spectradreammail

THERETRIEVALMODELFORORGANICCONTAMINATIONINWATERSUSINGOPTICALABSORPTIONPROPERTIESOFCDOMHUANGMiaofena,*,XINGXufenga,SONGQingjunb,a,SHENZhufeicaDalianOceanUniversity,Dalian116023,China--hmf808@163.
com,xxf@dlou.
edu.
cnbNationalSatelliteOceanApplicationService,SOA,Beijing100081,China--kingdream@mail.
nsoas.
gov.
cncZhejiangInternationalStudiesUniversity,Hangzhou310012China--szf@zjei.
netAbstract:YellowSubstances(alsochromophoricdissolvedorganicmatter:CDOM)isthepoolofabsorbingsubstanceinwater.
Inthefieldsofwatercolourremotesensing,itsopticalabsorptionpropertiesaremainlydescribedwithabsorptioncoefficientandspectralslope.
IthasshowedbypresentresearchthattheinfluenceoforganiccontaminationinwaterincludingpetroleumpollutantandCOD(ChemicalOxygenDemand)onspectralabsorptioncoefficientismainlyrepresentedonabsorptionspectralfeaturesofyellowsubstance.
Theexperimentaldatain-situisobtainedattheregionsofPanjincity,LiaoningprovinceofChinainMay2008,August2009andJune2010.
Withthemeasurementdata,first,theexponentialspectralslope(S)wasdetermined;Secondly,themodelofretrievalag(440)basedontheHJ-1satellitewascreated;Thirdly,theremotesensingmodeltoretrieveorganiccontaminationinwaterincludingpetroleumpollutantandCODwereestablished,thenthevalidationwasdone;Finally,themodelsofretrievingorganiccontaminationfromHJ-1satellitewereappliedintheregionofLiaodongbay.
Keywords:CDOM;absorptioncoefficient;spectralslope;petroleumconcentration;COD;retrievalmodels.
1.
INTRODUCTIONOrganiccontaminationsbelongtochemicalcontaminationandincludenaturalorganiccompoundsandsynthesizedorganiccompounds.
Naturalorganiccompoundsexistintheformofcarbohydrate,protein,aminoacidandfat.
Mostorganiccontaminationscanbedegradableandavailablebymicroorganism,andbecalledoxygenconsumptionorganicmatterbecauseinthecoursedissolvedoxygenwillbeconsumed.
TheconcentrationoforganiccontaminationsismeasuredbyCOD(ChemicalOxygenDemand),whichistheoxygendemandwhenorganiccontaminationsareoxygenatedbyoxidant.
Potassiumpermanganateandpotassiumdichromatearecommonoxidant.
Potassiumpermanganatemethod(abbreviatedtoCODMn)workstomeasureseawater.
Potassiumdichromatemethod(abbreviatedtoCODcrorCOD)workstomeasurewatersampleswhicharepollutedverymuch.
Thechemicaloxygendemand(COD)isaveryimportantparametertoreflectthepolluteddegreeoforganiccontaminationonwater.
Theorganiccontaminationisageneralproblemintheregionofoil-fielddevelopment,industrialproductionandcitizenlife.
ThetraditionaldeterminationofCODinwaterisnotonlytime-consumingandstrenuousbutalsoexpensive.
Moreoveritisdifficulttosampleintheriverwhichresultinthedifficultytoreflectsynchronouslythewholeregionstatus.
Remotesensinghastheadvantageoflargearea,speediness,dynamicandlow-costintheobtainingregionalinformation,soatpresenttherearesomescholarstodevotethemselvestoextractinformationofCODusingsatellitedata(Zhouetal.
,2009;Twardowskietal.
,2004;Bricaudetal.
,1981;Yacobietal.
,2003).
Petroleumsubstancesbelongtohydrocarbonanditsmaincomponentsisconstitutedofalkane,naphthenichydrocarbon,olefinandaromatichydrocarbon.
Floatingoil,dispersedoil,emulsifiedoilanddecompositionoilarethefourformsofpetroleumpollutioninwater.
Theemulsifiedoilordecompositionoiloradsorbedinthesurfaceofparticulatematterwillbeoxy-genatedanddecomposedbymicroorganism,sopetroleumpollutioncanbeclassitiedintoorganiccontamination.
ThetwooilspillsoccurringinMexicoBayandDalianBayin2010havebeendisasterstothemarineenvironment,whichneedsdecadestoberemovedandcostsbillionsofdollars.
Duetothefluidityofwater,howmuchwillthespreadofpetroleumsubstanceandthedurationofitsexistenceinfluenceontheoceanicenvironmentandeconomyTheassessmentneedsurgentlydataoftimeandspatialcontinuity.
Althoughthepresentordinaryobservationsofoceanandlandwaterincludethepollutionofpetroleum,theyareallpointobservationandlackspatialobservationpointsandobservingtimes,whichcanhardlyimplementacontinuousobservationofspatialandtimevariation(Huangetal.
,2007).
Atpresent,monitoringofpetroleumpollutionbyusingremotesensingmainlyfocusesonoilspill,whoseremotesensingmechanismandmodelshavebeensignificantlyimproved(Króletal.
,2006;Otrembaetal.
,2002).
However,researchesonthecomponentsofpetroleumsubstanceinwaterarerarelyreported.
Soresearchedonitsremotesensingmechanismandidentifymodelareofgreatsignificanceandapplicationvalue(Huangetal,2009).
Inthewatercolorremotesensing,ag(440)isusuallyusedtorepresenttheconcentrationofCDOMinwater,spectralslopeStorepresentthedegreeofattenuationofCDOMabsorptioncoefficient.
ThecurrentresearchesshowthatspectralslopeShasnorelationwithCDOMconcentration,butisrelatedtoitscomponent,thesimulationbandrangebyestablishingamodelandreferencewavelength(Wangetal,2007).
Petroleumsubstancesaremainlyhydrocarbons,includingalkanes,cycloalkanesandaromatic.
Huangetal.
(2010)foundinexperimentsthatinthewaterpollutedbypetroleum,theamplitudeofCDOMabsorptioncoefficientincreaseswiththeconcentrationofpetroleum,theshapeofspectradoesn'tchange,obeyingexponentialdecaylaw,butthespectralslopeSchanges,whichmeansthepetroleumsubstancesarecloselyrelatedtoCDOMconcentration(representedbytheabsorptioncoefficientag(440))andspectralslopeS,soitisavailabletoretrievethepetroleumconcentrationinwaterbyusingthecoefficientsofCDOMopticalcharacteristics.
ByusingtheexperimentdataobtainedinLiaoheoilfieldofPanjincity,LiaoningprovinceofCHINAinMay2008,August2009andJune2010,theremotesensingmodesofretrievingwaterpetroleumpollutionandCODbasedonCDOMopticalcharacteristicsareestablished,whichcanprovidetechnicalsupportfordetectingorganiccontaminationbyremotesensingdata.
2.
EXPERIMENTANDDATAACQUISITION2.
1ExperimentalDescriptionTheexperimentalfieldworkwasconductedinPanjincity,LiaoningprovinceofChina.
IntheregionthereareLiaoheoilfieldwhichisthethirdlargestoilfieldinChinaandreedwetlandwhichislargestinAsia.
InNovembereachyearreedisprocessedinthepapermillsafterharvested.
Theactivityoffield'sproducingandpaper-makingisinevitablytobringaboutorganiccontamination.
TheexperimentaltimeisinMay2008,August2009andJune2010.
Theexperimentaldatain-situincludedtheabsorptioncoefficientofCDOM,apparentopticalproperties,CODandpetroleumpollutantsconcentration.
ThemeasuringmethodisshowedintableAandseventysampleswereobtainedTableAthemeasuringdataandmethodsItemsContentsMethodReferencestandardapparentopticalpropertiesremotesensingreflectancespectrometry(ASD)inherentopticalpropertiesabsorptioncoefficientofCDOMUV-3900,350-900nm(HITACHI)NASAstandardwaterqualityparameterpetroleumJK-951GB11914-89waterqualityparameterCODdichromicacidGB11914-892.
2MeasurementsofCDOMabsorptioncoefficientsThemeasurementofCDOMabsorptioncoefficientsisusingGF/Fmembranetofilterwatersamplesandthenmeasuringtheirabsorbancebyspectrophotometer.
Firstly,soakedthepolycarbonatemembraneof0.
22maperturewith10%hydrochloricacid,andthenfilteredwatersampleswithvacuumpumptoobtainCDOMsamples,usedsamplepoolsof10cmlengthandmadereferencebyMilli-Qwater(purewater),thespectrophotometerisU-3010ofHitachi,putthe10cmcuvettefilledwithwaterandsamplecuvetteintothetwoopticalpathsofspectrophotometer,measuringtheopticaldensityODbs(λ)ofthereferencewater(dimensionless);removedthesamplecuvette,discardedthewater,filledwithsamples,measuringtheopticaldensityofCDOMrelatedtowaterODs(λ)(dimensionless).
Inaccordancewiththespecificationofmarineopticalmeasurements(Mitchelletal.
,2000)givenbyNASA,CDOMabsorptioncoefficientiscalculatedas,ag(λ)=2.
303/L*[[ODs(λ)-ODbs(λ)]-ODnull](1)Where:Listheopticalpathofcuvette(typically0.
1m);ODs(λ)istheopticaldensityrelatedtothereferencewatersamples(dimensionless);ODbs(λ)istheopticaldensityofblankpurewaterprocessedbythesamplehandlingprocedurerelatedtothereferencepurewater(dimensionless);ODnullistheapparentresidualopticaldensityatthelongwavelengthbandofvisiblelightoratthebandsofnear-infraredwheretheabsorptionofdissolvedsubstancescanbeassumedtobezero(dimensionless),whichistheresidualabsorptionoflongwavelengthvisiblelightornearinfraredband.
InthespecificationofmarineopticalmeasurementsgivenbyNASAitisrecommendedthatODnulltakestheaverageof590-600nm,duetotheturbidwaterbodycaseII,takingtherecommended590~600nmbandasaresidualcorrectionbandwillresultina"toolow"estimateofCDOMabsorptioncoefficients,mainlybecausetheabsorptionofthedissolvedsubstancesofhighturbidityinwatercaseIIat590~600nmbandisstrongandsmallparticlesinfilteredclearliquidmaycausescattering,sointhispaper,theresidualcorrectionmethodusing750nmwavelengthbandmadeforturbidwaterbodiesbyBricaudetal.
(1981)isapplied,whichiscalculatingtheabsorptioncoefficientsofeachwavelengthλ(nm)byformula(2),andthencorrectingscatteringbyformula(3).
ag(λ')=2.
303/L*[ODs(λ)-ODbs(λ)](2)ag(λ)=ag(λ')-ag(750)·λ/750(3)Where,ag(λ)istheabsorptioncoefficientofCDOM(m-1)atwavelengthλ;ag(λ')istheuncorrectedabsorptioncoefficientofCDOM(m-1)atwavelengthλ;ag(750)istheabsorptioncoefficientofCDOM(m-1)atwavelength750nm;themeaningofλ,ODs(λ),ODbs(λ)andListhesameasformula(1).
2.
3ThemeasurementofpetroleumsubstanceconcentrationThemethodsofdeterminingpetroleumpollutionincludeGravimetricmethod,ultravioletspectrophotometry,fluorescenceandinfraredmethod.
Infraredmethodsincludingnon-dispersiveinfraredmethodandinfraredspectroscopyuseabsorptioncharacteristicsatabout3-4m,butnon-dispersiveinfraredmethoddoesnotconsideraromatics,thusnon-dispersiveinfraredmethodwillreducethemeasuredvalueofpetroleumpollution,especiallythosecontainingaromatichydrocarbonoils.
Sothisstudyusedinfraredspectrophotometry(GB16488-96)todetermineoilpollution.
TheuseddeviceistheJK-951Multi-functioninfraredoilcontentanalyzer.
2.
4ThemeasurementofchemicaloxygendemandCODThemethodofapplyingmicrowavetosealinganddissipatingwasintroducedtomeasurethevalueofCOD(GB11914-89).
ThebasicprinciplesofthismethodaretocalculatethevalueCODcrafterwatersamplewasheatedbymicrowaveandittitrateswithferrousammoniumsulfate,usingtheferroinsulfatesolutionasacatalystandwithH2So4andK2Cr2O7digestion.
2.
5WaterspectrometrymeasurementTheaimofwaterspectrometrymeasurementistoreverseCDOMbyusingremotesensingreflectance.
Forwaterspectrometrymeasurementweappliedvisiblenearinfraredspectroradiometer(ASDFieldSpec3350-2500nm)producedbyASDcompanyoftheUSA,thereferenceboardisthe30%reflectancestandardboard.
Themeasurementmethodwasabovethewatersurface,using(40°,135°)relativeobservationgeometry,whereobservationnadiranglewas40°,theanglebetweenequipmentobservationazimuthandsolarwas135°.
Ateachobservationpoint,measuredintheorderofthegrayboard,water,sky,grayboard,grayshadingboard,eachgroupmeasured15spectra,eachspectralintervalwas1second.
2.
6HJ-1A/1BCCDThesatellitesenvironmentA&B(HJ-1A/1B)wassuccessfullylaunchedonSeptember6,2008,eachHJ-1AandHJ-1BwereloadedwithtwoCCDcamerasofexactlythesamedesignedprinciple,whichare4-bandwide-covermulti-spectralvisiblelightcameraCCD.
TableBshowsthetechnicalparametersoftheCCDonHJ-1Aand1B.
HJ-1AandHJ-1Barenotcolorsatellite,buttheCCDcamerasstillhassomeinformationofwater,whichmayserveasawatercolorremotesensingdeviceofenvironment(Huangetc,2010).
TableBthetechnicalparametersofHJ-1A/1BCCDItemPerformanceNadirgroundpixelresolution(m)30B10.
43~0.
52B20.
52~0.
60B30.
63~0.
69Bandset(m)B40.
76~0.
90Quantifiedvalue(bit)83.
.
.
.
THEINVERSIONMODELOFORGANICCONTAMINATION3.
1AnalysisofCDOMopticalabsorptioncharacteristicsPreviousstudies(Bricaudetal,1981)showedthat,CDOMabsorptionspectrumdecreasedexponentiallywithwavelengthfromUVtovisiblelight,undernormalcircumstances,usingtheformula(4)todescribetheCDOMabsorptioncoefficientchangeswithwavelength,ag(λ)=ag(λ0)exp[-S(λ-λ0)](4)Whereag(λ)istheCDOMabsorptioncoefficient(m-1)atwavelengthλ,λ0isthereferencewavelength(nm),Sisthespectrumslopecoefficient(nm-1).
InthestudyofCDOMopticalcoefficientproperties,theabsorptioncoefficientat440nm(ag(440))andspectralslopeSaretwoimportantparameters.
About28sampleswasusedofmixproportiontextinAugust2009andJune2010,took440nmasreferencebandλ0,substitutingintoformula(4)fittingoutspectralslopeSinwaterwithorganiccontamination.
ItisshowedthatinthisexperimentthespectralslopeSarebetween0.
0086nm-1~0.
014nm-1,theaverageis0.
010892nm-1.
3.
2TheretrievalmodelofCDOMconcentrationCurrently,theCDOMconcentrationretrievalmodelsestablishedforwatercolorsensorsaremostlyusedthebandratiomethod(Shen,etc.
,2006).
Bowersetal(2004)proposedtheremotesensingmodelfortheestuaryregiontousethereflectanceratioat670nmand490nmestimatingCDOMabsorptioncoefficientat440nm,thatis,ag(440)=1.
45*(R(670)/R(490))-0.
488(5)Whereag(440)isCDOMabsorptioncoefficientatwavelength440nm(m-1);R(670)andR(490)areremotesensingreflectanceatwavelength670nmand490nmrespectively.
TakingintoaccountthatHJ-1/CCDbandsetincludes670nm(B3)and490nm(B1),thestudyareaislocatedinestuaries,sobasedontheBower'smodel,weusedthetestdatainAugust2009,combiningwithHJ-1/CCDbandresponsefunctiontosimulatetheremotesensingmodelofretrievingCDOMwaterbyusingthetwowide-bands(B1:430~490nmandB3:630~690nm)ofnon-water-colorsatelliteHJ-1/CCD,thatisag(440)=2.
47*(R3/R1)-0.
27(6)Whereag(440)isthesameasinFormula(5);R1andR3aretheremotesensingreflectancecorrespondingtobandB1andB3ofHJ-1/CCD(dimensionless).
3.
3Correlationanalysisag(440)usuallyrepresentstheconcentrationofCDOM,whilethevariationofSmainlydependsonthecompositionofyellowsubstances,sothesetwofactorsshouldbebothconsideredbyestablishingorganiccontaminationremotesensingretrievalmodelbasedonCDOMopticalabsorptionpropertiestoimprovetheretrievalaccuracy.
ThecorrelationcoefficientbetweenorganiccontaminationandthevariouscombinationsofSandag(440)areshowedintableCandtableD(Saveraged0.
010892).
TableCcorrelationbetweenpetroleumconcentrationandyellowsubstanceabsorptionspectralparameters(n=70)S+ag(440)S*ag(440)S/ag(440)ag(440)/S0.
8177640.
8351720.
732592-0.
767891TableDcorrelationbetweenpetroleumconcentrationandyellowsubstanceabsorptionspectralparameters(n=70)ItcanbeseenfromtableC,theconcentrationofpetroleumpollutionhasthehighestcorrelationwithS*ag(440),theircorrelationcoefficientis0.
835132.
ItcanalsobeseenfromtableD,thevalueofCODhasthehighestcorrelationwithS*Ln(ag(440)),theircorrelationcoefficientis0.
927975.
Soweusedthesecombinationsasindependentvariablestoestablishthemodelofretrievingorganiccontaminationinwater.
3.
4TheestablishmentofretrievingmodelsIntableEtheyaretheestablishedrelationshipsmodelsofS*ag(440)andtheconcentrationofpetroleumsubstances.
Itcanbeseenfromthetable,intherelationalmodelofpetroleumcontentandag(440)*S,theR2ofpowerismaximum,upto0.
7418,theR2oflinear,logarithmicandexponentialrelationshiparealllessthan0.
7,sowechosepowerasthemodelofretrievingwaterpetroleumpollutionbasedonCDOMopticalabsorptionspectra,thatis,y=127.
566x1.
3288(7)Whereyistheconcentrationofpetroleumpollutioninwater(mg/L),xisag(440))*S,themeaningsofag(440))andSarethesameasbefore;ag(440)iscalculatedinformula(6).
TableEcorrelationmodelsbetweenpetroleumconcentrationandyellowsubstanceabsorptionspectralparameters(n=70)TypeofmodelsPowerLinearLogarithmicExponentialFormulay=127.
566x1.
3288y=240.
64x-0.
5815y=5.
7608Ln(x)+27.
858y=1.
4527e40.
796xR20.
74180.
69740.
6180.
6989*where,yistheconcentrationofpetroleumpollutioninwater;xisag(440))*S.
AccordingtothecorrelationofCDOMandCOD,therelationalmodelofretrievingCODwasestablishedbasedontheparametersofln(ag(400))*S,thatis,y=0.
583936X2-9.
5481X+0.
52127(R2=0.
876)(8)WhereyisthevalueofCODinwater(mg/L),xisln(ag(400))*S,themeaningsofag(440)andSarethesameasbefore.
.
4.
THEVERIFICATIONANDAPPLICATIONOFTHEMODEL4.
1TheverificationofthemodelSubstitutedthevalueofCODandpetroleumcontaminationdeterminedinanaturalriverinMay2008andAugust2009,andtheabsorptioncoefficientvalueatag(440)ofyellowsubstancesinwaterintoformula(7)and(8)toestimatethevalues(31Samples).
Theresultsshowedthatestimatedvaluesareveryclosetomeasuredvalues,therelativeerrorsare6.
8847%and7.
6695,respectively.
Thisindicatesthatformula(7)and(8)asthemodelofremotesensingretrievingorganiccontaminationconcentration,theretrievalresultsarerelativeaccurate,whichprovidesanewwaytoestimateorganiccontaminationconcentrationinwaterwithremotesensing.
4.
2TheapplicationofthemodelApplyingthemodelofRemotesensingretrievingpetroleumsubstancesestablishedinthispapertothetwoHJ-1transitimagesonSeptember22,2009,asshowninFig.
1below.
Fig.
1showedthattheconcentrationofpetroleumsubstancesfromtheoffshoretothehighseamaintainsahigh-to-lowtendency.
Sag(440)ln(ag(440))ln(ag(440))*S0.
04640.
90830.
9250.
9279Fig.
1ThespatialdistributionofthepetroleumconcentrationinLiaodongbayonSeptember22,2009ApplyingthemodelofRemotesensingretrievingpetroleumsubstancesestablishedinthispapertothetwoHJ-1transitimagesonSeptember22,2009,asshowninFig.
2below.
Fig.
2showedthatthevalueofCODisamong52.
1-168.
2mg/Landfromtheoffshoretothehighseamaintainsalow-to-hightendency.
Fig.
2ThespatialdistributionofthevalueofCODinLiaodongbayonSeptember22,20095.
CONCLUSIONSCODreflectsthedegreeoforganiccontaminationonwateranditsinfluenceonwaterabsorptioncoefficientmainlyreflectonthatofCDOM.
Petroleumcontaminationbelongstotheoneoforganiccontamination.
ThusitisfeasiblethatusingopticalabsorptionpropertiesofCDOMretrieveorganiccontaminationinwater.
Throughthemixproportiontestinfield,theremotesensingmodelsofretrievingorganiccontaminationinwaterbasedontheCDOMabsorptionspectralisproposed.
Themodelsarenotonlysimpleandeasytooperate,butalsosuitableasabusinessrunning.
AslongastheCDOMconcentrationag(440)iscalculatedwithremotesensingreflectancetodeterminearegionalexponentialslopeS,thevalueoforganiccontaminationinwatercanbefiguredout.
Theestablishmentofandprovideawaytomonitorwaterorganiccontaminationwithremotesensingtechnology.
Themodelscanpromotetheapplicationofwatercolorremotesensinginmonitoringcomponentsofenvironmentalpollution.
REFERENCESZhouGuang-hua,TANGJun-wu,TIANGuo-liang,LIJingandLIUQin-huo.
2009.
UncertaintyAnalysisofInlandWaterQualityRemoteSensing:AReview.
AdvancesinEarthScience,24(2):150-158.
TwardowskiMichaelS,BossEmmanuel,SullivancJamesM.
2004.
Modelingthespectralshapeofabsorptionbychromophoricdissolvedorganicmatter[J].
MarineChemistry,89:69-88.
BricaudA,MorelAandPrieurL.
1981.
AbsoptinbyDissolvedOganicMatterofTheSea(YellowSubstance)intheUVandVisibleDomains[J].
LinmalOceanogr,26:43-53.
YacobiYZ,AlbertsJJ.
andTakdcsM.
2003.
AbsorptionspectroscopyofcoloreddissolvedorganiccarboninGeorgia(USA)rivers:theimpactofmolecularsizedistribution[J].
JournalofLimnology,62(1):41-46.
HUANGMiao-fen,QIXiao-ping.
,YUWu-yi,ZHANGYi-mim.
2007.
IdentificationModeofPetroleumPollutioninWaterBasedonRemoteSensingTechniqueandItsApplication[J]RemoteSensingTechnologyandApplication,(3):314-320.
KrólT.
,StelmaszewskiA.
andFredaW.
2006.
Variabilityintheopticalpropertiesofacrudeoil–seawateremulsion[J].
OCEANOLOGIA,48(S):203–211.
OtrembaZ.
andKrólT.
2002.
ModelingoftheCrudeOilSuspensionImpactonInherentOpticalParametersofCoastalSeawater[J].
PolishJournalofEnvironmentalStudies,11(4):407-411.
HUANGMiao-fen,SONGQing-jun,TANGJun-wuetal.
.
2009.
Analysisofbackscatteringpropertiesofpetroleumpollutedwater:acasestudyattheLiaoheRiverandtheRaoyangRiverinLaoningProvince,China[J].
ActaOceanologicaSinica,31(3):12-20.
Wanglin,ZhaoDongzhi,FuYunna,YangJianhong,etal.
2007.
Correlationbetweenabsorptioncoefficientag(440)andtheslopeSgofCDOM[J].
JournalofDalianMaritimeUniversity,33(S2):179-182.
HUANGMiao-fen,TANGJun-wu,SONGQing-jun.
2010,AnalysisofWaterAbsorptionSpectralFeaturesofPetroleumPollution[J].
JournalofRemoteSensing,14(1):140~156.
Mitchell,B.
G.
,BricaudA.
,CarderKandClevelandJ.
2000:Determinationofspectralabsorptioncoeddicientsofparticles,dissolvedmaterialandphytoplanktonfordisscretewatersamplrs,In:Fargion,G.
S.
andJ.
L.
Mueller,[Eds.
]OceanOpticsProtocolsforSatelliteOceanColorSensorValidation,Revision2.
NASA/TM-2000-209966,NASAGoddardSpaceFlightCenter,Greenbelt,MD.
Chapter12,pp125-153.
HuangMiaofen,NiuShengli,SunZhongpingi,etal.
2010.
AnalysisonCharacteristicoftheWaterBodyInformationCollectedby"HJ-1"satelliteMultispectralCCDSensors[J].
RemoteSensingInformation,(4):68-75.
SHENHong,ZHAODong-zhi,FUYun-na,etal.
2006.
AnOverviewofOpticsCharacteristicandRemoteSensingofCDOM[J].
JOURNALOFREMOTESENSING,10(6):949-954.
BowersDG,EvansDandThomasDN.
2004.
Interpretingthecolourofanestuary[J].
Estuarine,CoastalandShelfScience,59:13-20.
ACKNOWLEDGMENTSThisworkwasfundedbyNationalOceanicPublicGoodProject(GrantNo.
201005025-04),NationalNaturalScienceFoundationofChina(GrantNo.
40771196)andKeyProjectsintheNationalScience&TechnologyPillarProgramduringtheEleventhFive-YearPlanPeriod(GrantNo.
2008BAC34B05-5).