Oyster Aquaculture Site Selection Using Landsat 8-derived ...
Post on 15-Jan-2022
9 Views
Preview:
Transcript
The University of MaineDigitalCommons@UMaine
Electronic Theses and Dissertations Fogler Library
Summer 8-18-2017
Oyster Aquaculture Site Selection Using Landsat8-derived Sea Surface Temperature, Turbidity, andChlorophyll a.Jordan SnyderUniversity of Maine, Orono, jordan.snyder@maine.edu
Follow this and additional works at: http://digitalcommons.library.umaine.edu/etd
Part of the Oceanography Commons
This Open-Access Thesis is brought to you for free and open access by DigitalCommons@UMaine. It has been accepted for inclusion in ElectronicTheses and Dissertations by an authorized administrator of DigitalCommons@UMaine. For more information, please contactum.library.technical.services@maine.edu.
Recommended CitationSnyder, Jordan, "Oyster Aquaculture Site Selection Using Landsat 8-derived Sea Surface Temperature, Turbidity, and Chlorophyll a."(2017). Electronic Theses and Dissertations. 2722.http://digitalcommons.library.umaine.edu/etd/2722
OYSTERAQUACULTURESITESELECTIONUSINGLANDSAT8–DERIVED
SEASURFACETEMPERATURE,TURBIDITY,ANDCHLOROPHYLLA
By
JordanSnyder
B.S.UniversityofCalifornia,Davis,2013
ATHESIS
SubmittedinPartialFulfillmentofthe
RequirementsfortheDegreeof
MasterofScience
(inOceanography)
TheGraduateSchool
TheUniversityofMaine
August2017
AdvisoryCommittee:
EmmanuelBoss,ProfessorofMarineScience,CoAdvisor
DamianBrady,AssistantProfessorofMarineScience,CoAdvisor
AndrewThomas,ProfessorofMarineScience
OYSTERAQUACULTURESITESELECTIONUSINGLANDSAT8-DERIVED
SEASURFACETEMPERATURE,TURBIDITY,ANDCHLOROPHYLLA
ByJordanSnyder
ThesisAdvisors:Dr.EmmanuelBoss,DamianBrady
AnAbstractoftheThesisPresented
inPartialFulfillmentoftheRequirementsfortheDegreeofMasterofScience
(inOceanography)
August2017
Remote sensing data is useful for selection of aquaculture sites because it can providewater-quality
productsmappedwithnocosttousers.However,thespatialresolutionofmostoceancolorsatellitesis
toocoarsetoprovideusabledatawithinmanyestuaries.ThemorerecentlylaunchedLandsat8satellite
hasboththespatialresolutionandthenecessarysignaltonoiseratiotoprovidetemperature,aswellas
oceancolorderivedproductsalongcomplexcoastlines.ThestateofMaine(USA)hasanabundanceof
estuarine indentations (~3,500miles of tidal shoreline within 220miles of coast), and an expanding
aquaculture industry,whichmakes itaprimecase-study forusingLandsat8data toprovideproducts
suitable foraquaculturesite selection.Wecollected theLandsat8 scenesovercoastalMaine, flagged
clouds,atmosphericallycorrectedthetop-of-the-atmosphereradiances,andderivedtimevaryingfields
(repeattimeofLandsat8is16days)oftemperature(100mresolution),turbidity(30mresolution),and
chlorophyll-a (30 m resolution). We validated the remote-sensing-based products at several in situ
locationsalongtheMainecoastwheremonitoringbuoysandprogramsareinplace.Initialanalysisofthe
validated fields revealed promising areas for oyster aquaculture. The approach used and the data
collectedtodateshowpotentialforotherapplicationsinmarinecoastalenvironments,includingwater
qualitymonitoringandecosystemmanagement.
iii
ACKNOWLEDGEMENTS
ThankyouEmmanuelBoss,DamianBrady,AndyThomasandCarterNewellforguidingandinstructing
me.ThankyouRyanWeatherbeeforpatientlyworkingwithmeandhelpingmewiththesatellitedata
processing. Thank you to Catherine Coupland, Nicolas DelPrete, TiegaMartin, Chris Rigaud,Matthew
Grey,andRobbieDownsforyourassistancemaintainingtheLOBObuoysattheDarlingMarineCenter.
Thank you to Nils Haëntjens for assistance with data processing and editing. Thank you to Jocelyn
RunnebaumandKevinStaples for thehelpfuledits. Thankyou to theSEANETprojectatUniversityof
MaineforprovidingLOBObuoydataandtravelsupport.ThankyouDanaMorse,BethBissonandMaura
Thomasforsupport.ThankyouKelly,LeAnnandAliforbeingspectacularlabmates,andthankyoutomy
familyforyourcontinuedencouragement.
iv
TABLEOFCONTENTS
ACKNOWLEDGEMENTS...............................................................................................................................iii
LISTOFTABLES............................................................................................................................................viLISTOFFIGURES.........................................................................................................................................viiLISTOFEQUATIONS....................................................................................................................................ixLISTOFABBREVIATIONS...............................................................................................................................xChapter1.INTRODUCTION.....................................................................................................................................12. METHODS…………………………….………………………………………………………………………………………………………….6 2.1StudyArea…................................................................................................................................6 2.2ProcessingofSeaSurfaceTemperature.....................................................................................7
2.3OceanColor.................................................................................................................................9 2.4AtmosphericCorrectionfor𝑅"#................................................................................................10 2.5RetrievalofTurbidity................................................................................................................11 2.6RetrievalofChlorophyll-a..........................................................................................................11 2.7Validationwithinsitudata……..................................................................................................12 2.8SatelliteImageryforanOysterSuitabilityIndex………………………………………………………..………....13
3. RESULTS……………….............................................................................................................................15 3.1Satelliteretrievedvalidationwithinsitudata...........................................................................15 3.2SatelliteImageryforOysterGrowthConditions........................................................................204.DISCUSSION………....................................................................................................................................22
4.1SatelliteImagery…….................................................................................................................22 4.2LimitationsinValidationProcess…………………....……………………………………………………………………22
v
4.3OysterSuitabilityIndex……………………………......……………………………………………………………………….244.4FutureWork……………………………………….....……………………………………………………………………………..25
5.CONCLUSION……....................................................................................................................................27REFERENCES...............................................................................................................................................28APPENDICES...............................................................................................................................................34
AppendixA.AssessmentofAtmosphericCorrection......................................................................34AppendixB.OysterSuitabilityIndex...............................................................................................37AppendixC.AveragedMonthlySatelliteData………........................................................................38AppendixD.StandardDeviationofMonthlyClimatologyMaps…..................................................41
BIOGRAPHYOFTHEAUTHOR.....................................................................................................................44
vi
LISTOFTABLES
TableA.1. MeasuredValuesinHumicPond............................................................................34
TableA.2. ValuesfromliteratureforequationA(1)................................................................35
TableA.3. DilutionseriesofArizonaDuststandardwithHachandLOBOWQMturbidity
measurements........................................................................................................35
TableB.1.CriteriaforOysterSuitabilityIndex….....................................................................37
TableB.2.OysterSuitabilityIndexscoresandaverageJulySSTatexistingandprospectiveoyster
aquaculturesitesinMaine.........................................................................................................…37
vii
LISTOFFIGURES
Figure1. Mapofmid-coastMaine,USA...................................................................................7
Figure2. Landsat8-derivedSeasurfacetemperaturemapofmid-coastMaineon
July14,2013............................................................................................................15
Figure3. TypeIIlinearregressionformatch-upsbetweenLandsat8seasurface
temperatureandseasurfacetemperaturemeasuredbyoceanographicbuoys…..16
Figure4. Landsat8-derivedturbidityalongmid-coastMaineonJuly14,2013.....................17
Figure5. TypeIIlinearregressionbetweenLandsat8turbidityandturbiditymeasured
byLOBObuoys........................................................................................................18
Figure6. Landsat8-derivedchlorophyll-aalongmid-coastMaineonJuly14,2013..............19
Figure7. TypeIIlinearregressionbetweenLandsat8chlorophyll-aandchlorophyll-a
measuredbyLOBObuoys.......................................................................................20
Figure8. Oystersuitabilitymapbasedonphysicaloceanographicparameters:
seasurfacetemperature,turbidity,andchlorophyll-a............................................21
FigureA.1. TypeIIlinearregressionbetweenLandsat8chlorophyll-aandchlorophyll-a
measuredbyLOBObuoysatnighttime..................................................................36
FigureC.1. SeasurfacetemperaturederivedfromLandsat8dataaveragedoverall
imagesinJulyfrom2013to2016.…….....................................................................38
FigureC.2. TurbidityderivedfromLandsat8dataaveragedoverallimagesinJuly
from2013to2016.…….............................................................................................39
FigureC.3. Chlorophyll-aderivedfromLandsat8dataaveragedovereachimageinJuly
from2013to2016.……............................................................................................40
FigureD.1. StandarddeviationofmonthlyaveragedseasurfacetemperaturedatainJuly
from2013to2016…................................................................................................41
viii
FigureD.2. StandarddeviationofmonthlyaveragedturbiditydatainJulyfrom
2013to2016….........................................................................................................42
FigureD.3. Standarddeviationofmonthlyaveragedchlorophyll-adatainJulyfrom
2013to2016….........................................................................................................43
ix
LISTOFEQUATIONS
Equation1. Retrievalofturbiditywithremotesensingreflectance....................................................11
Equation2. Retrievalofchlorophyll-awithOC3……………….................................................................11
Equation3. CalculationofOysterSuitabilityIndex………………………………………………………………………….14
EquationA.1. Relationshipbetween𝑅"#andabsorptionandbackscatteringcoefficients………………….34
x
LISTOFABBREVIATIONS
SST-SeaSurfaceTemperature
T–Turbidity
SPM–SuspendedParticulateMatter
Chla–Chlorophyll-a
AVHRR–AdvancedVeryHighResolutionRadiometer
𝑅"#–Remotesensingreflectance
OSI–OysterSuitabilityIndex
DRE–DamariscottaRiverEstuary
CDOM–coloreddissolvedorganicmatter
NERACOOS–NortheasternRegionalAssociationofCoastalOceanObservingSystems
LOBO–Land/OceanBiogeochemicalObservatory
NTU–Nephalometricturbidityunits
𝑏–backscattering
𝑏%&–backscatteringofparticles
𝑏%'–backscatteringofwater
𝑎'–absorptionofwater
𝑎&∗–absorptionofparticles
𝑎*–absorptionofdissolvedsubstances
1
CHAPTER 1
INTRODUCTION
Oyster aquaculture of the American oyster,Crassostrea virginica, is an expanding industry in coastal
Maine,USA,withlandingsworth$4.8milliondollarsin2015,upfrom$0.6millionin2003andincreasing
by250%between2011and2015(MaineDMRcommerciallandings2016,www.maine.gov/dmr).Tomeet
thegrowingdemandforhighqualityoysters,identificationofnewsiteswiththemostoptimalbiophysical
conditionsforoystergrowthisneeded.Todecreasetheriskofchoosinganunproductivesite,itiscrucial
thatgrowershavetherighttoolsforsiteselection.Currently,themethodforselectingasuitablesitefor
bivalveaquacultureislargelybasedonproximitytoexistingsitesortrialanderror.Thesemethodsare
inefficientbecausetheymaynotconsiderthespecifictemperatureandnutritionalconditionsneededfor
thespeciestogrow,eachofwhichaffecthowfast it takestoreachmarketsize(Hawkinsetal.,2013;
Rheault&Rice,1996).Recentadvancesinremotesensingtechniquesenablesatelliteimagerytohelpin
site selection (e.g. Thomas et al., 2011). By visually inspecting information products calculated from
processedLandsat8satellite images,estuaries that reachrelativelywarmtemperatures (above20°C),
supporthighlevelsofchlorophyllinthesummer(above1μgChlL-1),andexhibitlowturbidity(below8
NTU)canbeefficientlyidentifiedaspotentialoystergrowingareas.
Thespatialresolutionofstandardglobaloceancolorsatellites(typically1kmx1km)istoocoarse
toprovideusabledatawithinthemanyestuariesandembaymentsalongcoastalMainewheremuchof
the suitable habitat for oyster aquaculture is located. The Thermal Infrared Sensor (TIRS) and the
OperationalLandImager(OLI)aresensorsontheLandsat8satellite,launchedFebruary11,2013.These
sensorshaveboththespatialresolution(100mforinfrareddataand30mformultispectralvisibledata)
and the necessary signal to noise ratio to provide useful temperature aswell as ocean color derived
products along the Maine coastline (Vanhellemont and Ruddick, 2014). In this paper, we used a
2
combination of empirical and analytical approaches to derive temperature, turbidity and chlorophyll
productsfromLandsat8OperationalLandImager(OLI)andThermalInfraredSensor(TIRS)dataforthe
coastofMaine.
Although it was designed for terrestrial monitoring, Landsat 8 data can be used for marine
applications ifa reliableatmospheric correction isapplied.Anatmospheric correction isnecessary for
satelliteremotesensingbecauseinthevisiblewavelengthsthesignalobservedbythesatelliteisreflected
from gas and aerosol particles in the atmosphere (Mobley et al, 2016).We used theNASA software
platformSeaDAS,andalgorithmsimplementedwithinit,togetherwithanempiricalapproachtoderive
chlorophyllandturbidity.
Aswithanyinstrument,therearelimitationstousingLandsat8imageryforcoastalmonitoring.
Comparedtosatellites,suchasAVHRRandMODIS,thathavedailycoverage,thetemporalresolutionof
Landsat8coverageislow.The16dayrepeatcoveragemakesitinsufficienttoobserveshort-termchanges
(duetoweather,stormevents,etc.),butit isusefulfordescribingpatternssuchasseasonalaverages,
whichisinformativeformonitoringlong-termconditionsandrelativespatialdifferencesbetweensites.
Additionally,cloudcoverdecreasestheprobabilityofclearoverpasses;mostoftheimagesweretrieved
comefromsummerandfallmonths(JunethroughNovember)whentherewastheleastamountofcloud
cover.Fortunately,thisisalsothecriticaltimeofyearforoysteraquacultureasitoverlapsmostofthe
growingseason.
Ocean color measurements can be used to describe components of water quality, such as
turbidityandchlorophyll-a(Chla)concentration(O’Reillyetal.,1998).Algorithmshavebeendeveloped
that can estimate concentrations of these components by 1) retrieving radiant flux from the target
surface,2)correctingforthesignalfromtheatmosphere,3)transformingradiantenergycollectedbythe
satellite sensor into remote sensing reflectance (R,-), and 4) converting R,- values into products.
3
Reflectanceintheredwavelengthsoflightisusedtoestimatesuspendedparticulatematter(Dogliottiet
al.,2015;VanhellemontandRuddick,2014),whilereflectanceintheblueandgreenwavelengthsisused
toestimateofChlabiomass(aproxyofphytoplanktonbiomass)(Franzetal.,2015;Mobleyetal.,2016).
Thesemethodshavebeenusedformonitoringinseveralsitesaroundtheworld(Aguilar-Manjarrezand
Crespi,2013;Gernezetal.,2014;Radiartaetal.,2008;Thomasetal.,2011;Wangetal.,2010)toassess
theimpactsofturbidityandChlaonaquaculture.
Optimalconditionsforoystergrowthhavebeendeterminedprimarilythroughtheuseofvarious
types of ecophysical models. Habitat suitability models were first applied to the restoration of the
Americanoyster,Crassostrea virginica, on thewarm southeastAtlantic coastof theU.S. (Cake, 1983;
SoniatandBrody,1988;Barnesetal.,2007).Thesemodels consideredbottomsubstrateandsuitable
salinities to maximize oyster survival in relation to siltation and protozoan parasites. More recently,
Radiartaetal.(2008)usedsatelliteimageryofChlaandseasurfacetemperature(SST),andweightedbio-
physical, social-infrastructural constraint criteria and amodel builder inARCGIS to identify sites best
suitedforhangingcultureofscallops(i.e.highfoodavailability,minimaldistancetosupportservices,and
favorabledepth).CarrascoandBaron(2010)usedsatelliteimagerytomaptemperatureswhichdefined
thepotentialrangeforPacificoysterpopulationsinSouthAmerica.Thomasetal.(2011)usedsatellite-
derivedSSTandChla inMontSaint-MichelBay,France,topredictmusselgrowthbasedonadynamic
energybudgetmodel.Statisticalmodelsrelatingorganismgrowth,biomassandeconomicyieldillustrate
theimportanceofsitespecificenvironmentalvariables(watervelocity,foodconcentration)onfarmyields
(Pérez-Camachoetal.,2014).Powelletal. (1992)andHoffmanetal. (1992)modeledAmericanoyster
filtrationrateandgrowthasafunctionofanimalsize,watertemperatureandtotalparticulatematter,
withanegativeeffectathighsuspendedloads,althoughselectionfororganicmatterbytheoysterwhen
producingpseudofeceswasnotconsidered(NewellandJordan,1983).Gernezetal.(2014),used300m
4
pixel-size SPM distributions derived from MODIS to provide a spatial picture of the impact of SPM
concentrationonoyster-farmingsites.
Crassostrea virginica is somewhat unusual in that its filtration rate is a strong function of
temperature from8°C to amaximumat 30°C compared to other bivalveswhere the filtration rate is
relatively independentofwater temperature (Loosanoff,1958). Therefore, temperature isofprimary
importanceinsiteselectionforoysteraquacultureintherelativelyheterogeneousandstronglyseasonal
seasurfacetemperatureregimeofthecolderMainewaters.Bivalvefeedingandgrowthisalsoapositive
function of phytoplankton concentration (Hawkins et al., 2013), so Chla is considered the nextmost
importantfactorforsiteselection.Ingeneral,totalsuspendedparticulatematterhasanegativeeffecton
bivalvegrowthbydilutingtheorganicmatterathighlevels(Widdowsetal.,1979;Barilleetal.,1997).In
someareas,thereisarelativelyhighproportionofinorganicparticlesinresuspendedsediments,andin
others,sedimentsconsistofbothinorganicmatterandparticlesthatcontainchlorophyll.Forbivalves,the
proportionofphytoplanktoninthesuspendedparticlesisakeyaspectofsitesuitability,(Newelletal.,
1989).
Another important factor in oyster site selection is water velocity, which delivers food to
populationsofoystersatcommercial-scaledensities.Congletonetal.(1998)developedaGISsystemthat
included water velocity and intertidal elevation to predict optimal locations for clam (Mya arenaria)
mariculture.Withinacoastalbay,ShellGISusedthegrowthmodelShellsimtopredictoystergrowthand
yieldasafunctionofwaterquality(temperature,salinityandfoodconcentration),husbandryandseeding
density,andwatervelocityona50mfarmscale(Newelletal.,2013;Hawkinsetal.,2013).Watervelocity
isnotalimitingfactorinthecoastofMainewheretidalamplitudesandcurrentsarelarge.Hence,the
primaryscreeningtoolsoftemperature,chlorophylla,andturbidityareeffectivetoolstoidentifysuitable
5
locationsonthebayscale,andprovidenovelopportunitiesformappingpotentialzonesforaquaculture
developmentoverlargecoastalregionssuchasMaineorAlaskaintheU.S.
WepresentanddemonstrateamethodologytoobtainSSTandcalibratedwaterqualityproducts
fromtheTIRSandOLIsensorsonboardLandsat8,andvalidatethemwithmeasurementsincoastalMaine
waters. We computed uncertainties based on match-ups between local data and that derived from
satellites anddiscusshow temporal and spatial samplingandadjacencyeffects affect theaccuracyof
remote sensing products. These processed satellite products were then used for mapping oyster
aquaculturesites,andprovedusefulbecausetheyverifiedgoodconditionsatexistingfarms,andrevealed
otherlocationsalongthecoastofMainewithsimilarlyoptimalconditionsthatcouldbedevelopedfor
oysteraquaculture.
6
CHAPTER2
METHODS
2.1.StudyArea
ThecoastofMaine includesa seriesof fjards (shallowerandbroader fjords)and jaggedembayments
carved by receding glaciers during the Pleistocene epoch. In situ samples were collected and ocean
monitoring buoy systems were maintained in two of these estuaries, the Damariscotta River and
HarpswellSound,overthecourseofseveralyearsandweusedthemheretovalidateLandsat-8derived
productsontheMainecoast(trianglesonFigure1).WechosetofocusontheDamariscottaRiverbecause
ithasexistingaquacultureoperations (currently75%of theoystersproduced inMaine, (MaineDMR,
2015))andsuitablesamplingaccess.TheDamariscottaRiverEstuaryis29kilometerslong,hasamean
summerflushingtimeof4to5weeks,andisdominatedbystrongtideswithamplitudesofupto3.35m
(McAlice,1977).Sedimentresuspensioninthisestuaryishighestatlowtide,andlowestathightide.The
estuary is highly saline, ranging from 25 to 32.5 psu,with a small amount of freshwater input from
DamariscottaLakeintoSaltBayatitsnorthernreach.Thesubstrateisasoft,muddybottomcomposed
ofclaytosandysiltswithanaveragedepthof15.25m.TheseattributesmaketheDamariscottaRiveran
idealplaceforgrowingmarket-sizeoystersandotherbivalvespecies,andmakeitanexcellentreference
pointforexpandingtheaquacultureindustryalongthecoastofMaine.
7
Figure1.Mapofmid-coastMaine,USA.Trianglesindicatelocationsofvalidationbuoys.Freshwaterlakesusedfortheatmosphericcorrectionarelocatedatapproximately44'N,-69.5'E.
2.2.ProcessingofSeaSurfaceTemperature
AllapplicablerawdatafromLandsat8wasdownloadedfromtheUSGSEarthExplorerwebsitefromthe
period2013to2016(USGS,2016).TocalculateSST,weusedbrightnesstemperaturevaluesfromLandsat
8’sThermalInfraredSensor(TIRS)Band10.TherearestraylightissuesassociatedwiththetwoTIRSbands
(Band10andBand11)duetothecurvatureoftheoptical lens(Montanaroetal.,2014).Ofthesetwo
bands,wechosetousethermalBand10becauseithaslesserissuesofthetwo(seeDiscussionsection).
EachimagewasprocessedintheNASASeaDASplatformuptolevel2toretrievelatitudeandlongitude
8
arrays,ageo-registeredimage,andtheassociatedland/cloudmask(georeferencingismaintained,asitis
providedfromUSGS).
RegressionsbetweencoincidentatmosphericallycorrectedAVHRRsatellitederivedSSTandthat
derivedfromLandsat8’sbrightnesstemperaturewereusedtocreateanSSTproductfortheLandsat8
imagery (similar to Thomas et al., 2002). This regression process, de-facto, acts as the atmospheric
correction for theLandsatSSTproduct1)assuming that theatmospheredoesnotchange in the time
intervalbetweenAVHRRandLandsatoverlappingimageand2)theatmosphereishomogenousacross
theLandsatscene.ExampledatafromthisprocedurearedisplayedonFigure2below.Ofthefourtoeight
AVHRRimagescapturedonthesamedayasLandsat8,wesubjectivelychosetheimagewiththeleast
amountofcloudcoverandpoorlymaskedpixels,bestgeolocation,andcleanestSSTpatterns, for the
regression(seeAppendixAfordetaileddescription).Thedatafortheregressionwasselectedfromcloud
free and offshore areas to accommodate the lower AVHRR resolution (1 km versus Landsat 8 100m
resolution).ThebestresultswereachievedusingcloudfreeareaswithahighdynamicrangeinSST.The
resultingregressionequationbetweenthesignalofLandsat’sBand10andtheAVHRR-basedtemperature
wasthenappliedtoprovideSSTforthefullresolutionLandsat8image.
Ingeneral,thereareapproximatelyfourAVHRRimagesperday.Duetochangingcloudcoverand
orbitconfigurationbetweenavailableAVHRRimages,itwassometimesnecessarytouseanimagemore
distantintime(butlesscloudy)fromtheLandsat8overpass,despiteatemporallymoreproximateone
beingavailable.However,becauseGulfofMaineSSTpatternschangeslowly(lessthan0.4oCover12hours
atbuoy44005,www.neracoos.org),weconsiderthisanacceptabletradeofftomaximizethenumberof
qualityAVHRRpixelsthatwillbeusedintheregression.ThemeanoffsettimebetweentheLandsat8and
AVHRRoverpasseswas6.8hours,withaminimumoffsetof2.3hours,maximumoffsetof30.2hours,and
astandarddeviationof5.8hours.TheentireareaofoverlapbetweenAVHRRoceanpixelsandLandsat8
9
oceanpixelsisusedformostscenes.Landsat8imagesweresubsampledtoevery10thpixelinbothxand
ydimensionstoreducethedatavolumefortheregressions,andAVHRRimageswereresampledtomatch
the30m(interpolatedfrom100m)resolutionoftheLandsatB10usingnearestneighborresamplingin
MATLAB.Dependingonthedistributionofclouds, theregressionareawasrestrictedtoareaswithout
cloudcontamination(orpoorlymaskedclouds)insomeinstances.Cloudandlandweredilatedbytwo
pixels in the AVHRR image to reduce occurrences of cloud ringing artifacts and nearshore land
contamination.Theregressionprocesswas iterative.Aftereach iteration,allLandsat8andcoincident
AVHRRpixelsthatweregreaterthanonestandarddeviationfromthelinearbestfitlineoftherelationship
wereremovedandtheregressionwasre-calculatedwiththeremainingdata.Theiterationprocesswas
repeateduntil thePearsoncorrelationcoefficient for the twodatasetsstabilizedorstarted toworsen
(whichisduetouncertaintiesintheapproach).Thefinalregressionequationwasthenappliedtoeach
Landsat8B10pixelatthefull30mresolutiontoobtainaLandsatSSTimage.
2.3.OceanColor
Oceancolormultispectraldata,whichrespondstotheeffectsofoceanicparticlesanddissolved
matter,aremeasuredfromspacebytheOperationalLandImager(OLI)radiometeronboardLandsat8.
Theradiancemeasuredincludescontributionsfromthetarget(thesurfacewatercolumn),theairwater
interface,andthebackground(particlesandgasesfromnearbypixelsandparticlesintheatmosphere)
(Mobleyetal.,2016).Toobtaininformationontheoceanicconstituents,theatmosphericcontributionto
the signal needs to be removed (a process known as ‘atmospheric correction’ see below). From the
correctedwater-leavingradiance,wecomputedthereflectance(denotedasR,-)fromwhichtheproducts
ofturbidityandChlaarederived.
10
2.4.AtmosphericCorrectionfor𝐑𝐫𝐬
Giventhelowturbidityinourareaofinvestigation(seeSection2.5below),wechosetouseacombination
oftheNearInfrared(NIR)andShortWaveInfrared(SWIR)channelsforatmosphericcorrectioninSeaDAS.
TheNIRwasimportanttousebecauseof itshighersignal/noiseratio(NIRbandshadratiosof6and7
whileSWIRbandshadratiosof9and10)inlowturbiditywaters,andtheSWIRwasimportantbecauseit
hasthestrongestabsorptionforwaterwhichhelpsdifferentiatein-watersedimentsfromatmospheric
aerosolparticles(Franzetal.,2015;Pahlevanetal.,2014).Applyingthisatmosphericcorrectionovera
sceneresultedinaper-pixelcorrection,eachwithitsownangstromcoefficient.Theangstromcoefficient
istheslopeofthespectralaerosolopticalthickness,whichisderivedrelativetoareferencewavelength
(usually443nm/865nmasoutputfromSeaDAS).Weadjustedthiscoefficientbecausetheautomaticper-
pixelretrievalsprovidedbySeaDASresultedinnegativevaluesinseveralfreshwaterareasthatwereblack
bodytargetsforouratmosphericcorrectionschemeandshouldhavenear-zeroorpositiveretrievals.The
primary reason for adjusting the angstrom is that the aerosolmodels used for processing data from
satellitessuchasSeaWiFsandMODIS(Ahmadetal.,2010),donotrepresenttheaerosolconditionsfor
ourstudyarea,thecoastofMaine(Pahlevanetal.,2017).Wethenchoseasingleangstromcoefficient
perscene(fromwithinthedistributionofinvertedangstromvalues),byrequiringthattheminimalvalue
ofR,-(443)inascene,measuredinaveryhumiclake,benearzero.Mostfreshwaterlakesonthecoast
ofMainearehumicandhavehighlevelsofchromophoricdissolvedorganicmatter,CDOM,whichgives
themabrownhueandattenuateslightquickly(RoeslerandCulbertson,2016;Rasmussen,1989).Several
freshwaterlakeswithhighCDOMwithinourstudyregion(MuddyPond,BiscayPond,andDamariscotta
PondcircledinFig.1)wereselectedassuitablereferencetargetstocorrecttheentireLandsat8scene.In
each individualsatellite image,thedarkest lake(whereR,-(443) isnearzero)wasusedtodetermine
angstrom coefficient. Analysis of a sample of water from one of these lakes verified that the
expectedR,-(443) is zero within the uncertainty of the measurement (Appendix Table B1). We
11
subsequentlyappliedtheretrievedangstrominSeaDAStotheentirescenetorecalculateR,-atevery
wavelength.R,-valueswerethenusedtocomputeturbidityandchlorophyll.
2.5.RetrievalofTurbidity
Turbidity,T,(notethat1gL-1ofSPMisequivalent,withintherangeofvaluesfoundinourstudyarea,to
aturbidityof1NTU(PfannkucheandSchmidt,2003))wascalculatedovertheentiresatellitesceneusing
atmosphericallycorrected𝑅"#(655)andthefollowingequationfromNechadetal.(2010):
𝑇 = 𝐴;𝜌'
1 − 𝜌'/𝐶;[𝑔𝑚DE](1)
where 𝜌' = 𝑅"#(655) ∗ 𝜋 and 𝜌'is the atmospherically corrected and derived water leaving
reflectance,𝐴;=289.1and𝐶;=16.86(Nechadetal.,2010).
2.6.RetrievalofChlorophyll-𝐚
Chl𝑎wascalculatedusingthestandardOC3algorithm(O’Reillyetal.,1998)fromtheNASAOceanBiology
ProcessingGroup,usingtheabove-calculated𝑅"#:
logLM 𝑐ℎ𝑙𝑜𝑟_𝑎 = 𝑎M + 𝑎U logLM𝑅"# 𝜆%WXY𝑅"# 𝜆*"YYZ
U[
U\L
(2)
whereaManda_aresensorspecificcoefficients,andR,- λabcd andR,- λe,ddf arethegreatestofvalues
from443>483and555nm,respectively,ontheOLIsensoraboardLandsat8(NASA,2016).(Note:SeaDAS
applies coefficients to convertbroadband Landsat8-basedR,- to11nmnarrowbands forwhich this
equationwasdeveloped).
12
2.7.Validationwithinsitudata
Validationwascarriedoutforphysicalandbiogeochemicalparameters(SST,turbidity,andChl𝑎)using
data from water samples and three oceanographic buoy observing systems. Historical data was
downloaded from the NERACOOS (Northeastern Regional Association of Coastal Ocean Observing
Systems) buoys E01 and I01 operated by the University of Maine, Orono, in the Gulf of Maine, a
Land/OceanBiogeochemicalObservatory(LOBO)buoyatBowdoinCollegeinHarpswellSound,andtwo
LOBO buoys at the University of Maine’s Darling Marine Center in the Damariscotta River (Fig. 1,
NERACOOSBuoyI01notpictured).TheLOBObuoyswereequippedwithsensorsthatremainatadepth
of1.5metersandmaintainedandcleanedtopreventbiofoulingeverytwoweeks.Temperaturedatawere
collectedfromallthreeobservingsystemsandcomparedtoLandsat8SST.Atotalof52matchupswere
identifiedoriginatingfrom31clearoverpassesfrom2013to2016.
Insituturbiditywasusedtovalidatesatellite-derivedturbidityduringeightoverpassesin2015
and 2016. Data were collected from the UMaine LOBO buoys in the Damariscotta River, and were
measuredbyaWETLabsWQMsensorcapableofmeasuringturbidityfrom0–25NTU(thatmeasurelight
scattered in the back direction at a 20 nm bandwidth around 700 nm). This sensor was vicariously
calibratedagainstaHachturbiditysensor(whichisanISO7027:1999turbiditystandard).Thebuoydata
werecorrectedbyaregressionbetweenHachturbiditysamplesandtheLOBOturbiditywithaslopefactor
of1.58(AppendixTableB2).
InsituChl𝑎wasusedtovalidatesatellite-derivedChl𝑎duringthesameeightoverpassesin2015
and2016.InsituChl𝑎dataweremeasuredbytheDamariscottaRiverLOBObuoys’WETLabsfluorescent
sensorcapableofmeasuringChl𝑎concentrationsfrom0–50µgL-1.Thebuoydatawascorrectedbya
regressionbetweenextractedChl𝑎samplesandtheLOBOChlawithaslopefactorof1.71.Watersamples
werecollectedintriplicate,atthesurface,andinopaquebottleswithin30minutesofeachoverpassand
13
filteredforextractiononaTurner10AUfluorometerperstandardprotocol(Holm-HansonandRiemann,
1978).
2.8.SatelliteImageryforanOysterSuitabilityIndex
AnOysterGrowthSuitabilityIndexwasdesignedusingthesatellite-derivedSST,turbidity,andChl𝑎.A
weightingandindexingprocedureofthesethreephysicalparametersdescribedideal,moderate,andpoor
conditions forgrowingmarket sizedoysters in surfaceculture.Thecriteria for the indexwerechosen
basedonpublishedstudiesofenvironmentaleffectsonoystergrowth,recognizingthattheconcentration
oforganicdetritus,knowntobeanimportantcomponentofoysterdiet,wasnotavailable.Temperature
isthemost importantvariable inoystergrowth,especially inthecoldwatersofcoastalMainesince it
controlsthefiltrationrateofoysters(andthereforegivenanimportanceweightfactorof80%;Loosanoff,
1958;Hoffmannetal.,1992;EhrichandHarris,2015).Oysterclearanceofalgaeisapositivefunctionof
algaeconcentration,aslargeamountsofpseudofecesareproducedathighalgalconcentrations.Because
ofthis,weweightedChlaat15%,withpoorconditionsbeinglessthan1μgL-1orgreaterthan10μgL-1,
moderateconditionsbeing1to3μgL-1,andidealconditionsasto10μgL-1(EpifanioandEwart,1977;
Hawkinsetal.,2013).Turbidity,asestimatedbysuspendedparticulatematter,hasanegativeeffecton
oyster feeding at high concentrations, by diluting algal cellswith largely inorganicmatter.Haven and
Morales-Alamo (1966) observed large amounts of pseudofeces production by Eastern oysters at
concentrationsofsuspendedparticulatematterabove10mgL-1,thuswegaveturbidityaweightof5%
anddesignatedpoorconditionsasgreaterthan10μgL-1,moderateconditionsbetween8and10μgL-1
and idealconditionsunder8μgL-1. Hoffmanetal. (1992)alsomodeledoyster filtrationasapositive
functionofwatertemperatureandanegativefunctionofhighsuspendedloads.
Theserelativeweightswerechosenasastartingpointfortheindexandcouldberefinedinfuture
iterations to optimize the index (Gong et al., 2012), by doing a sensitivity analysis of the relative
14
importance of the factors concomitant with growth measurements and growth model outputs. The
resultingOysterSuitabilityIndexisthesumoftheweightedconditionsonascalefrom0to1,wherepixels
withavalueof1representwaterswhereanoysterislikelytogrowtomarketsizewithin2years:
𝑂𝑆𝐼 = 𝑆𝐼U×𝑤U
Z
U\L
(3)
where𝑆𝐼U isthevalueoftheenvironmentalvariablei,𝑤Uistheweightofthevariablei,andnisthenumber
ofenvironmental variables.Wecombined images fromeachyearduring the samemonth to createa
monthlyaveragedindex.Note:thisindexdoesnotincludeinformationaboutsiteclosures,bottomdepth,
orresidentialrestrictions.Futureworkshouldincludethisinformationforamorecomprehensiveindex.
15
CHAPTER3
RESULTS
3.1.Satelliteretrievedvalidationwithinsitudata
TheLandsat8brightnesscalculationcorrelatedwellwithinsitutemperatures(RMSDis0.82°C,RRMSD
is4%,r2=0.94)with,onaverage,1°ChigherSSTvaluesthanthosemeasuredbythebuoysensors,
especiallyinwarmerwaters(Figure2,3).However,variabilityofthebuoymeasurementsislargerat
highertemperatureswhenhorizontalgradientsintemperaturewerealsolarger.
Figure2.Landsat8-derivedSeasurfacetemperaturemapofmid-coastMaineonJuly14,2013.Spatialresolutionis30meters(interpolatedfrom100meters).
16
Figure3.TypeIIlinearregressionformatch-upsbetweenLandsat8seasurfacetemperatureandseasurfacetemperaturemeasuredbyoceanographicbuoys.Differentsymbolsrepresentmeasurementsbythethreedifferentobservingsystems.Verticalerrorbarsarethestandarddeviationabouta5x5pixelboxcenteredattheinsitumeasurement.Horizontalerrorbarsarethestandarddeviationofdailytemperatureateachbuoy.Rootmeansquareerroris0.82°C,rootmeansquarerelativedifferenceis4%,r2=0.94andthegreylineis1:1.
TheLandsat8turbidityestimatescorrelatedwellwithinsituturbidities(RMSD0.49NTU,RRMSD3%,
maxabsolutedeviationis0.96andmaximalrelativedeviationis15%,r2=0.88),withanuncertaintyof
17
0.5NTU,onaverage(Figure4,5).Uncertaintiesarelargerathigherturbiditiesforboththebuoyandthe
satellitealgorithm.
Figure4.Landsat8-derivedturbidityalongmid-coastMaineonJuly14,2013.Spatialresolutionis30meters.
18
Figure5.TypeIIlinearregressionbetweenLandsat8turbidityandturbiditymeasuredbyLOBObuoys.Verticalerrorbarsarethestandarddeviationabouta5x5pixelboxcenteredattheinsitumeasurement.Horizontalerrorbarsarethestandarddeviationofturbidityforfourhoursateachbuoy.Rootmeansquareerrorbetweenthetwodatasetsis0.49NTU,rootmeansquarerelativedifferenceis3%,maxabsolutedeviationis0.96andmaximalrelativedeviationis15%,r2=0.88andthegreylineis1:1.
Landsat-8based chlorophyll didnot correlatewellwith in situ Chla (RMSD is 1.75μgChl L-1,
RRMSDis110%,maxabsolutedeviationis3.14μgChlL-1,maxrelativedeviationis156%,r2=0.31).Below
5μgL-1,theOC3algorithmproducedhigherChl𝑎valuesthanthosemeasuredbythebuoysensors(Figure
6,7).Above5μgL-1,thebuoymeasurementswerehigherthanthesatellite-derivedChl𝑎.Uncertainties
19
arelargerathigherChlaforthebuoysandthesatellitealgorithm.Outofthethreeparametersderived
from Landsat 8, this algorithm has the highest relative deviation of 156%, with an average relative
differenceof110%,whichissignificantlyworsethantheaveragerelativedifferenceof30%forchlorophyll
algorithmsintheopenocean(butseeDiscussion).
Figure6.Landsat8-derivedchlorophyllaalongmid-coastMaineonJuly14,2013.Spatialresolutionis30meters.
20
Figure7.TypeIIlinearregressionbetweenLandsat8chlorophyll-aandchlorophyll-ameasuredbyLOBObuoys.Verticalerrorbarsarethestandarddeviationabouta5x5pixelbox.Horizontalerrorbarsarethestandarddeviationof chlorophyll-a for four hours at eachbuoy. Buoy chlorophyll-awas correctedbychlorophyll extraction samples. Rootmean square error is 1.75 μg Chl L-1, rootmean square relativedifferenceis110%,maxabsolutiondeviationis3.14μgChlL-1,relativedeviationis156%,r2=0.31andthegreylineis1:1.
3.2.SatelliteImageryforOysterGrowthConditions
Monthlymaps of anOyster Suitability Index (Figure 8)were created using averagedmonthly
satelliteimages(AppendixC).Mostexistingoysteraquacultureareas(indicatedbyredstarsonFig.8)fall
withinthehighestsuitability indexduringthemonthofJuly.Areascoloredbrightyellowindicatesites
21
thatareoptimalforfastgrowingjuvenileoysters(hightemperature,lowturbidity,andmoderateChl𝑎).
Areasingreenindicatesitesthataremoderatelysuitedforgrowingoysters,andareasinblueindicate
waters that are least suitable for oyster growth. Suitabilitymaps forAugust and September exhibit a
similarpatternofideal,moderate,andpoorgrowingareasasthemapforJuly(Fig.8),but, ingeneral,
withslightlylowervaluesduetocoldertemperatures(averagemonthlytemperatureswerehighestduring
July).TheOysterSuitabilityIndexmapprovidestwoimportantfindings:1)itconfirmstheDamariscotta
Riverasasuitableplacetogrowoystersinaquacultureandthereforeanimportanttestandverification
siteforusingremotesensingtools,and2)itmapsmanynewlocationsalongthecoastthathostsimilar
conditions(AppendixTableB2).
Figure8.Oystersuitabilitymapbasedonphysicaloceanographicparameters:seasurfacetemperature,turbidity,andchlorophyll-a.MapisanaverageofallimagesinthemonthofJuly.Yellowareasindicateidealconditions,greenareasindicatemoderateconditions,andblueareasindicatepoorconditions.Redstars indicate existing oyster farms. Index criteria is given in Appendix Table C, standard deviation ofaveragedparametersinJulyaregivenasfiguresinAppendixD.
22
CHAPTER4
DISCUSSION
4.1.SatelliteImagery
The correspondence between the Landsat 8 satellite derived products and in situ measurements
demonstratesthecapabilityofgeneratingSST,turbidity,andChlamapsalongthejaggedcoastofMaine.
Whilethesedatashowencouragingresults,thereareseveralfactorsfromourstudythatcouldimprove
thepresentalgorithms.Straylightissuesariseifthetemperaturefromanobjectoutsideofthefieldof
viewof the imageraffectsapixelwithin the fieldofview.Fortunately,mostwateralong thecoastof
Maineisvigorouslytidallymixed(~3mtidalrange),andthusdatafromthecenterofchannelscanbe
usedtoinfertheSSTconcentrationsthroughoutthosechannels(ThorntonandMayer,2015).Withinthe
estuaries, however, a TIRS pixel (which is three times as wide as an OLI pixel) next to landmay be
incorrectlycolder(ifthelandiscolder)orwarmer(ifthelandiswarmer).However,thematch-upswith
temperature and turbidity products suggest adjacency and stray light have not degraded the data
significantly,anddifferencesarelikelyduetonoiseasopposedtosystematicbias.
4.2.LimitationsinValidationProcess
Validationof Landsat8 imagerywith in situmeasurements isnecessary toassess theaccuracyof the
algorithms for retrieving bio-physical products. Someof thediscrepancy between in situ and satellite
matchupscanbeexplained,whileothers require further investigation.Onereason thatLandsat8SST
valuesmaybehigherthanmostbuoymeasurements(Fig.3)isbecausetheSSTestimatescomefromlight
emitted from the top fewmicrometers of the sea surface,while the buoy sensors are located about
1.5mbelowthesurface.Inthedaytimeimages,thesubsurfacewaterislikelycoolerthanthesurface
skindue tophysicalandenvironmental factors (Padulaetal.,2010;Donlonetal.,2002;Ward,2006).
23
Despitethisbias,theLandsat8SST(derivedbyregressingwithAVHRR)performedwellalongthecoastof
Maine(Fig.3)andourresultssuggestthatourapproachcouldbeusedasatoolformeasuringSSTwhere
highspatialresolutionisdesired.
Avigoroussemi-diurnaltidecharacterizestheDamariscottaRiveranddeliversshelfwater into
theupperreachesoftheestuary.Thetidalcyclewasevidentinthedailyturbiditysignal(notshown)from
the LOBObuoys: at low tide, there are elevated levels of turbiditywhereas at high tide there is less
turbidity(duetotheincreaseinturbidityfromthemouthtotheendoftheestuary).Thehorizontalerror
barsinFigure5representthevariabilityduringafour-hourperiodaroundeachsatelliteoverpasstime,
and highlight the importance of simultaneous sampling for in situ - satellitematchups. The turbidity
algorithmperformswellwithinouruncertaintiesinthiscontext.
Landsat 8 Chl a often differs significantly from the LOBO buoy measurements. There are
significantuncertaintiesassociatedwithbothmeasurementtechniques(Cullen,2008).Landsat8Chlais
retrievedfromR,-usinganalgorithmcalibratedintheopenocean,whereastheLOBObuoysmeasure
Chlafluorescencewhichisregressedagainstwatersamples.EstimatingChlafromfluorescenceisthe
mostcommonwaytomeasureChlabutisaffectedbyseveralprocessesthatcontributetouncertainty.
Theseincludechangesinfluorescenceyieldduetovariabilityinthealgaltaxonomy,nutrientstress,and
photo-acclimation,tonameafew(Cullen,1982).Inaddition,concentrationsofphytoplanktonhavebeen
observedintheDamariscottaRivertovaryontimescalesofhours(ThompsonandPerry,2006).
Non-photochemical quenching (NPQ; when phytoplankton decrease their fluorescence at a
maximumlightharvestinglevel,e.g.Cullen,1982)contributestovariability.However,wefindnighttime
measurements to be comparable to daytimemeasurements (Appendix Fig. B.1) for theDamariscotta
River.Therefore,theoffsetinChlaislikelynotduetoerrorsinducedbyNPQ.Anotherpotentialerroris
associatedwiththeOC3algorithm,whichestimatesChlaasaratioof𝑅"#intheblueandgreenchannels.
24
The blue channel is especially affected by colored dissolved organic material (CDOM). Independent
changesofCDOMwillaffecttheOC3chlorophyllestimate(Siegeletal.,2005).AlongthecoastofMaine,
where there are coastal forests andmarshes, CDOM is in high concentration and variable (Roesler&
Culbertson,2016). Incoastalareasandestuaries rich inCDOMit is likely thatabsorptionbydissolved
organicmatterwouldbias theOC3algorithm. It is likely thata localalgorithm that takes localCDOM
concentrationintoaccount,couldimproveChlaretrievalsfromLandsat8.
4.3.OysterSuitabilityIndex
TheOysterSuitabilityIndexprovidedinthispaperisintendedasasupplementtoothertoolsthat
determineoptimaloystergrowingareas.Firstly,thesatelliteimagesprovideonlyaclimatologicalmonthly
snapshotofcoastaltemperatureproducts,whichprovideslesstemporalresolutionthanacomprehensive
daydegreemodelfortemperature-dependentshellfishgrowth.Secondly,moreimportantenvironmental
factorssuchassalinity,waterdepth,bottomtypeandwatervelocity(necessaryforoystergrowing),are
not considered.Organicdetritus is known tobean important componentofbivalvediets (Dameand
Patten,1981;Bayneetal.,1993;Barilleetal.,1997),butcurrentlycannotbemeasuredusingsatellite
imagery.Ourindexthereforeprovidesguidanceonsuitablewaterqualityconducivetorapidgrowth,but
notsufficientinformationtomodelsitespecificproductioncapacityforsuspendedorbottomculture.
Althoughsatellitethermaldataisonlysensitivetothetemperatureofthetopfewmicrometers
ofwater,andoceancolorissensitiveonlytooneopticaldepth(whichvaries,butontheMainecoastis
usuallythetoponeortwometers),thesedataarerelevanttothewholewatercolumnifthewatercolumn
is often vertically well-mixed. Indeed, many estuaries on the Maine coast are well-mixed (e.g. the
SheepscotandMedomakRivers,ThorntonandMayer,2015;Mayer,1996),whichcoincidentlyqualifyfor
oysteraquacultureinoursuitabilityindex(AppendixTableB2).Finally,localknowledgeisinvaluablefor
25
the expansion of an existing industry on the coast of Maine, and stakeholder input is essential for
improvingsuchanindexwithlocalinformationsuchassiteaccessibility,townordinances,etc.
4.4.FutureWork
Continuedsamplingduringthespringandsummerof2017willprovideamorecompletedataset
foroptimizingoceancolorproducts inMaine.A localalgorithmforLandsat8Chlaalongthecoastof
Mainecouldbeconstructedwithadditional insitusamplescollectedduringsatelliteoverpasses.There
areseveralapproachestotunealocalalgorithm.Anempiricalapproach,suchastheOC3algorithm,uses
arelationshipbetweeninsitumeasurementsandratiosofthesatellitesensorbands.Asecondmethod
involvesusingageneralizedinherentopticalpropertiesinversion(GIOP,Werdelletal.,2013).Thismethod
solvesforChla,SPM,andCDOMusingknownspectralshapesofopticalproperties(forphytoplankton
and non-algal absorption and backscattering by particles) and known values of absorbance and
backscatteringofwater(whichareweakfunctionsofsalinityandtemperature).Databasesofcollection
sites locatedintheDamariscottaRiverandHarpswellSoundcouldtunetheshapesofinherentoptical
propertiesfortheGIOPalgorithmandprovideanestimateofChlainthesetwoestuaries.Furthermore,
insitusamplesfromvariouslocationsalongthecoastwillvalidatethelocalalgorithmsothatitsusecan
beexpandedfromtheDamariscottaRivertootherplacesalongthecoast.
ObtainingmoreparametersfromLandsat8,suchascoloreddissolvedorganicmatter(CDOM),
wouldprovideadditional information to growers aswell as environmentalmonitoringandecosystem
managers.Franzetal.,(2015)andSloneckeretal.,(2015)describethepotentialofusingLandsat8for
remotesensingofCDOMinconjunctionwithinsitumeasurements.AreliableCDOMproductwouldalso
improvethealgorithmforChla,asthepresenceofCDOMoftencontributestoanoverestimationofChl
a.Furthermore,highlevelsofCDOMarecorrelatedwithlowsalinityincertainestuariesinMaine(Carder
etal.,1989;D’Saetal.,2002;Mayer,L.,2017personalcommunication).Itwouldbehelpfultoidentify
26
areas with significant freshwater influx because these often bring concentrations of bacteria that
negativelyaffectclammingandotherfisheries(Kleindinstetal.,2014;Shumwayetal.,1988).
27
CHAPTER5
CONCLUSION
OursatellitedataderivedOysterSuitabilityIndexcanactasapowerfultoolforoysteraquaculture
site selection and the expansion of the shellfish farming industry. It shows that suitable biophysical
conditionsforoystergrowthexistinmanyareasoftheMainecoast.Suitabilityindicesforotherbivalve
species,suchasmussels,scallops,andfinfishalongthecoast,orotherapplicationsrequiringhighspatial
resolution,canbedevelopedbasedonthealgorithmspresentedhere.SST,turbidity,andChlaretrieved
fromLandsat8issufficientlyvalidatedbyinsitumatchups(within+/-1°CforSST;maxabsolutedeviation
is0.96NTUand relativedeviation is15% for turbidity;andmaxabsolutedeviation is3μgChl L-1and
relativedeviationis156%forChla).OurresultsshowthatLandsat8imageryisusefulforretrievingSST,
turbidity,andChlaincoastalwatersofMaine,USA,andcanbeappliedtoothernarrowestuariesaround
theworld.ThenoveltyofusingLandsat8inthiscontextoffersauniqueopportunitytomapandmonitor
coastal waters at an unprecedented spatial resolution. Inclusion of data from other satellites with
complimentarysensorsuitessuchasSentinel2A,andtherecentlylaunchedSentinel2B,couldimprove
boththespatialandtemporalcoverageofcoastalwaters,astheywillprovidefive-dayorbettercoverage
(unfortunately,Sentinel2AandBdonothavethermalbands).SSTdatafromLandsat8isespeciallyuseful
for aquaculture site prospecting. We recommend adding thermal bands to future high resolution
missions,asmorefrequentSSTdatawillassistbothsiteselectionforaquacultureandotherapplications.
Futureworkimprovingbiogeochemicallocalalgorithms,refiningtheatmosphericcorrection,andadding
other parameters such as CDOM,will further advance the use of high resolution remote-sensing for
coastalapplications.
28
REFERENCES
Ahmad,Z.,Franz,B.a,McClain,C.R.,Kwiatkowska,E.J.,Werdell,J.,Shettle,E.P.,&Holben,B.N.(2010).Newaerosolmodelsfortheretrievalofaerosolopticalthicknessandnormalizedwater-leavingradiancesfromtheSeaWiFSandMODISsensorsovercoastalregionsandopenoceans.AppliedOptics,49(29),5545.
AguilarManjarrez,J.&Crespi,V.(2013).FAOAquacultureNewsletter55,September2016.
Barillé,L.,Prou,J.,Héral,M.andRazet,D.,(1997).Effectsofhighnaturalsestonconcentrationsonthefeeding,selection,andabsorptionoftheoysterCrassostreagigas(Thunberg).Journalofexperimentalmarinebiologyandecology,212(2),pp.149-172.
Barnes,R.A.,Holmes,A.W.,&Esaias,W.E.(1995).StrayLightintheSeaWiFSRadiometer.SeaWiFSTechnicalReportSeries,31(July),79.
Barnes,T.K.,Volety,A.K.,Chartier,K.,Mazzotti,F.J.andPearlstine,L.,(2007).Ahabitatsuitabilityindexmodel for theeasternoyster (Crassostreavirginica), a tool for restorationof theCaloosahatcheeEstuary,Florida.JournalofShellfishResearch,26(4),pp.949-959.
Bayne,B.L.,Iglesias,J.I.P.,Hawkins,A.J.S.,Navarro,E.,Heral,M.andDeslous-Paoli,J.M.,(1993).Feedingbehaviourofthemussel,Mytilusedulis:responsestovariationsinquantityandorganiccontentoftheseston.JournaloftheMarineBiologicalAssociationoftheUnitedKingdom,73(04),pp.813-829.
Buitrago,J.,Rada,M.,Hernández,H.andBuitrago,E.,(2005).Asingle-usesiteselectiontechnique,usingGIS,foraquacultureplanning:choosinglocationsformangroveoysterraftcultureinMargaritaIsland,Venezuela.EnvironmentalManagement,35(5),pp.544- 556.
CakeJr.,E.W.,(1983).Habitatsuitabilityindexmodels:GulfofMexicoAmericanoyster(No.82/10.57).U.S.FishandWildlifeService.
Carrasco,M.F.andBarón,P.J., (2010).Analysisof thepotentialgeographic rangeof thePacificoysterCrassostrea gigas (Thunberg, 1793) based on surface seawater temperature satellite data andclimatecharts:thecoastofSouthAmericaasastudycase.BiologicalInvasions,12(8),pp.2597-2607.
Carder,K.L.,Steward,R.G.,Harvey,G.R.,&Ortner,P.B.(1989).Marinehumicandfulvicacids:Theireffectsonremotesensingofoceanchlorophyll.LimnologyandOceanography,34(l),68–81.http://doi.org/10.4319/lo.1989.34.1.0068
Congleton,W.R.,Pearce,B.R.,Parker,M.R.andBeal,B.F.,(1999).Mariculturesiting:aGISdescriptionofintertidalareas.EcologicalModelling,116(1),pp.63-75.
Cullen,J.J.(1982).Thedeepchlorophyllmaximum:Comparingverticalprofilesofchlorophylla.CanadianJournalofFisheriesandAquaticSciences,39(5),791-803.doi:10.1139/f82-108.
29
Cullen,J.J.,(2008).Observationandpredictionofharmfulalgalblooms.InM.Babin,C.S.RoeslerandJ.J.Cullen[eds.],Real-timecoastalobservingsystemsforecosystemdynamicsandharmfulalgalblooms,UNESCO.
Dame,R.F.andPatten,B.C.,(1981).Analysisofenergyflowsinanintertidaloysterreef.Mar.Ecol.Prog.Ser,5(2),pp.115-124.
D'Sa,E.,Hu,C.,Muller-Karger,F.,&Carder,K.(2002).Estimationofcoloreddissolvedorganicmatterandsalinityfieldsincase2watersusingSeaWiFS:Examplesfrom FloridaBayandFloridaShelf.JournalofEarthSystemScience,111(3)197-207.
Dogliotti,A.I.,Ruddick,K.G.,Nechad,B.,Doxaran,D.,&Knaeps,E.(2015).Asinglealgorithmtoretrieveturbidityfromremotely-senseddatainallcoastalandestuarinewaters.RemoteSensingofEnvironment,156,157–168.http://doi.org/10.1016/j.rse.2014.09.020
Donlon,C.J.,Minnett,P.J.,Gentemann,C.,Nightingale,T.J.,Barton,I.J.,Ward,B.,&Murray,M.J.(2002).Towardimprovedvalidationofsatelliteseasurfaceskintemperaturemeasurementsforclimateresearch.JournalofClimate,15(4),353–369.http://doi.org/10.1175/1520-0442(2002)015<0353:TIVOSS>2.0.CO;2
Ehrich,M.K.,&Harris,L.A.(2015).Areviewofexistingeasternoysterfiltrationratemodels.EcologicalModelling,297,201–212.http://doi.org/10.1016/j.ecolmodel.2014.11.023
Epifanio,C.E.andEwart,J.,1977.MaximumrationoffouralgaldietsfortheoysterCrassostreavirginicaGmelin.Aquaculture,11(1),pp.13-29.
Estapa,M.L.,Boss,E.,Mayer,L.M.,&Roesler,C.S.(2012).Roleofironandorganiccarboninmass-specificlightabsorptionbyparticulatematterfromLouisianacoastalwaters.LimnologyandOceanography,57(1),97–112.http://doi.org/10.4319/lo.2012.57.1.0097
Franz,B.a.,Bailey,S.W.,Kuring,N.,&Werdell,P.J.(2015).OceancolormeasurementswiththeOperationalLandImageronLandsat-8:implementationandevaluationinSeaDAS.JournalofAppliedRemoteSensing,9(1),96070.http://doi.org/10.1117/1.JRS.9.096070
Gong,C.,Chen,X.,Gao,F.,&Chen,Y.(2012).Importanceofweightingformulti-variablehabitatsuitabilityindexmodel:Acasestudyofwinter-springcohortofOmmastrephesbartramiiintheNorthwesternPacificOcean.JournalofOceanUniversityofChina,11(2),241–248.http://doi.org/10.1007/s11802-012-1898-6
Haven,D.S.andMorales-Alamo,R.,1966.Aspectsofbiodepositionbyoystersandotherinvertebratefilterfeeders.Limnol.Oceanogr,11(4),pp.487-498.
Hawkins,a.J.S.,Pascoe,P.L.,Parry,H.,Brinsley,M.,Cacciatore,F.,Black,K.D.,…Zhu,M.Y.(2013).ComparativeFeedingonChlorophyll-RichVersusRemainingOrganicMatterinBivalveShellfish.JournalofShellfishResearch,32(3),883–897.http://doi.org/10.2983/035.032.0332
30
Hawkins,A.J.S.,Pascoe,P.L.,Parry,H.,Brinsley,M.,Black,K.D.,McGonigle,C.,Moore,H.,Newell,C.R.,O'Boyle,N.,Ocarroll,T.andO'Loan,B.,(2013).Shellsim:agenericmodelofgrowthandenvironmentaleffectsvalidatedacrosscontrastinghabitatsinbivalveshellfish.JournalofShellfishResearch,32(2),pp.237-253
Hoffmann,E.E.,Powell,E.N.,Klinck,J.M.andWilson,E.A.(1992).ModelingoysterpopulationsIII.Criticalfeedingperiods,growth.JournalofShellfishResearch,11(2),pp.399-416.
Holm-Hansen,O.,andB.Riemann.(1978).Chlorophylladetermination:improvementsinmethodology.Oikos,30:438-447.
Hudson,A.S.,Talke,S.A.,&Jay,D.A.(2016).UsingSatelliteObservationstoCharacterizetheResponseofEstuarineTurbidityMaximatoExternalForcing.EstuariesandCoasts,1–16.http://doi.org/10.1007/s12237-016-0164-3
Kleindinst,J.L.,Anderson,D.M.,McGillicuddy,D.J.,Stumpf,R.P.,Fisher,K.M.,Couture,D.A.,…Nash,C.(2014).CategorizingtheseverityofparalyticshellfishpoisoningoutbreaksintheGulfofMaineforforecastingandmanagement.Deep-SeaResearchPartII:TopicalStudiesinOceanography,103,277–287.http://doi.org/10.1016/j.dsr2.2013.03.027
Gernez,P.,Lerouxel,A.,Mazeran,C.,Lucas,A.,&Barill,L.(2014).Remotesensingofsuspendedparticulatematterinturbidoyster-farmingecosystems.JournalofGeophysicalResearch:Oceans,7277–7294.http://doi.org/10.1002/2014JC010055.
Loosanoff, V.L., (1958). Someaspectsof behavior of oysters at different temperatures.TheBiologicalBulletin,114(1),pp.57-70.
Mayer,L.M.,UniversityofMaineatOrono.,&UniversityofMaine/UniversityofNewHampshireSeaGrantCollegeProgram.(1996).TheKennebec,SheepscotandDamariscottaRiverEstuaries:Seasonaloceanographicdata.Orono,Me:Dept.ofOceanography,UniversityofMaine.
McAlice,B.J.(1977).ApreliminaryoceanographicsurveyofDamariscottaRiverEstuary,LincolnCounty,Maine.Washington:U.S.NationalOceanicandAtmosphericAdministration.
Mobley,C.D.,Werdell,J.,Franz,B.,Ahmad,Z.,&Bailey,S.(2016).AtmosphericCorrectionforSatelliteOceanColorRadiometryATutorialandDocumentationNASAOceanBiologyProcessingGroup.
Montanaro,M.,Gerace,A.,Lunsford,A.,&Reuter,D.(2014).Straylightartifactsinimageryfromthelandsat8thermalinfraredsensor.RemoteSensing,6(11),10435–10456.http://doi.org/10.3390/rs61110435
NASA.(2016).https://oceancolor.gsfc.nasa.gov/atbd/chlor_a/
Nechad,B.,Ruddick,K.G.,&Park,Y.(2010).Calibrationandvalidationofagenericmultisensoralgorithmformappingoftotalsuspendedmatterinturbidwaters.RemoteSensingofEnvironment,114(4),854–866.http://doi.org/10.1016/j.rse.2009.11.022
31
Newell,R.I.E.andJordan,S.J.,(1983).PreferentialingestionoforganicmaterialbytheAmericanoysterCrassostreavirginica.Marineecologyprogressseries.Oldendorf,13(1),pp.47-53.
Newell,C.R.,S.Shumway,T.L.CucciandR.Selvin.(1989).Theeffectsofnaturalsestonparticlesizeand typeon feeding rates, feeding selectivity and food resource availability for themusselMytilusedulisLinnaeus,1758atbottomculturesitesinMaine.J.Shellfish.Res.8:187-196.
Newell, C.R., A.J.S. Hawkins, K. Morris, J. Richardson, C. Davis and T. Getchis. (2013).ShellGIS:adynamictoolforshellfishfarmsiteselection.WorldAquaculture.44:50-53.
O’Reilly, J.E.,Maritorena,S.,Mitchell,B.G.,Siegel,D.A.,Carder,K.L.,Garver,S.A.,…McClain,C.R.(1998).Oceancolor chlorophyll algorithm forSeaWiFS. JournalofGeophysicalResearch,103(Cll),24937–24953.
Pahlevan,N.,Lee,Z.,Wei,J.,Schaaf,C.B.,Schott,J.R.,&Berk,A.(2014).On-orbitradiometriccharacterizationofOLI(Landsat-8)forapplicationsinaquaticremotesensing.RemoteSensingofEnvironment,154,272–284.http://doi.org/10.1016/j.rse.2014.08.001
Pahlevan,N.,Sarkar,S.,&Franz,B.A.(2016).Uncertaintiesincoastaloceancolorproducts:Impactsofspatialsampling.RemoteSensingofEnvironment,181,14–26.http://doi.org/10.1016/j.rse.2016.03.022
Pahlevan,N.,Sheldon,P.,Peri,F.,Wei,J.,Shang,Z.,Sun,Q.,…Loveland,T.(2016).Calibration/validationoflandsat-derivedoceancolourproductsinBostonharbour.InternationalArchivesofthePhotogrammetry,RemoteSensingandSpatialInformationSciences-ISPRSArchives,41(July),1165–1168.http://doi.org/10.5194/isprsarchives-XLI-B8-1165-2016
Pérez,O.M.,Ross,L.G.,Telfer,T.C.anddelCampoBarquin,L.M.,(2003).WaterqualityrequirementsformarinefishcagesiteselectioninTenerife(CanaryIslands):predictivemodellingandanalysisusingGIS.Aquaculture,224(1),pp.51-68.
Pérez-Camacho,A.,Aguiar,E.,Labarta,U.,Vinseiro,V.,Fernández-Reiriz,M.J.andÁlvarez-Salgado,X.A.,(2014).Ecosystem-basedindicatorsasatoolformusselculturemanagementstrategies.EcologicalIndicators,45,pp.538-548.
Pfannkuche,J.,&Schmidt,A.(2003).Determinationofsuspendedparticulatematterconcentrationfromturbiditymeasurements: Particle size effects and calibration procedures.Hydrological Processes,17(10),1951–1963.http://doi.org/10.1002/hyp.1220
Powell,E.N.,Hofmann,E.E.,Klinck,J.M.andRay,S.M.,(1992).Modelingoysterpopulations1.Acommentaryonfiltrationrate.Isfasteralwaysbetter?J.ShellfishRes.,1I:387-398.
Radiarta,I.N.,Saitoh,S.I.,&Miyazono,A.(2008).GIS-basedmulti-criteriaevaluationmodelsforidentifyingsuitablesitesforJapanesescallop(Mizuhopectenyessoensis)aquacultureinFunkaBay,southwesternHokkaido,Japan.Aquaculture,284(1–4),127–135.
32
Rasmussen,J.B.,Godbout,L.,&Schallenburg,M.(1989).Thehumiccontentoflakewateranditsrelationshiptowatershedandlakemorphometry.LimnologyandOceanography,34(7),1336-1343.
Rheault,R.B.,&Rice,M.A.(1996).Food-limitedgrowthandconditionindexintheeasternoyster,Crassostreavirginica(Gmelin1791),andthebayscallop,Argopectenirradiansirradians(Lamarck1819).JournalofShellfishResearch,15(2),
Roesler,C.,&Culbertson,C.(2016).LakeTransparency :AWindowintoDecadalVariationsinDissolvedOrganicCarbonConcentrationsinLakesofAcadiaNational.http://doi.org/10.1007/978-3-319-30259-1
Shumway,S.E.,Sherman-Caswell,S.,&Hurst,J.(1988).ParalyticshellfishpoisoninginMaine:monitoringamonster.JournalofShellfishResearch.
Siegel,D.A.,Maritorena,S.,Nelson,N.B.,&Behrenfeld,M.J.(2005).Independenceandinterdependenciesamongglobaloceancolorproperties:Reassessingthebio-opticalassumption.JournalofGeophysicalResearchC:Oceans,110(7),1–14.http://doi.org/10.1029/2004JC002527
Slonecker,E.T.,Jones,D.K.,&Pellerin,B.A.(2015).ThenewLandsat8potentialforremotesensingofcoloreddissolvedorganicmatter(CDOM).MarinePollutionBulletin,107(2),518–527.http://doi.org/10.1016/j.marpolbul.2016.02.076
Soniat,T.M.andBrody,M.S.,(1988).FieldvalidationofahabitatsuitabilityindexmodelfortheAmericanoyster.EstuariesandCoasts,11(2),pp.87-95.
Thomas, A., Byrne, D., andWeatherbee, R. (2002). Coastal sea surface temperature variability fromLandsatinfrareddata.RemoteSens.Environ.81,262–272.
Thomas,Y.,Mazurié,J.,Alunno-Bruscia,M.,Bacher,C.,Bouget,J.F.,Gohin,F.,…Struski,C.(2011).Modellingspatio-temporalvariabilityofMytilusedulis(L.)growthbyforcingadynamicenergybudgetmodelwithsatellite-derivedenvironmentaldata.JournalofSeaResearch,66(4),308–317.http://doi.org/10.1016/j.seares.2011.04.015
Thompson,B.P.,&Perry,M.J.(2006).TemporalandSpatialVariabilityofPhytoplanktonBiomassintheDamariscottaRiverEstuary,Maine,USA.DepartmentofMarineSciences,Masterof.
Thornton,K.,&Mayer,L.(2015).MaineCoastalObservingAlliance,SummaryReport2014.
Tissot,C.,Brosset,D.,Barillé,L.,LeGrel,L.,Tillier,I.,Rouan,M.andLeTixerant,M.,(2012).Modelingoysterfarmingactivitiesincoastalareas:agenericframeworkandpreliminaryapplicationtoacasestudy.CoastalManagement,40(5),pp.484-500.
USGS.(2016).L8OLI/TIRSproduct.https://earthexplorer.usgs.gov
Vanhellemont,Q.,&Ruddick,K.(2014).TurbidwakesassociatedwithoffshorewindturbinesobservedwithLandsat8.RemoteSensingofEnvironment,145,105–115.
33
Wang,H.,Hladik,C.M.,Huang,W.,Milla,K.,Edmiston,L.,Schalles,J.F.,…Edmiston,L.(2010).Detectingthespatialandtemporalvariabilityofchlorophyll-aconcentrationandtotalsuspendedsolidsinApalachicolaBay,FloridausingMODISimagery.InternationalJournalofRemoteSensing,1161(December2016).http://doi.org/10.1080/01431160902893485
Wang,P.,Boss,E.S.,&Roesler,C.(2005).Uncertaintiesofinherentopticalpropertiesobtainedfromsemianalyticalinversionsofoceancolor.AppliedOptics,44(19),4074–4085.http://doi.org/10.1364/AO.44.004074
Ward,B.(2006).Near-surfaceoceantemperature.JournalofGeophysicalResearch:Oceans,111(2),1–18.http://doi.org/10.1029/2004JC002689
Widdows,J.,Fieth,P.andWorrall,C.M.,(1979).Relationshipsbetweenseston,availablefoodandfeedingactivityinthecommonmusselMytilusedulis.MarineBiology,50(3),pp.195-207.
34
APPENDIXA:ASSESSMENTOFATMOSPHERICCORRECTION
WatersampleswithhighCDOM(measuredwithaCary-50)andlowturbidity(measuredwitha
Hach-3)wereusedtoverifytheselectionofafreshwaterpondasatargetwithinsignificantreflectancein
thebluefollowinganatmosphericcorrection.
Beginning with a relationship between 𝑅"#(443), and the absorption and backscattering
coefficients (e.g. Wang et al., 2005) of water (subscript w), dissolved substances (subscript g) and
particulatesubstances(subscriptp):
𝑅"# 443 = 0.095 ∗𝑏𝑏𝑤(443)+𝑏𝑏𝑝(443)
∗
𝑎𝑤+𝑎𝑔(443)+𝑎𝑝∗(443)+𝑏𝑏𝑤(443)+𝑏𝑏𝑝(443)∗ (A1)
Wemeasuredduringthesummerof2016theabsorptioncoefficientandturbidityoftwohumic
ponds(BiscayandMuddyponds,TableA1).Togetherwithvaluesofwaterabsorptionandbackscattering
fromtheliterature(TableA2)andrelationshipbetweenparticulatepropertiesandturbidity,wederived
reflectancevalues(TableA1)thatarenotsignificantlydifferentfromzerogivenLandsat8signaltonoise
ratio(Pahlevanetal.,2016).
TableA1MeasuredValuesinHumicPond
Variable BiscayPond MuddyPond
𝑎* 443 [m-1] 4.4 7.0
𝑇[NTU] 2.3 8.7
𝑅"# 443 [sr-1] 7.29*10-4 4.56*10-4
35
TableA2Valuesfromliteratureforequation(A1)
absorptionofwater 𝑎'(443) 0.006(Sullivanetal.2006,Masonetal.2016)
massspecificabsorption 𝑎&∗(443) 0.06m2/g(Estapaetal.,2012,Figure3)
backscatteringofinorganicparticles
𝑏%&∗(443) 0.034(𝑏%&=(0.03)𝑇;Twardowskietal.,2001)
backscatteringofwater 𝑏%'(443) 0.003(Zhangetal.2009)
TableA3
DilutionseriesofArizonaDuststandardwithHachandLOBOWQMturbiditymeasurements.
AZdustadded[ml] Hach[NTU] BuoySensor[NTU]
0 2.26 0.85
2 4.39 1.77
4 7.29 2.87
7 10.43 4.38
15 18.93 6.46
36
FigureA1.TypeIIlinearregressionbetweenLandsat8chlorophyll-aandchlorophyll-ameasuredbyLOBObuoysatnighttime.Verticalerrorbarsarethestandarddeviationabouta5x5pixelbox.Horizontalerrorbars are the standard deviation of chlorophyll for four hours at each buoy. Buoy chlorophyll-a wascorrectedbychlorophyllextractionsampleswithaslopefactoror1.71.Rootmeansquareerroris1.0μgChll-1,r2=0.31andthegreylineis1:1.
37
APPENDIXB.OYSTERSUITABILITYINDEX
TableB1
CriteriaforOysterSuitability Index.Theweightsareadditive,exceptwhenat leastoneparameterhaspoorconditions,inwhichcasetheentirecriteriaisthenmultipliedbyzero.
Physicalparameter
Idealconditions(1)
Moderateconditions(.6)
Poorconditions(0)
Importancefactor
SST[˚C] SST>22 22>SST>20 20>SST 0.8
Turbidity[NTU] 8>Turbidity 10>Turbidity>8 Turbidity>10 0.05
Chla[μg/l] 10>Chla>3 3>Chla>11>Chla
Chla>100.15
TableB2
OysterSuitabilityIndexscoresandaverageJulySSTatexistingandprospectiveoysteraquaculturesitesinMaine.
Upper
DamariscottaRiver
MedomakMaquoit
BayNew
Meadows
UpperSheepscott
River
CousinsIsland
OSIscore 0.94 0.90 1.0 0.84 0.84 0.78
SST[˚C] 24 22 23 24 22 22
38
APPENDIXC.AVERAGEDMONTHLYSATELLITEDATA
FigureC.1.SeasurfacetemperaturederivedfromLandsat8dataaveragedoverallimagesinJulyfrom2013to2016.
41
APPENDIXD.STANDARDDEVIATIONOFMONTHLYCLIMATOLOGYMAPS
FigureD.1.StandarddeviationofmonthlyaveragedseasurfacetemperaturedatainJulyfrom2013to2016.Highervariabilitynearrivermouthsindicatedifferencesintemperatureduetoriverineoutput.Darkareasmaybepoorlymaskedcloudsoratmosphericartifacts.
42
FigureD.2.StandarddeviationofmonthlyaveragedturbiditydatainJulyfrom2013to2016.Highvariabilityinthebottomleftcornerrevealastripingeffectinoneofthefourcompiledimages.Darkareasmaybepoorlymaskedcloudsoratmosphericartifacts.
43
FigureD.3.StandarddeviationofmonthlyaveragedchlorophylladatainJulyfrom2013to2016.Variabilityofchlorophyllaintheupperestuariesishigherthanvariabilityofchlorophyllaoffshore.
44
BIOGRAPHYOFTHEAUTHOR
JordanSnyderwasborninAnaheim,CaliforniaonDecember13,1990.Shewasraisedin
HuntingtonBeach,CaliforniaandgraduatedfromHuntingtonBeachHighSchoolin2009.Sheattended
theUniversityofCalifornia,Davisandgraduatedin2013withaBachelor’sdegreeinGeology.She
movedtoMaineandenteredtheOceanographygraduateprogramatTheUniversityofMaineinthe
summerof2015.Sheenjoyssurfing,SCUBAdiving,running,hiking,camping,gardening,cookingand
laughing.JordanisacandidatefortheMasterofSciencedegreeinOceanographyfromtheUniversityof
MaineinAugust2017.
top related