
panels.TheslopeandinterceptforO3vs.NOXare–0.61±0.05and
21.45±0.82(r2=0.88),respectively.TheslopeandinterceptforOX
vs.NOXare0.39±0.05and21.45±0.82(r2=0.75),respectively.Our
regional(i.e.NOX–independent)contributiontooxidantis21.4
ppb,muchlowercomparedtothosederivedat∼35ppbaveraged
frommultiplesitesinJulyandAugustintheUK(ClappandJenkin,
2001).Thelocal(i.e.NOX–dependent)sourceofOXinRichlandis
muchhigherthanthoseinthesamestudybyClappandJenkin
(2001).Forexample,theoxidantslopesforJulyandAugustinthis
UKstudyare∼0.1,whereasourslopeis∼0.4.Thisindicatesthat
theregionalbackgroundofoxidantinRichlandisfairlylow.Clapp
andJenkin(2001)alsoshowedthatthelocalNOX–dependent
oxidantcontributionhasremarkablylittlevariationwithseason,
i.e.∼10%ofNOXlevelthroughouttheyear.Therefore,the
observationmadeinthesummercouldbequiteusefulin
estimatingthelocaloxidantamountforothertimesoftheyearin
Richland.
OneofthemostimportantlocalNOX–dependentcontri‐
butionsisroadtrafficexhaust.Thisisprobablythemostimportant
sourceofOXinRichland.Anothersourceoflocaloxidantcomes
fromNOreactionwithoxygentoformNO2,whichisdependenton
NOconcentration.ThismaynotbeasignificantsourceinRichland,
sincetheNOconcentrationremainedfairlylow(i.e.avgerage0.6
ppb)throughtheentiresamplingtime.Sunlight–initiatedfree
radicalcatalyzeddegradationofVOCinthepresenceofNOXleads
totheoxidationofNOtoNO2.Thisdependsonthephotolysisrates
ofradicalformationandchainlengthofNO2formation.
UnfortunatelyVOCandphotolysisrateswerenotmeasured,sono
furtherconclusioncanbedrawnfromtheothertwopotential
sourcesoflocaloxidants.
SeveralfactorsmayexplainwhyO3mixingratiosremained
lowinRichland.Rapidfreeradicalproductionislikelytoconsume
muchoftheO3duetothehighsolarfluxobservedinRichland
duringthesummer.Solarflux,measuredasshortwaveradiationin
W/m2,wasplottedwithO3mixingratios,relativehumidity,and
temperature(Figure3).Watervaporenhancestheremovalrateof
freeradicalproduction,particularlyinlow–NOXenvironments,so
thelackofwatervaporinRichlandasrepresentedbylowrelative
humiditysupportstheaforementionedfreeradicalhypothesis
(WalcekandYuan,1995).Thehourlyaveragesolarflux,O3,NOX,
CO,andSO2areplottedinFigure5.Ithasbeendemonstratedthat
O3andsolarradiationarerelated(VukovichandSherwell,2003).
Thehighozonedaysgenerallyconcurwithhighsolarradiation
measurementsonthesurface.Highradiationusuallyresultsin
highersurfacetemperatures.Itisnotsurprisingtoobserveaweak
positivecorrelationbetweensolarradiationandozone,asthe
linearleast–squaresfitproducesaslopeof0.0083±0.0006and
interceptof12.6±0.3withr2=0.18.(Figure4).Summersolar
radiationinRichlandisgreaterthaninmuchofthecountrydueto
lackofcloudcover(NCDC,2008),andthiscouldhavefacilitated
freeradicalproduction.Ontheotherhand,ozonecanbeformed
anddestroyedbyfreeradicals,suchashydroxylorperoxyradicals,
formedfromphotolysis.Pastcalculationsindicatedthatozone
reductionispossibleinmostareaswhentheUVradiationinthe
troposphereincreases,whichgiveschangesinthedissociationrate
forozoneyieldingradicalssuchas(O1D).Thisincreasein
photodissociationinducestropospherichydroxylradicals.While
thelevelsofhydroxylradicalsandhydrogenperoxideincrease,the
levelsoftroposphericozonearegenerallyreduced(Fuglestvedet
al.,1994).Sincethephotolysisrateswerenotmeasured,quantifi‐
cationofthiseffectisprohibited.
EPAusesthefourthhighestdaily8–houraverageO3
concentrationsmeasuredwithinanareaovereachyeartosetup
thenationalambientairqualitystandards(EPA,1998).Ozonecon‐
centrationsinRichland(metropolitanareapopulationof∼200000)
werecomparedtothoseinLosAngeles,CA(16million),Seattle
(3.5million)andSpokane,WA(∼400000),andMission,OR(town
populationof∼1000)in2007(USCBb,2008).Spokanereporteda
fourthmax8–hourO3valueof64ppb,Seattlereported46ppb,
andvalueswerecommonlyover100ppbintheLosAngelesarea
(EPA,2008).Missionisonly54milestothesouth–southeastof
RichlandandistheonlyotherEPAtracegasobservingstationin
thelowerColumbiaBasin.Thefourthmax8–hourO3valueat
Missionwas57ppb.Bycontrast,eventhehighest8–houraverage
96Mioduszewskietal.–AtmosphericPollutionResearch2(2011)89‐98
Figure9.ScatterplotsofO3vs.NOxandOxvs.NOx.ThetwotoppanelsdepicthourlyaverageO3vs.NOxandOxvs.NOx,respectively.Thesolidlinesarelinear
least‐squaresfits.ThebottomthreepanelsareO3vs.NOxandOxvs.NOxusing10minaverageddataduringepisodesI,II,andIII,respectively.
inRichlandwaswellbelow40ppb.Theozoneconcentrationin
RichlandislowercomparedwithotherlocationsalongthePacific
region,wheretheaverageO3valueintheozonegrowingseason
(MaythroughSeptember)rangesfrom19–58ppbin1995and20–
65ppbin1996.Comparedwiththefewlocationswithozone
monitoringinthestateofWashington(WA),theaverageozonein
Richlandisabout5ppblower,i.e.,19–21ppbforotherstationsin
1995,and20–25ppbin1996inWA.Specifically,comparableto
valuesobservedinClassIairsheds,suchastheMountRainier
NationalPark(MOR),theaverageozoneconcentrationinRichland
(e.g.15ppb)issimilartothatinMOR(e.g.17.5ppb)inJulyin
2007.Otherruralareas,suchasGrandCanyonNationalPark(52
ppb),havehigheraverageozoneconcentrationinthesamemonth
thaninRichland(NPS,2010).
EventhoughRichlandandMissionareingeographically–
similarlocations,O3mixingratiosinMissionweremuchhigher
thanRichland.Althoughwecannotdefinitelyestablishan
explanationforthehighervaluesinMission,wewouldnotethatit
iseightkilometersdirectlyeastofamajoragriculturalexchange
Mioduszewskietal.–AtmosphericPollutionResearch2(2011)89‐9897
community(Pendleton,OR).Thereforeitisnotinconceivablethat
therichNOXandVOCsmixtureassociatedwithtruckstopstothe
westmaybeasourceofthehigherO3reportedatthissite.In
addition,bothPendletonandMissionlieontheUmatillaRiver,
whichcanactasanaturalcorridordownwhichpollutantsreadily
travel.Consequently,MissionlikelyreceivesmuchofitsO3by
advectionfromthewest,whereasRichland’sO3isprimarily
generatedin–situwhenthereislittleairmotion.Elevationis
similarbetweenthesetwolocations.Theaverageelevationin
Pendletonis300–365mabovesealevel,andtheelevationin
Missionis370mabovesealevel.Therefore,theeffectofaltitude
onsolarfluxisinsignificantbetweenthesetwolocations(Dvorkin
andSteinberger,1999).WhilethedatafromRichlandcover∼5
weeks,thissamplingperiodcorrespondstoatimeoftheyear
whenO3typicallyreachesitshighestlevels.GiventhelevelsofO3
duringtheentiresamplingperiodatRichland,itisclearthat
photochemicalO3isnotthesameprobleminRichlandasinalarge
urbanarealikeLosAngeles,andthatRichlandisoneofthecleaner
sitesinthePacificNorthwestregion.
4.Conclusions
Thisworkprovidesapracticalexampleinintegrating
instrumentsandapplyingthemtoin–situmeasurements.
Operatingandmaintainingtheinstrumentsforanextended
amountoftimeinthefieldisausefultestwhichservestobolster
confidenceintheirabilitytocollectqualitydata.Moreover,the
datacollectedprovideusefulcharacteristicsoftracegasesinthe
underrepresentednon–urbanenvironment.Thisisofparticular
interesttopollutantstudiesonRichlandbecauseoftheanticipated
growthofthearea.Furthermore,theverylowmixingratiosof
manyofthetracegasspeciesmaybecomeatopicofinterestinthe
futureifpollutantscontinuetoincreaseinthisnon–urbanarea.
Acknowledgements
ThisworkwassupportedbytheOfficeofScienceand
EngineeringEducationandtheAtmosphericScienceandGlobal
ChangeDivisionatthePacificNorthwestNationalLaboratory,
DepartmentofEnergy.JohnMioduszewskiwouldliketothank
KarenWiedaofPNNL’sScienceUndergraduateLaboratory
Internshipprogramforherfinancialsupportoftheinternshipfor
JohnMioduszewski.SupportwasalsofromtheOfficeofScience
(BER),U.S.DepartmentofEnergy,undertheauspicesofthe
AtmosphericScienceProgram,underContractsDE–AC05–
76RL01830atthePacificNorthwestNationalLaboratory.
SupportingMaterialAvailable
Informationon“Gasanalyzerintegrationmethodology”,A
schematicdiagramoftheCR–23XMicrologger’sconnectionstothe
tracegasanalyzers(FigureS1),Aschematicdiagramofthewiring
connectionsbetweenthedataloggerandspecificpinsonthe
terminalblocksofeachinstrument(FigureS2).Thisinformationis
availablefreeofchargeviatheInternetathttp://www.
atmospolres.com.
References
Atkinson,R.,2000.AtmosphericchemistryofVOCsandNOX.Atmospheric
Environment34,2063‐2101.
C.S.DataloggerSupportSoftware,2002.http://www.campbellsci.com.au/
documents/lit/b_loggernet21.pdf.
Clark,T.L.,Karl,T.R.,1982.Applicationofprognosticmeteorological
variablestoforecastsofdailymaximumone‐hourozone
concentrationsinthenortheasternUnitedStates.JournalofApplied
Meteorology21,1662‐1671.
Clapp,L.J.,Jenkin,M.E.,2001.Analysisoftherelationshipbetweenambient
levelsofO3,NO2andNOasafunctionofNOXintheUK.Atmospheric
Environment35,6391‐6405.
Cohen,A.,2007.Seattleareagrew7.2%in6years,butTri‐Citiespopulation
up17.8%.SeattlePost‐Intelligencer.http://seattlepi.nwsource.com/
local/310361_census05.html.
Dvorkin,A.Y.,Steinberger,E.H.,1999.Modelingthealtitudeeffectonsolar
uvradiation.SolarEnergy65,181‐187.
Draxler,R.R.,Rolph,G.D,2003.HYSPLIT(HYbridSingle‐ParticleLagrangian
IntegratedTrajectory)ModelaccessviaNOAAARLREADYWebsite,
http://www.arl.noaa.gov/ready/hysplit4.html.NOAAAirResources
Laboratory,SilverSpring,MD.
Draxler,R.R.,Hess,G.D.,1998.AnoverviewoftheHYSPLIT_4modelling
systemfortrajectories,dispersion,anddeposition.Australian
MeteorologicalMagazine47,295‐308.
ElAmraoui,L.,Attie,J.L.,Semane,N.,Claeyman,M.,Peuch,V.H.,Warner,
J.,Ricaud,P.,Cammas,J.P.,Piacentini,A.,Josse,B.,Cariolle,D.,
Massart,S.,Bencherif,H.,2010.Midlatitudestratosphere–
troposphereexchangeasdiagnosedbyMLSO3andMOPITTCO
assimilatedfields.AtmosphericChemistryandPhysics10,2175‐2194.
EPA’sAirDataBase,2008.http://www.epa.gov/air/data.
EPA,2009.TransferStandardsforCalibrationofAirMonitoringAnalyzers
forOzone.Technicalassistancedocument,November2010.www.epa.
gov/ttnamti1/files/ambient.
EPA(EnvironmentalProtectionAgency),1998.Guidelineondatahandling
conventionsforthe8‐hourozoneNAAQS.http://www.epa.gov/
ttncaaa1/t1/meta/m25245.html.
Finlayson‐Pitts,B.J.,Pitts,J.N.,2000.ChemistryoftheUpperandLower
Atmosphere:Theory,Experiments,andApplications,AcademicPress.
SanDiego,pp.583‐585.
Fuglestved,J.S.,Jonson,J.E.,Isaksen,I.S.A.,1994.Effectsofreductionsin
stratosphericozoneontroposphericchemistrythroughchangesin
photolysisrates.TellusB46,172‐192.
Gerasopoulos,E.,Kouvarakis,G.,Vrekoussis,M.,Donoussis,C.,
Mihalopoulos,N.,Kanakidou,M.,2006.Photochemicalozone
productionintheeasternMediterranean.AtmosphericEnvironment
40,3057‐3069.
Ghude,S.D.,Jain,S.L.,Arya,B.C.,Kulkarni,P.S.,Kumar,A.,Ahmed,N.,2006.
TemporalandspatialvariabilityofsurfaceozoneatDelhiand
Antarctica.InternationalJournalofClimatology26,2227‐2242.
HPC(HydrometeorologicalPredictionCenter),2007.Dailyweathermaps.
http://www.hpc.ncep.noaa.gov/dailywxmap/.
Itano,Y.,Bandow,H.,Takenaka,N.,Saitoh,Y.,Asayama,A.,Fukuyama,J.,
2007.ImpactofNOXreductiononlong‐termozonetrendsinanurban
atmosphere.ScienceoftheTotalEnvironment379,46‐55.
Jacob,D.J.,Logan,J.A.,Gardner,G.M.,Yevich,R.M.,Spivakovsky,C.M.,
Wofsy,S.C.,Sillman,S.,Prather,M.J.,1993.Factorsregulatingozone
overtheUnitedStatesanditsexporttotheglobalatmosphere.Journal
ofGeophysicalResearch–Atmospheres98,14817‐14826.
Jenkin,M.E.,2004a.Analysisofsourcesandpartitioningofoxidantinthe
UK–PartI:theNOX‐dependenceofannualmeanconcentrationsof
nitrogendioxideandozone.AtmosphericEnvironment38,5117‐5129.
Jenkin,M.E.,2004b.Analysisofsourcesandpartitioningofoxidantinthe
UK–PartII:contributionsofnitrogendioxideemissionsand
backgroundozoneatakerbsidelocationinLondon.Atmospheric
Environment38,5131‐5138.
Kleinman,L.I.,2005.Thedependenceoftroposphericozoneproduction
rateonozoneprecursors.AtmosphericEnvironment39,575‐586.
Mazzeo,N.A.,Venegas,L.E.,Choren,H.,2005.AnalysisofNO,NO2,O3and
NOxconcentrationsmeasuredatagreenareaofBuenosAiresCity
duringwintertime.AtmosphericEnvironment39,3055‐3068.
Nair,P.R.,Chand,D.,Lal,S.,Modh,K.S.,Naja,M.,Parameswaran,K.,
Ravindran,S.,Venkataramani,S.,2002.Temporalvariationsinsurface
ozoneatThumba(8.6°N,77°E)‐atropicalcoastalsiteinIndia.
AtmosphericEnvironment36,603‐610.
NCDC(NationalClimaticDataCenter),2008.ClimateMapsoftheUnited
States.http://cdo.ncdc.noaa.gov/climaps/skyc5007.pdf.
98Mioduszewskietal.–AtmosphericPollutionResearch2(2011)89‐98
NationalParkService(NPS),2010.Ozonemonitoringoverview–Ozone
pollutantmonitoringinthenationalparks.http://www.nature.nps.
gov/air/monitoring/network.cfm.
Platt,U.,1999.Modernmethodsofthemeasurementofatmospherictrace
gases.PhysicalChemistryChemicalPhysics1,5409‐5415.
Seinfeld,J.H.,2004.Airpollution:ahalfcenturyofprogress.AIChEJournal
50,1096‐1108.
USCBa(UnitedStatesCensusBureau),2008.http://www.census.gov/
population/www/estimates/metroarea.html.
USCBb(UnitedStatesCensusBureau),2008.http://factfinder.census.gov.
Vukovich,F.M.,Sherwell,J.,2003.Anexaminationoftherelationship
betweencertainmeteorologicalparametersandsurfaceozone
variationsintheBaltimore‐Washingtoncorridor.Atmospheric
Environment37,971‐981.
Vukovich,F.M.,1995.Regional‐scaleboundarylayerozonevariationsinthe
easternUnitedStatesandtheirassociationwithmeteorological
variations.AtmosphericEnvironment29,2259‐2273.
Walcek,C.J.,Yuan,H.H.,1995.Calculatedinfluenceoftemperature‐related
factorsonozoneformationratesinthelowertroposphere.Journalof
AppliedMeteorology34,1056‐1069.
Wang,Y.,Ridley,B.,Fried,A.,Cantrell,C.,Davis,D.,Chen,G.,Snow,J.,
Heikes,B.,Talbot,R.,Dibb,J.,Flocke,F.,Weinheimer,A.,Blake,N.,
Blake,D.,Shetter,R.,Lefer,B.,Atlas,E.,Coffey,M.,Walega,J.,Wert,
B.,2003.Springtimephotochemistryatnorthernmidandhigh
latitudes.JournalofGeophysicalResearch–Atmospheres108,art.no.
8358.
WashingtonStateDepartmentofTransportation(WSDOT),2010.WSDOT
GeodataDistributionCatalog.http://www.wsdot.wa.gov/mapsdata/
geodatacatalog/.