Measurements of spectral optical properties and their relation to biogeochemical variables and proce

CRATERLAKE,OREGON

Measurementsofspectralopticalproperties

andtheirrelationtobiogeochemicalvariablesand

processesinCraterLake,CraterLakeNationalPark,OR

EmmanuelS.BossÆRobertCollierÆ

GaryLarsonÆKatjaFennelÆW.S.Pegau

ÓSpringerScience+BusinessMediaB.V.2007

AbstractSpectralinherentopticalproperties(IOPs)havebeenmeasuredatCraterLake,OR,anextremelyclearsub-alpinelake.IndeedPurewaterIOPsaremajorcontributorstothetotalIOPs,andthustothecolorofthelake.VariationsinthespatialdistributionofIOPswereobservedinJuneandSeptember2001,andreflectbiogeo-chemicalprocessesinthelake.AbsorptionbycoloreddissolvedorganicmaterialincreaseswithdepthandbetweenJuneandSeptemberintheupper300m.Thispatternisconsistentwithanet

GuestEditors:GrayL.Larson,RobertCollier,andMarkW.Buktenica

Long-termLimnologicalResearchandMonitoringatCraterLake,Oregon.

E.S.Boss(&)

UniversityofMaine,5741LibbyHall,Orono,ME04469,USA

e-mail:emmanuel.boss@maine.edu

R.CollierÆW.S.Pegau

OregonStateUniversity,COAS,104Ocean.Admin.Bldg.,Corvallis,OR97331,USA

G.Larson

USGSForestandRangelandEcosystemScienceCenter,3200JeffersonWay,Corvallis,OR97331,USA

K.Fennel

IMCS,RutgersUniversity,71DudleyRoad,NewBrunswick,NJ001,USA

releaseofdissolvedorganicmaterialsfrompri-maryandsecondaryproductionthroughthesummeranditsphoto-oxidationnearthesurface.Watersfedbyatributarynearthelake’srimexhibitedlowlevelsofabsorptionbydissolvedorganicmaterials.Scatteringismostlydominatedbyorganicparticulatematerial,thoughinorganicmaterialisfoundtoenterthelakefromtherimfollowingarainstorm.Severalsimilaritiestooceanicoligotrophicregionsareobserved:(a)TheBeamattenuationcorrelateswellwithpar-ticulateorganicmaterial(POM)andtherela-tionshipissimilartothatobservedintheopenocean.(b)Thespecificabsorptionofcoloreddissolvedorganicmaterialhasavaluesimilartothatofopenoceanhumicmaterial.(c)Thedis-tributionofchlorophyllwithdepthdoesnotfol-lowthedistributionofparticulateorganicmaterialduetophoto-acclimationresultinginasubsurfacepigmentmaximumlocatedabout50mbelowthePOMmaximum.

KeywordsCraterLakeÆOpticsÆ

BiogeochemistryÆBackscatteringcoefficient

Introduction

Novelcommercialin-situspectralopticalinstru-mentation,developedinthepastdecade,hasopenedoceansandlakestoexplorationof

123

150biogeochemicalprocessesatsub-1mscales,scales,whichhavenotbeenaccessiblepreviously.Fromadatapoorfield,withonlyahandfulofabsorptionspectrameasuredperwatercolumn,hydrologicalopticshasbecomeadatarichfieldwithspectrageneratedatsub-Hzandthusatsub-1mresolution.Untiltherecentdevelopmentofacommercialin-situspectralabsorptionandattenuationsensor(Mooreetal.,1997)andbackscatteringsensors(Maffione&Dana,1997;Zaneveldetal.,2003),absorptionandbackscat-teringwerenotmeasuredinhighresolution,routinely,andin-situ.

Theabsorptionandbackscatteringcoefficientsareimportantinherentopticalproperties,inparticularsincetheyaretightlylinkedtothecolorofawaterbody(e.g.,Gordonetal.,1988).Thelinkbetweenthecolorofawaterbodyanditsinherentopticalproperties(propertiesnotaffectedbytheilluminationconditionssuchasabsorption,scattering,andattenuation,hereafterIOP)isprovidedbytheradiative-transferequa-tion(e.g.,Mobley,1994);knowledgeoftheIOPandtheilluminationconditionsissufficienttopredictthecolorofthelakethatwillbeperceivedbyanobserver.

IOPareroutinelyinvertedtoprovideinfor-mationabouttheconcentrationandnatureoftheparticulatesanddissolvedsubstancesinaquaticenvironments.InversionsofIOPhavebeenusedtoestimatethevolumeoftotalsuspendedmatterfrombeamattenuation(Spinrad&Zaneveld,1982),particulateorganicmatterconcentrationfrombeamattenuation(Bishop,1999),dissolvedorganicmatterconcentrationfromabsorptionbydissolvedmaterial(e.g.,reviewbyBlough&Green,1995),chlorophyllconcentrationfromfluorescenceorparticulateabsorption,particulatesizedistributionfromspectralparticulatebeamattenuation(Bossetal.,2001a),andbulkpartic-ulatecompositionfromtheparticulatebackscat-teringratio(Twardowskietal.,2001;Bossetal.,2004).

MeasurementsofIOPinlakeshavemostlybeenlimitedtosingle-beamtransmissometersandland-basedspectrophotometry(e.g.,reviewbyBukataetal.,1995).InthispaperwepresentdataofspectralIOPmeasuredin-situinCraterLake,OR.

123

Hydrobiologia(2007)574:149–159

CraterLakeisasub-alpinelake(Altitude1,883m),5mdeep(deepestintheUS),andenclosedbyavolcano’scalderawitharadiusofnearly3kmatthelake’ssurface(Klimasauskasetal.,2002).CraterLakehasbeenanaturallaboratoryforhydrologicalopticsfornearly70years.Pettit(1936)performedmeasurementsofspectralbackscatteringthereandfoundthemtobecomparabletodistilledwater.Smith&Baker(1981)useditaspartoftheirdatasettodeterminetheabsorptionoftheclearestnaturalwaters.InthispaperwepresentthefirstspectralIOPmeasurementsinCraterLake.TheclarityofCraterLakeprovidesachallengetothenovelspectralIOPinstrumentation,sincethesignalisveryweak.IfthisinstrumentationisfoundtoworksatisfactorilyatCraterLake,itislikelytoperformwellinmostnaturalconditions(assum-ingthatthepath-lengthisreducedinveryturbidwaters).

MaterialandmethodsSampling

Datawerecollectedduringtwosamplingperiods.InJune2001wesampledfollowingarainstorm.Wesampledattwostationsinthelake.Sta.13,overthedeepestareaofthelake(42°56¢N,122°06¢W),wassampledontwoconsecutivedays(June28–29)downto300mdepth(asubsetofthesedatawasanalyzedinFennel&Boss,2003).OnJune28wealsosampledashallow(6m)stationwhereaturbidwhitishstreamofwaterflowedintothelakefromthecalderawalls.WereturnedtoSta.13onSeptember19thandsam-pledtowithin20mofthelake’sbottom(596m).Datacollectionandmethodofanalysis

Spectralabsorptionandattenuation,hydro-graphicpropertiesandspectralbackscatteringweremeasuredonawinchedpackagewithasingleWETLabs’ac-9,aSeaBirdSBE25andaHobiLabsHydroscat-6(HS-6),respectively.Additionally,aWETLabs’Eco-VSFwasusedtomeasurebackscatteringinSeptember.

Hydrobiologia(2007)574:149–159ChlorophyllfluorescencewasmeasuredwithaWETLabsWetStar.Ateachstationtworepeatedcastsweretaken,oneofwhichhad0.2lmfilters(GelmanSuporcap100)oneattachedtoeachintakesoftheabsorption,andattenuationtubesoftheac-9forthemeasurementsofCDMabsorption(ColoredDissolvedMaterial,opera-tionallymaterialthatgoesthrougha0.2lmfilter).InSeptemberwealsouseda1lmfilter(WhatmanCapsule)ateachintakeoftheac-9.Newfilterswereusedforeachfieldcampaign.Datawerebinnedto1mbinsbycomputingthemedianofthepropertiesinthatbin.Partic-ulatepropertieswerecomputedbysubtractingthebinnedmeasurementinaverticalprofileofspectralabsorptionandattenuationwiththe0.2lmpre-filterfromaprofileperformedwithoutafilter.Someerrorsmaybecausedbythetem-poraldeparturebetweenthetwomeasurements(approximately30min),giventhepresenceofinternalwavesconfirmedintemperatureprofiles.However,thesedeparturesarelikelytoresultinsmalluncertaintiesinthederivedparticulateabsorptionandattenuationduetotheslowlyvaryingprofileofCDMwithdepth(seebelow).Notethatdifferencesbetweentwosuccessivemeasurements,withandwithoutfilter,resultinaparticulateprofilethatisindependentoftheinstrumentcalibration(assumingwesamplethesamewaters).Thusthelikelyerrorsofthemea-surementsofabsorptionandattenuationarelimitedbytheprecisionoftheinstrument(±0.001m–1)andtheenvironmentalvariabilityandnottheaccuracyofthecalibration(per-formedwithpurewaterbeforeandafterthecruiseattheauthor’slabfollowingthemethodofTwardowskietal.,1999).Notethatanadditionaluncertaintymaybecausedbychangesinfilterefficiencywithuse.ConsistencyofCDMvaluesbetweenconsecutiveprofiles(notshown)sug-gestedthatthiswasnotasignificantproblemduringthedeployments.

BackscatteringbyparticleswiththeHS-6(measuresscatteringatoneangle,centeredaboutascatteringangleof140°)wascomputedusingthemethodoutlinedinBoss&Pegau(2001).Thevolumescatteringfunctionmeasured(ofwaterandparticles)wasmultipliedbyafactorof1.12toaccountforacorrectioninthecalibrationfactor

151

duetoSpectralonreflectivitythathasbeendetectedinDec.2002(D.Dana,personalcom-munication,2004).ApplicationofthisfactorresultedinanimprovedagreementbetweenthedifferentwaysofestimatingthebackscatteringcoefficientthanachievedinBossetal.(2004)(seebelow).TheEco-VSF(measuresscatteringatthreeanglesinthebackdirection,centeredat100,125,and150°)wasprocessedfollowingZaneveldetal.(2003).WehadproblemswithtwowavelengthsoftheHS-6;the555-channelwasobservedtoshiftvalueswithincastsinamanneruncorrelatedwiththeotherwavelengthsandtohavethelargeststandarddeviationduringcali-bration.The676nmchannelwashighlycorre-latedwithchlorophyllconcentration(probablyduetoitsrecordingofsomechlorophyllfluores-cenceexcitedat676andemittedat681nm),unliketheremainingfourchannels.Wethereforepresentonlyfourchannels(440,488,530,620nm)oftheHS-6data(Fig.1).

Itisinterestingtonotethatthecontributionofmolecularbackscatteringbypurewatertothetotalbackscatteringcoefficientinourmeasure-ments(thesumofwaterandparticulateback-scattering)variesfromnearly60%to30%

0-50-100b]bw(620nm)m[ hHS 440tp-150HS 488edHS 530HS 620-200Eco 530bbw(488nm)b-250bw(440nm)bbw(530nm)-30000.511.522.533.5Backscattering by particles [m-1] x1000 Fig.1Verticaldistributionoftheparticulatebackscatter-ingcoefficientobtainedatfourwavelengths(HS-6)andatasinglewavelength(Eco-VSF)onSep.19,2001atSt.13.Thestraightlines,denotesthevaluesofbackscatteringbypurewaterbasedonMorel(1974).Theregressionbetweenthebackscatteringcoefficientsat530nmbetweenthemeasurementsbythetwobackscatteringinstrumentsisbbp(530,EcoVSF)=0.97bbp(530,HS6),andthecorrela-tioncoefficientbetweenthemis0.97.ThespikeintheHS-6(530nm)couldnotbeexplained123

152betweenblueandredwavelengthsatdepth(Fig.1).Inthecalculationofparticulateback-scatteringthecontributionfromwaterwassub-tractedoutofthetotalbackscatteringassumingthebackscatteringbywatertobeconstantandequaltothevaluesofMorel(1974).Theremainingbackscatteringisattributedtoparticlesasitiscustomary,toassumeanegligiblecontri-butionofCDMtoscatteringandbackscattering(e.g.,Mobley,1994).

Chlorophyllconcentrationwasestimatedinthreeways:1.

Chlorophyllfluorescence(WETLabsWet-Star)calibratedagainstchlorophyllestimatesfromdiscretesamplesanalyzedwithHPLC.2.

Theline-heightofthechlorophyllabsorp-tionpeakinthered:([chl]={ap(676)–(39/65Æap(650)+26/65Æap(715))}/0.014),e.g.,Davisetal.(1997),whereapdenotespar-ticulateabsorption,andthewavelengthisgiveninbrackets.Thevalueofthechlorophyll-specificabsorption,a*(676)=0.014m2gchl–1usedhereisconsistentwithpublishedliteraturevaluesfortheoceanicassemblagesofphytoplankton(e.g.,Sosik&Mitchell,1995)andwaschosenherebasedonregressionofHPLCdetermina-

0-100]m-200[ htpe-300D-400-500-6000.100.10.20.30.40.50.60.70.8chlorophyll [mg/m3]Fig.2Comparisonbetweenthreeestimatesofchloro-phyll;WetStarfluorometer(boldblackline),chlorophyllestimatedfromparticulateabsorptionlineheight(black),andHPLC(asterisks)basedonmeasurementsatSt.13,Sep.19,2001.Notethat,unlikefluorescence,theabsorp-tionestimateofchlorophyllconcentrationdoesnotsufferfromnon-photochemicalquenchingnearthesurface123

Hydrobiologia(2007)574:149–159

tionof[chl]versusabsorptionbased[chl].a*(676)islikelytovarywithdepth,aspackagingwithincellschangesitsvalue.Publishedvaluesofa*(676)varyfrom0.008m2gchl–1to0.023m2gchl–1(Bri-caudetal.,1995;Sosik&Mitchell,1995).3.

DiscretechlorophyllconcentrationobtainedwithHPLC.

ThesquaredcorrelationcoefficientbetweenthechlorophyllfluorescenceandabsorptionisR2=0.83(Fig.2,N=570).Theabsorptionesti-mated[chl]hastheadvantageofnotsufferingfromnear-surfacenon-photochemicalquenchingwhichreducesthefluorescence-yieldofphyto-planktonexposedtohighlightnearthesurface.Theuncertaintyfortheabsorption-estimated[chl]is±0.2lgl–1(~mgm–3)basedontheuncertaintyinac-9measurements(seebelow,andnottakingintoaccountthepossiblebiasina*whichislikelytoaddanduncertaintyof±30%).Negativevalues(~–0.05lgl–1)inabsorptionbasedchlorophyllestimatesobservedinFig.2areindicativeoftoobigaremovalofthedetritalabsorptionfromtheparticulateabsorptioninthecalculationofphytoplanktonabsorptionat676nm.Notethatthesevaluesarewellbelowtheuncertaintyintheabsorption-basedchlorophyll.Slopesoftheparticulatesizedistribution(the‘‘PSDslope’’)wereestimatedfromtheattenuationspectrumbyfittingahyperbolicfunctiontotheparticulateattenuationspectrum(e.g.,Bossetal.,2001a).ThismethodprovidesanestimateforthebroadPSDslopeassumingahyperbolicparticulatePSD(seeBossetal.,2001b,foracriticalanalysisoftheassumptions).Thismethodprovidesanesti-mateofthespatialchangesinthemeanparticlesize.Wefittedasimilarfunctiontotheparticulatebackscatteringspectrumtoassesswhetherthereisanyrelationbetweenthisspectrumandthatofparticulatebeamattenuation,andfoundnosig-nificantrelationbetweenthetwo(notshown).Iftheparticleswerenon-absorbingandthemea-surementswereerrorfreewewouldhaveexpectedtheslopestobesimilarbasedonMietheory(Morel,1973).

Bulkparticulateindexofrefractionwasesti-matedusingthemethodofTwardowskietal.

Hydrobiologia(2007)574:149–159(2001).Thebulkindexofrefractionprovidesaninsightintotheparticulatecomposition;phyto-planktonanddetritustendtohavealowindexofrefraction(n=1.02–1.1)duetoalargewaterfractionwhileinorganicparticleshavealargerin-dexofrefraction(n=1.12–1.24)(seeTwardowskietal.,2001;Bossetal.,2004,fordetaileddiscus-sion).Twardowskietal.(2001)developedamethodtoobtaininformationonthebulkindexofrefractionfromknowledgeoftheparticulatesizedistributionandtheparticulatebackscatteringra-tio(theratioofparticulatebackscatteringtopar-ticulatescattering).Assumingtheparticulatesizedistribution(PSD)tobehyperbolic,arelationshiphasbeenfoundbetweenthehyperbolicslopeofthePSDandthespectralslopeofparticulatebeamattenuation(e.g.,Bossetal.,2001a,andreferencestherein).

Discretesamplesforchlorophyllandparticulateorganiccarbon(POC)werecollectedinSeptem-berandprocessedwiththesamemethodsasinUrbachetal.(2001).Inaddition,cellswerecol-lectedinSeptemberforFlow-cytometricanalysis.Uncertaintiesinmeasurements

Whilewithcalibration,theerroroftheac-9areassumedtobeO(0.005m–1)(Twardowskietal.,1999)weestimatetheparticulateabsorptionandattenuationmeasurementpresentedheretobeaccuratetowithinO(0.002m–1).Thisisbecausethevaluesfortheparticulateattenuationandabsorptionweremeasuredusingasingleinstru-mentdeployedwithandwithoutapre-filter.Assumingthein-waterpropertiestostayconstantbetweenthetwoconsecutiveprofiles(30min)theuncertaintyisonlylimitedbytheinstrumentpre-cision(0.001m–1)andstabilityoftheinstrumentmeasuringthesamewatersO(0.002m–1).Notethatchangesinfilterefficiencywithtimewouldincreasetheseuncertainties.Asnotedabove,wehavenotnoticedchangesinCDMconcentrationatdepthbetweenconsecutiveprofiles,indicatingnonoticeablechangeinfilterperformance.

Note,however,thattothisdatethereisalargeuncertaintyintheparticulateabsorptionintheblueregionofthespectrumduetoissuesassoci-153

atedwiththescatteringcorrectionofparticulateabsorption.Differentmethodsofthe‘‘Scatteringcorrection’’(e.g.,Zaneveldetal.,1994)deviatebyupto20%intheirestimateofabsorptionintheblue,whileconvergingtowardstheredpartofthespectrum.Thescatteringcorrectionmethodusedherewasmethod3ofZaneveldetal.(1994),inwhichafixedportionofscatteringisremovedfromthemeasuredabsorption.Theproportionofscatteringthatisremovedequalsthemeasuredratioofabsorptionandscatteringat715nm,consistentwithassumingzeroabsorptionbypar-ticlesat715nm.Uncertaintyincoloreddissolvedmaterialabsorption(CDM)is0.005m–1,asizableuncertaintyrelativetotheabsoluteCDMabsorptionmeasuredinCraterLake.Toreducethisuncertaintythetwoindependentmeasure-mentsofCDMabsorptionobtainedbythea-andc-sidesoftheac-9wereaveraged,reducingtheuncertaintyto0.0035m–1.

Backscatteringerrorsareassumedtobesmallerthan15%ofthesignalbasedonuncer-taintiesinconvertinglightbackscatteredat140tothebackscatteringcoefficient(Maffione&Dana,1997;Bossetal.,2004).Anadditionaluncertaintybasedonvariabilityintheinstruments’calibra-tionhistoryis±0.0002m–1(D.Dana,2004,per-sonalcommunication).Comparisonofparticulatebackscatteringestimatesat530nmbasedonthetwobackscatteringinstrumentsrevealsthatthetwoarehighlycorrelated(R2=0.94,N=298),withtheEco-VSFbeingonaverage0.97timestheHS-6andwithanoffsetthatisnotsignifi-cantlydifferentfromzero(Fig.1).Thisisanexcellentagreement,inparticulargiventhatthetwoinstrumentsarecalibrateddifferentlyandthatthebackscatteringcoefficientiscomputeddifferentlyfromeachmeasurement(Zaneveldetal.,2003).Italsoindicatesthattheuncertain-tiesquotedaboveareprobablytooconservative.Uncertaintiesintheparticulatebackscatteringratio,theratioofbackscatteringtototalscattering,arelikelytobelessthan20%,basedonpropagatingtheerrorsofscatteringandbackscattering.Uncer-taintiesinthediscretemeasurementswereesti-matedfromreplicatesamplesandare0.05mgl–1forchlorophyllandnearly30%forPOC.

123

154

Fig.3Measurementsoftemperature,particulateattenuationat650nm,andchlorophyll(absorptionbased)asfunctionofdepth(toprow)andofparticulateattenuationat650nm,andchlorophyllasfunctionoftemperature(bottomrow).MeasurementsweretakenatSt.13onJune28(thin)andJune29(bold),20010-50Hydrobiologia(2007)574:149–159

Depth [m]-100-150-200-25046800.10.200.511.5Temperature98cp(650) [m-1]987654Chlorophyll [mg/m3]Temperature7654300.050.10.150.2300.51cp(650) [m-1]Chlorophyll [mg/m3]Results

Spatialandtemporalvariabilityinparticulateproperties

TheprofilesofparticulatepropertiesinJunevarystronglybetweenthe2daysofsampling(Figs.3,4).Thisvariabilityismostlyduetotheactivityoflargeinternalwaves(amplitude>20m)setupbyastormthatpassedbyadaypriortosampling.Whenplottedagainsttemperature,bothchloro-phyllandbeamattenuationareverysimilarforbothsamplingdays,indicatingthe(mostly)con-servativebehaviorofthesepropertiesinthepresenceofinternalwavesoverasingleday.ParticulatepropertiesvaryinthewatercolumnasfunctionofdepthanddifferbetweentheJuneandSeptembersamplingperiods(Figs.4,5).TheredoesnotseemtobearelationbetweenparticulatepropertiesandstratificationexceptfortheelevatedphytoplanktonbiomassnearthesurfaceinSeptember.TheverticaldistributionofthebackscatteringcoefficientissimilartothebeamattenuationinSeptemberbutnotinJune.Theresultingchangeinthebackscatteringratio

withdepthinJune,suggestsachangeinthebulkparticulatecompositionwithdepth.Beam-candPOC

Whenregressingthediscreteparticulateorganiccarbonmeasurements(POC,mmolCm–3)andtheparticulatebeamattenuationmeasurements(cp,m–1,extrapolatedto660frommeasurementsat650and676)wefind:

cpð660Þ¼POCÃ0:032À0:024ðR2¼0:996Þ:ThisregressionissimilartothatfoundbyGun-dersonetal.(1998)fortheArabianSea(0.031and–0.007,forslopeandintercept,respectively).Whilethecorrelationcoefficientishigh,theconfidenceintheslopeandinterceptislowduetothe30%differencebetweenreplicatemeasure-mentsofPOC(weusedthemeanofthereplicatestoderivetheregression).Giventhatlargeuncertainty,theregressionisalsoconsistentwithdatafromStationAlohanearHawaii(Fennel&Boss,2003)andtheSouthernocean(Gardner

123

Hydrobiologia(2007)574:149–159Fig.4Measurementsoftemperature(4casts),chlorophyll(absorptionbased),particulatebackscattering,particulatebeamattenuation,backscatteringratio,andthespectralslopeofbeamattenuationatSt.13onJune28(thin)andJune29(bold),2001.TwoofthefourtemperaturecastsandthechlorophyllandbeamattenuationdataarethesameasthoseinFig.3155

0-50Depth [m]-100-150-200-250468Temperature [C]00.511.5Chlorophyll [mg/m3]1.5bback,particles(530) x 1000 [m-1]0.510-50Depth [m]-100-150-200-25000.10.2cparticles(650) [m-1]0.0050.010.0150.02bback,particles/bparticles(530)0.511.5Attenuation spectral slopeetal.,2000)wheretheslopeswerefoundtobeO(0.02[m–1(mmolCm–3)–1]).Chlorophyllversusbiomass

Inbothsamplingperiodsweobservedchlorophyllandparticulateattenuationtobedecoupled.Similardecouplingbetweenphytoplanktonbio-volumeandchlorophyllhasbeenobservedinpastmeasurementsinCraterLake(McIntireetal.,1996:Fig.5).Thedecouplingisduetophyto-planktonphoto-acclimation,whichischaracter-izedbyatypicalnearlyexponentialincreaseinthechlorophyll/cpratioandtheratioofchloro-phyll/phytoplankton-bio-volumefromthesurfacedowntothechlorophyllmaximum(e.g.,Kitchen&Zaneveld,1990;Fennel&Boss,2003:Fig.7).ThisrelationshipisdiscussedinmoredepthinFennel&Boss(2003),andillustratestheproblemassociatedwithusingchlorophyllasaproxyphytoplanktonbiomasswhenanalyzingverticalprofiles.

Contributionofsub-micronparticles

Sub-micronparticlesarefoundtocontributenearly30%ofchlorophyllatthechlorophyllmaximuminSeptember(Fig.5),thoughtheircontributiontoabsorptionorattenuationisless

than20%.Flow-cytometricanalysisofsamplesfromthelake(Sherr&Sherr,personalcommu-nication,2002)indicatesthatthephytoplanktonsampledonthesamedayareneitherSynecho-coccusnorProchlorococcus,thedominatingsub-micronnano-phytoplanktonintheoceans(basedonfluorescenceandscatteringcharacteristics).Wecurrentlydonotknowthetaxonomyofthesesub-micronparticlesthatareobservedtocon-tributesignificantlytotheopticalpropertiesofCrateLake.McIntireetal.(1996)observedthepresenceofseveralspeciesofpicoplanktonwithdiametersontheorderof1lmbutcouldnotidentifymicroscopicallysmallercells.Sizedistribution

ChangesintheslopeofthebeamattenuationbetweenJuneandSeptemberareconsistentwith,ingeneral,smallerparticlesizeinSeptemberthaninJune,exceptrightnexttothesurface.ThisisconsistentwiththesurfacewatersatCraterLakebeingregularlydominatedbytherelativelylargeandelongated(about70lmlongandafewmi-cronwide)diatomNitzschiagracilisinAugustandSeptember(McIntireetal.,1996).Inbothsamplingseasons,thechlorophyllmaximumregionisdominatedbyparticleswithasimilarsizedistribution(cpslope~1).

123

156

0-100Depth [m]-200-300-400-500510Temperature [C]0-100Depth [m]-200-300-400-50000.10.2cparticles(650)[m-1]0.010.02bback,particles/bparticles(530)00.20.40.6Chlorophyll [mg/m3]Hydrobiologia(2007)574:149–159

2bback,particles(530) x 1000[m-1]010.511.5Attenuation spectral slopeFig.5Measurementsoftemperature,chlorophyll(absorptionbased),particulatebackscattering,particulatebeamattenuation,particulatebackscatteringratio,andthespectralslopeofparticulatebeamattenuationatSt.13onSep19,2001.Thethinlinerepresentstotalparticulatemeasurementswhiletheboldlinerepresentsthefractionsmallerthan1lm.Starsrepresentdiscretechlorophyll-a(HPLC)andPOCmeasurements.HydroScat-6backscat-teringisdenotedbyablacklineandextendsonlyto300m.Eco-VSFbackscatteringisdenotedingrayParticulatecompositionandtheirbulkindexofrefraction

SamplinginJunenearthecaldera’swall,inwatersfedtothelakebyatributary,revealsalevelofparticulateattenuationnearlyfivetimeshigherthanthatfoundnearthesurfaceatSt.13.Thesewatersaredominatedbyparticulatematerial(Fig.6)witharatioofparticulatescatteringtoparticulateattenuationgreaterthan0.93.Thismaterialhasabackscatteringratiogreaterthan0.017andabeamattenuationslopeofnearly1.2.BasedontheanalysisofTwar-dowskietal.(2001;Fig.7here)thisislikelytobeinorganicmaterial,characteristicofclayminerals.ThismaterialislikelycontributingtotherelativelyhighbackscatteringratioobservedforthesurfacewatersandthewatersbelowthechlorophyllmaximaatSt.13(Fig.7).At50m,however,thewaterofSt.13hasabackscat-teringratiocharacteristicoforganicmaterialsuchasphytoplankton(0.005).

ThebulkindexofrefractionatSt.13inSep-temberisfoundtobelow(<1.1)consistentwithorganic,phytoplankton,anddetritus(Fig.7).Theparticleswiththehighestindexofrefractionarelocatedbelow200m.Thismaybeduetoamixofdetritalparticles(e.g.,inorganicphytoplanktonshells,smallinorganicparticles)andheterotro-phicorganisms.ThevariabilityinestimatedPSDslopeandindexofrefractionthefullwatercol-umnofCraterLakeisindicativeofthedifferentassemblagesofparticlesthatoccupydifferentdepthsofthelake.

ColoreddissolvedmaterialandDOM

Absorptionbydissolvedmaterialat440nmin-creasesmonotonicallywithdepthinJunewhilehavingasubsurfacemaximumat130minSep-tember(Fig.8).BetweenJuneandSeptemberitincreasesmostlybetween100mand300m,belowthelayerofmaximumattenuation(maxi-mumPOCandphytoplanktonbiomass),andin

123

Hydrobiologia(2007)574:149–159Fig.6Spectraof

particulateattenuation(cp),particulatescattering(bp),particulate

absorption(ap),dissolvedabsorption(aCDM)andparticulatebackscattering(bbp,multipliedby50)averagedover1matastationneartheedgeofthecalderawhereastreamenteredintothelakeonJune28,2001(blacklines).Highestvaluesweremeasuredclosesttothebottomanddecreasedmonotonicallytowardsthesurface.Graylinesdenotethesamepropertiesintheupper10matthecenterofthelake(st.13)

157

bback,particles x 1000[m-1]252

151050

cparticles [m-1]aparticles [m-1]500

600

700

20

1.510.500.10.080.060.040.020400

500

600

700

0.02

adissolved [m-1]0.010

-0.01

400

wavelength [nm-1]wavelength [nm-1]

theareaofmaximumparticulategradient.Theobservedaccumulationisconsistentwithareleaseofdissolvedorganicmaterialasaby-productofprimaryandsecondaryproductionwithinthewatercolumnandnearthebottom(whereaslightincreaseinCDMisalsoobserved).ThedissolvedabsorptionmeasurednearthetributaryinJuneissimilarinmagnitudetothatatthecenterofthelake(Fig.6),implyingthatnosignificantcoloreddissolvedmaterial(andbyconjunctionnosignificantamountofdissolvedmaterial)isinputintothelakefromthecaldera’srimatthatlocation.Unfortunately,largeuncer-taintiesinCDMvalues(±0.0035m–1)didnotal-lowustocomputeareliableCDMspectralslope(e.g.,spectrainFig.6).

Pastmeasurementsoftotalorganiccarbon(TOC)atCraterLakesuggestabackgroundvalueofabout0.1gCm–3,dominatedbydis-solvedorganiccarbon(DOC)(Urbachetal.,2001;Hargreavesetal.,2007).ThismagnitudeofDOCisatleastafactoroffivelowerthanopenoceanvalues(comparedtovaluesfoundintheBermudaAtlanticandtheHawaiianOceantimeseries).Foranabsorptionvalueof

–10.01±0.003mat440nm(Fig.8)itimpliesaDOCspecificabsorptioncoefficientof

0.001±0.0005m2(gC)–1.ThisvalueoftheCDM

specificabsorptionisverylowandsimilartovaluesmeasuredformarinehumicacid(Blough&Green,1995).

0.030.025Backscattering ratio [bbp/bp]n=1.14 0.020.015n=1.1 0.01n=1.06 n=1.02 0.00503.63.844.24.44..85Particulate size distribution slope ~ 3 + cp-slopeFig.7Diagramdepictingthebackscattering-ratioagainsttheestimatedslopeoftheparticulatesizedistribution(computedas3+spectralslopeofcp)forprofiledatacollectedatSt.13,Sep.192001.Contoursrepresentlinesofequalindexofrefraction(n)basedonMietheory(seeTwardowskietal.,2001;Bossetal.,2004).The+-signdenotesthebackscatteringratiobasedonmeasurementswiththeHS-6fortheupper300m,whilethedotsdenotethebackscatteringratiobasedontheEco-VSF,bothat530nm.Lowvaluesofthebulkindexofrefraction(1.02–1.1)implydominancebyorganicmaterial

123

158

0

-100

-200

]m[ ht-300

Sep 19, 2000peJune 26, 27, 2000

D-400

-500

-600

00.0050.010.015

CDM absorption [m-1]

Fig.8Distributionofabsorptionbycoloreddissolvedmaterialat440nminJune(thin)andSeptember(bold)2001.TheJuneprofileisanaverageoftheprofilesmeasuredonbothdays(28and29June).NotethattheuncertaintyintheCDMmeasurementis0.0035m–1,whichislargecomparedtomeasurement(seeMethodssection)

Discussion

WefindmodernopticalsensorstoperformwelldespitetheclarityofCraterLake.Thesesensorsprovideuswithhighverticalresolutionmeasure-mentsofopticalproxiesofbiogeochemicalvari-ables,whichcannotbeachievedusingstandardtechniques.

Thepicturethatemergesfromthemeasure-mentsisthatthewatersofCarterLakearehighlystratifiedopticallyinJuneandSeptemberatdepthswheredensitystratificationisweak.Largeampli-tudeinternalwaveswereobservedandmaypro-videanimportantmechanismforbringingnutrientsclosertothesurface.Photo-acclimationinphytoplanktoncausesadecorrelationoftheparticulatebeamattenuation(andalsoPOCandphytoplanktonbio-volume)andchlorophyllcon-centration;Chlorophyllexhibitsamaximum40–50mdeeperthanthatofPOC.

Interestingly,opticalestimatesandverticaldis-tributionofPOC,andchlorophyllcomparewellwiththerelationshipsanddistributionsobservedinopenoceanenvironments.Inaddition,thespecificabsorptionofCDMisfoundcomparabletothatofoceanichumicmaterials.Thesesimilaritiessuggestthatbiogeochemicallakestudiesmayberelevantforunderstandingoceanicprocesses.

123

Hydrobiologia(2007)574:149–159

Sub-micronparticlesarefoundtocontributesignificantlytothechlorophyllconcentration.TothisdatewedonotknowwhataretheorganismsdominatingthisfractioninCraterLake(McIntire,2002,personalcommunication).CDMabsorption,aproxyfordissolvedorganicmaterial,isfoundtoincreasefromspringtofall,mostlikelyaby-productofprimaryandsecondaryproduction.Theinstrumentswehaveusedforthisstudylendthemselvestoautonomousdeploymentonmoor-ings.WestronglyencouragesuchdeploymentinCraterLakeandsimilarenvironmentstoobtainmeasurementsofbiogeochemicalproxieswithhightemporalresolution.Suchdeploymentswillallowthestudyofthelakeresponsetoforcingvaryingfromepisodiceventstoclimatechange.Real-timebroadcastofthedatacanbeusedforadaptivesampling,wherefieldsamplingisdrivenbyobservedchangesinthelakehydrographicandopticalproperties.

AcknowledgementsWethankF.Baratangeforassis-tanceinthefieldandlab.ThankstoL.Eisner,F.Prahl,E.andB.Sherrfortheanalysisofthediscretesamples.CommentsbyT.Swiftandtwoanonymousreviewersaregratefullyacknowledged.ThisworkasbeenfundedbytheUnitedStatesGeologicalSurvey.TheinstrumentationusedwaspurchasedwithfundingbyONRandNASAtoS.PegauandE.Boss.

References

Bishop,J.K.B.,1999.Transmissometermeasurementof

POC.Deep-SeaResearchI46:353–369.

Blough,N.V.&S.A.Green,1995.Spectroscopicchar-acterizationandremotesensingofnonlivingorganicmatter.InZeppR.G.&C.Sonntag(eds),TheRoleofNon-livingOrganicMatterintheEarth’sCarbonCycle.Willey,Chichester:23–45.

Boss,E.&W.S.Pegau,2001.Therelationshipoflight

scatteringatanangleinthebackwarddirectiontothebackscatteringcoefficient.AppliedOptics40:5503–5507.

Boss,E.,W.S.Pegau,W.D.Gardner,J.R.V.Zaneveld,

A.H.Barnard,M.S.Twardowski,G.C.Chang&T.D.Dickey,2001a.Thespectralparticulateattenuationandparticlesizedistributioninthebottomboundarylayerofacontinentalshelf.JournalofGeophysicalResearch106:9509–9516.

Boss,E.,M.S.Twardowski&S.Herring,2001b.The

shapeoftheparticulatebeamattenuationspectrumanditsrelatontothesizedistributionofoceanicparticles.AppliedOptics40:4885–43.

Hydrobiologia(2007)574:149–159

Boss,E.,W.S.Pegau,M.Lee,M.S.Twardowski,E.Shy-banov,G.Korotaev&F.Baratange,2004.Thepar-ticulatebackscatteringratioatLEO15anditsusetostudyparticlescompositionanddistribution.JournalofGeophysicalResearch109,C0101410.1029/2002JC001514.

Bricaud,A.,M.Babin,A.Morel&H.Claustre,1995.

Variabilityinthechlorophyll-specificabsorptioncoefficientsofnaturalphytoplankton:analysisandparameterization.JournalofGeophysicalResearch(C–Oceans)100:13,321–13,332.

Bukata,R.P.,J.H.Jerome,K.Y.Kondratyev&D.V.

Pozdnyakov,1995.OpticalPropertiesandRemoteSensingofInlandandCoastalWaters.CRCPress,BocaRaton,Florida.

Davis,R.F.,C.C.Moore,J.R.V.Zaneveld&J.M.Napp,

1997.Reducingtheeffectsoffoulingonchlorophyllestimatesderivedfromlong-termdeploymentsofopticalinstruments.JournalofGeophysicalResearch102:5851–5855.

Fennel,K.&E.Boss,2003.Subsurfacemaximaofphy-toplanktonandchlorophyll—steadystatesolutionsfromasimplemodel.LimnologyandOceanography48:1521–1534.

Gardner,W.D.,M.J.Richardson&W.O.Smith,2000.

SeasonalpatternsofwatercolumnparticulateorganiccarbonandfluxesintheRossSea,Antarctica.Deep-SeaResearchII47:3423–3449.

Gordon,H.R.,O.B.Brown,R.E.Evans,J.W.Brown,

R.C.Smith,K.C.Baker&D.C.Clark,1988.Asemianalyticmodelofoceancolor.JournalofGeo-physicalResearch96:10909–10924.

Gunderson,J.S.,W.D.Gardner,M.J.Richardson&I.D.

Walsh,1998.EffectsofmonsoonsontheseasonalandspatialdistributionofPOCandchlorophyllintheArabianSea.Deep-SeaResearchII45:2103–2132.Hargreaves,B.R.,S.F.Gardiner,M.W.Buktenica,R.W.

Collier,E.Urbach&G.L.Larson,2007.Ultravioletradiationandbio-opticsinCraterLake,Oregon.Hydrobiologia574:107–140.

Kitchen,J.C.&J.R.Zaneveld,1990.Onthenoncorre-lationoftheverticalstructureoflightscatteringandchlorophyllaincaseIwaters.JournalofGeophysicalResearch95:20237–20246.

Klimasauskas,E.,C.Bacon&J.Alexander,2002.Mount

MazamaandCraterLake:GrowthandDestructionofaCascadeVolcanoU.S.GeologicalSurveyFactSheet092-02.

Maffione,R.A.&D.R.Dana,1997.Instrumentsand

methodsformeasuringthebackward-scatteringcoef-ficientofoceanwaters.AppliedOptics36:6057–6067.McIntire,C.D.,G.L.Larson,R.E.Truitt&M.K.

Debacon,1996.TaxonomicstructureandproductivityofphytoplanktonassemblagesinCraterLake,Ore-gon.JournalofLakeandReservoirManagement12:259–280.

Mobley,C.D.,1994.LightandWater:RadiativeTransfer

inNaturalWaters.AcademicPress,SanDiego.

Moore,C.,E.J.Bruce,W.S.Pegau&A.D.Weideman,

1997.WETLabsac-9:fieldcalibrationprotocol,

159

deploymenttechniques,dataprocessing,anddesignimprovements.‘‘OceanOpticsXII’’,ProceedingsoftheSocietyofPhotoOpticsInstrumentationEngineers2963:725–730.

Morel,A.,1973.Diffusiondelalumiereparleseauxde

mer.Resultatexperimentauxetapprochtheorique.InAgardLectureSeries61onOpticsoftheSea~AdvisoryGroupforAerospaceResearchandDevel-opment;NATO,London:3.1.1–76.

Morel,A.,1974.Opticalpropertiesofpurewaterandpure

seawater.InJerlov,N.G.&E.S.Nielsen(eds),OpticalAspectsofOceanography.AcademicPress,NewYork.

Pettit,E.,1936.OnthecolorofCrateLakewater.Pro-ceedings,NationalAcademyofScience22:139–146.Smith,R.C.&K.S.Baker,1981.Opticalpropertiesofthe

clearestnaturalwaters(200–800nm).AppliedOptics20:177–184.

Sosik,H.M.&B.G.Mitchell,1995.Lightabsorptionby

phytoplankton,photosyntheticpigmentsanddetritusintheCaliforniaCurrentSystem.DeepSeaResearch42:1717–1728.

Spinrad,R.W.&J.R.V.Zaneveld,1982.Ananalysis

oftheopticalfeaturesofthenear-bottomandbot-tomnepholoidlayersintheareaoftheScotianRise.JournalofGeophysicalResearch87:9553–9561.

Twardowski,M.S.,J.M.Sullivan,P.L.Donaghay&

J.R.V.Zaneveld,1999.Microscalequantificationoftheabsorptionbydissolvedandparticulatematerialincoastalwaterswithanac-9.JournalofAtmosphericOceanicTechnology16:691–707.

Twardowski,M.S.,E.Boss,J.B.Macdonald,W.S.

Pegau,A.H.Barnard&J.R.V.Zaneveld,2001.AmodelforestimatingbulkrefractiveindexfromtheopticalbackscatteringratioandtheimplicationsforunderstandingparticlecompositioninCaseIandCaseIIwaters.JournalofGeophysicalResearch106:14129–14142.

Urbach,E.,K.L.Vergin,L.Young,A.Morse,G.L.

Larson&S.J.Giovannoni,2001.Unusualbacterio-planktoncommunitystructureinultra-oligotrophicCraterLake.LimnologyandOceanography46:557–572.

Zaneveld,J.R.V.,J.C.Kitchen&C.C.Moore,1994.

Scatteringerrorcorrectionofreflectingtubeabsorp-tionmeters.InAckleson,S.(ed.),OceanOpticsXII,Proc.SPIE2258:44–55.

Zaneveld,J.R.V.,S.Pegau&J.L.Mueller,2003.

Volumescatteringfunctionandbackscatteringcoefficients:instruments,characterization,fieldmeasurementsanddataanalysisprotocols.InMueller,J.L.,G.S.Fargion&C.R.McClain(eds.),OceanOpticsProtocolsforSatelliteOceanColorSensorValidation,Revision4,VolumeIV:InherentOpticalProperties:Instruments,Charac-terizations,FieldMeasurementsandDataAnalysisProtocols,NASATech.Memo.,2003-211621/Rev4-Vol.IV.Greenbelt:NASAGoddardSpaceFlightCenter:65–76.

123