The mussel path – Using the contaminant tracer, Ecotracer, in Ecopath to model the spread of pollutants in an Arctic marine food web

Paper II: Larsen L-H, Sagerup K, Ramsvatn S (2016) The Mussel Path - Using the contaminant tracer, Ecotracer, in Ecopath to model the spread of pollutants in an Arctic marine food web.

Ecological modelling 331:77-85. http://dx.doi.org/10.1016/j.ecolmodel.2015.10.011

EcologicalModelling331(2016)77–85

ContentslistsavailableatScienceDirect

Ecological Modelling

j o u r n al ho me p ag e :w w w . e l s e v i e r . c o m / l o c a t e / e c o l m o d e l

The mussel path – Using the contaminant tracer, Ecotracer, in Ecopath to model the spread of pollutants in an Arctic marine food web

Lars-Henrik Larsen

, Kjetil Sagerup, Silje Ramsvatn

Akvaplan-niva,FramCentre,Postbox6606,Langnes,9296Tromsø,Norway

a r t i c l e i n f o

Keywords:

PAHs PechoraSea Walrus Food-web Ecotoxicology Modelling

a b s t r a c t

Asthepolaricecapisreceding,shippingintheArcticseasbecomeseasier,andbothdestinationand Atlantic–Pacifictransitshippingisexpectedtoincrease.Thereby,theriskofaccidentsincrease.Immediate negativeimpactsareexpectedfromoilspillsthroughtheacutemortalityformarineorganisms,especially fromheavyfueloil(HFO).MarineDieseloil(MDO)isthereforesuggestedasapreferablefuelforships operatinginArcticwaters.However,PolycyclicAromaticHydrocarbons(PAHs)aretoxiccomponentsin bothtypesoffuel,arehighlybioavailableandcantransferupthefoodchain.AspillofMDOfollowing ashipwreckcouldthereforehaveimpactsbeyondthespillsiteandlongafterthedieselhasspread andevaporated.WemodelthespreadofPAHsfromafictitiousspillofMDOinthePechoraSea(South EasternBarentsSea)usingthecontaminanttracermoduleEcotracer,intheEcopathmodellingsoftware.

Weaddresstheeffectsonthefood-webincludinglongtermeffectsbycombiningtoxicologyandfood- webmodelling.Ecotracerassumesthatpollutantsfollowthebiomasspassivelythroughthesystem,and degradationofpollutantsisfollowinguserspecifiedrates.Bycombininginnaturameasurementsof PAHsinseawaterandinbluemussels(Mytilusedulis)recordedatanaccidentalMDOspillsite,with experimentsconductedontheredkingcrab(Paralithodescamtschaticus)andbluemussels,wederived valuesasinputsintothemodel.TheEcotracerpredictedthatthepollutioninthemusselswillspread throughoutthefood-web,especiallytothetoppredatorsofmussels,kingeider(Somateriaspectabilis) andAtlanticwalrus(Odobenusrosmarusrosmarus)andalsofromsnowcrab(Chionoecetesopilio)toseals andtoothedwhales.

©2015ElsevierB.V.Allrightsreserved.

1. Introduction

ThePechoraSea(67–71 ^◦N,44–60 ^◦E)in theRussian Arctic issituated inthesouth-easternBarentsSea(Fig.1)andis con- sideredtobeaseparateseaareabecauseofmarkeddifferences inenvironmentalconditionscomparedtotherestoftheBarents Sea. The area is an important spawning ground for Arcticfish andisrichinseabirdsandmammalsthatfeedonbenthicinver- tebrates(Boltunov et al., 2010). The Atlantic walrus(Odobenus rosmarus rosmarus) occur in the Pechora Sea, and one of very fewpopulationestimatesindicatesasummerpopulationof3943 animals(Lydersenetal.,2012).Walrusfeedonbenthicorganisms inshallowwaters,andhauloutoneitherlowelevationbeaches, orontheseaice.ThePechoraseaisidentifiedasanareaofhigh ecologicalimportancebytheArcticMonitoringandAssessment Programme(AMAP/CAFF/SDWG,2013)basedoncriteriasetbythe

∗Correspondingauthor.Tel.:+4748114233.

E-mailaddress:[email protected](L.-H.Larsen).

International Maritime Organization (IMO, 2006).The coastline includesmanylow-levelmarshesandwetlandswhichareexposed tofrequentandlong-termfloodingduringspringandsummer,as welltheabrasiveeffectsofseaice.Muddy,shallowwatercoasts arecharacterizedbyhighabundancesofmussels.Inthecaseofan oil-spill,shallow,softsedimentareasareknowntoaccumulateoil, asseen,forexample,afteraspillonthecoastofMassachusetts, USAin1969(Culbertsonetal.,2008).

Asthepolarseaicecaprecedes,shippingandindustrialactiv- itiesbecomeeasierintheArcticseas.Increasedactivitiesmean increasedriskof accidents.ThePechoraSeais oneoftheareas expectedtoholdlargesub-seabeddepositsofhydrocarbons.The firstoffshoreoilfieldinthePechoraSea,Prirazlomnoye,60kmoff theSiberiancoast(Fig.1)atawaterdepthof20metresstarted productionin2014.TheoilfromthePrirazlomnayainstallationis exportedbyicestrengthenedtankers.

This modelling exercise is based on a fictitious ship wreck (Larsenet al.,in press)where thecontainervesselMV“Oleum”

suffersanenginemalfunctionandrunsagroundonthenorthwest- erncorner ofVaygach (Fig.1), thereby releasingapproximating 200tonnesofMDO.Toassesstheenvironmentalimpactofanoil http://dx.doi.org/10.1016/j.ecolmodel.2015.10.011

0304-3800/©2015ElsevierB.V.Allrightsreserved.

78 L.-H.Larsenetal./EcologicalModelling331(2016)77–85

Fig.1.LocationandseabedtopographyofthePechoraSea,thesouth-easternpartoftheBarentsSea.Redlinesindicatemodelboundary.Modelareaisapproximately 125000km².Fictionalwrecksiteof“Oleum”isindicated.

spillwe are combiningecotoxicology and ecologicalmodelling.

Weapplyresultsfromecotoxicologystudiesdoneinourlabora- tory(MDOexposureofbluemussels(Mytilusedulis)andredking crab(Paralithodescamtschaticus)(Sagerupetal.,unpublisheddata), andverifyitsrealismbyapplyingmeasurementsperformedupon anaccidentalreleaseof180tonnesofMDOinSkjervøyharbour, NorthernNorway(70N,20E)inDecember2013(Sagerupand Geraudie,2014).

1.1. ToxicologyofPAHsandfoodwebmodelling

Polycyclicaromatichydrocarbons(PAHs)arelipophilichydro- carbon components and may therefore be susceptible to bioaccumulation (NB not biomagnification). PAH exposure is associatedwithnarcosis,mutagenicity,carcinogenicity,embryo- toxicity,genotoxicity,cellulardamage,endocrinedisruptionand reducedsurvivaloflarvalfish(Mooreetal.,1989;Baussantetal., 2009;Bechmann etal., 2010; Nahrganget al.,2010).PAHs are suspectedtoberesponsibleforseveralofthebiologicalimpacts recordedafterthe1989ExxonValdezspillinAlaska,USA,such asincreasedeggmortality,andreducedsurvivalratesandgrowth inpinksalmon(Oncorhynchusgorbuscha)(Petersonetal.,2003).In molluscs,populationeffectssuchasreducedrecruitment,increased mortalityandreducedproductionhavebeenshowninthefieldfol- lowingexposuretopetroleumhydrocarbonsinthesandgaper(Mya arenaria)(Gilfillan and Vandermeulen, 1978)and in mesocosm studiesforbluemussels(BakkeandSørensen,1985;Widdowsetal., 1985).

Toxicologicalmodelsgenerallyfocusonthedynamicsofthe chemical,thekineticsandthemodelmaybelimitedtothefate withinoneorganism.Thephysiologicalresponsesareextremely complex and therefore the food-web must be simplified (e.g., Thomann,1989).However,combiningecotoxicologyandfood-web

modellingisimportanttobeabletoaddresseffectsofpollutantsat anecosystemlevel.Ecopathhasbeenusedtomodelthespread and accumulationof pollutantsor toxins, e.g.,thefateof diox- ins(Carreretal.,2000),andtocomparehowecosystemstructure dictatesmercuryconcentration(FerrissandEssington,2014),how- everbothofthesestudieschosetonotusetheEcotracermodule.

BoothandZeller(2005)assessedtheimplicationsofmercuryaccu- mulationforhumanhealthusingtheEcotracermodule.Ourwork focusesontestingwhetherwecanuseEcotracerforamajorsin- gledischargeofpollutants,exemplifiedbyaMDOspillfromaship wreck.

ThemaininterestofthisworkistotestandevaluatehowEco- tracerworks.IsEcotracerabletomodelthespreadofPAHsinan Arcticmarinefoodwebatlevelslikelytooccurafteranaccidental spill?Whataremethodologicalchallengesanddespitechallenges, whatcanwelearnfromusingEcotracer?Toanswertheseques- tions,weinvestigatetheeffectsofMDOthatcontainbioavailable PAHsonthemarineenvironment,and assessthespread inthe food-web.

2. Materialsandmethods

AfictitiousscenariodescribingacargoshipvoyagefromHam- burg (Germany) to Yamburg (Russia) forms the basis for our modellingexercise.Thecasestudy(Larsenetal,inpress)describes therescueoperationandoutlinespotentialenvironmentaleffects fromthelossofMDOandcargofromrupturingcontainers.

To investigate the sensitivities of organisms representing ecosystemsalongcurrentandfutureArcticshippingroutes,labora- toryexperimentswithexposuretoMDOwereperformed(Sagerup etal.,unpublisheddata).Twospeciesofmussel,theIcelandicscal- lop(Chlamysislandica)andbluemussel,wereexposedtodispersed MDO.Aspartoftheexperimenttheexposure,trophictransferand

L.-H.Larsenetal./EcologicalModelling331(2016)77–85 79

recoverywerestudiedinredkingcrab.Impactsofdischargesof MDOwerealsostudiedinnatura,aftertheaccidentalreleaseof 180000lofMDOintheharbourofSkjervøy,14thDecember2013 (SagerupandGeraudie,2014).Bluemusselswereusedtoassess uptakeafterthespillbyanalysinglocalmusselsand byplacing uncontaminatedspecimensincagesintheharbourfivedaysafter thespill.TotalPAHlevelsweremeasuredinthemusselsafterfive days,onemonth,twomonths,threemonthsandoneyear(Sagerup etal.,unpublisheddata).

2.1. InputtoEcopathmodelforthePechoraSea

TheEcopathmodel(Polovina,1984;ChristensenandWalters, 2004)isamassbalancemodellingapproachbasedonasetoflin- earequationsrepresentingflowofbiomassesbetweengroupsin theecosystem.The“massbalance”termreferstothephysicalcon- straintofthemodel parametersdescribing thesystemtobein

“balance”.Thisoccurswhentheflowsintoagroupequaltheflows outofthegroupandmortalityforapreyequalsconsumptionby apredator.Ecopathisthebasemodelrepresentingasnapshotof thesystem.TheEcopathmodelbalanceslossesandgainsforeach functionalgroupusingEq.(1).

Bi∗

B

i∗EEi=BAi+Yi+

(j=1 to n)

Bj∗

B

j∗DCij (1)

where B is biomass, P/B is production per biomass and EE (ecotrophicefficiency) is thefraction of production transferred withinthemodel,BAisbiomassaccumulationandYismortali- tiesduetofisheriesandhunting.Q/Bisconsumptionperbiomass ratioandDCisthefractionofgroupiinthedietofgroupj.

Ourmodelarea(the PechoraSea)coversanareaofapproxi- mately125000km²(Fig.1).Wehavedefined27functionalgroups inthemodel,includingfourgroupsofmammals,twogroupsof birds,sevengroupsoffish,10invertebrategroups,twogroupsof primaryproducersandtwodetritusgroups(AppendixA).

Mostoftheproduction and consumption valuesare derived from,orcomparedto,anEcopathmodelmadefortheNorwegian andBarents Seasin2002(Dommasneset al.,2002).Recordings from the Pechora Sea were used to calculate the biomass of whales(Boltunovetal.,2014),seals(Boltunovetal.,2014),wal- rus(Lydersenetal.,2012),seabirds(Spiridonovetal.,2011)and ducksandeiders(Strømetal.,2000;Spiridonovetal.,2011).Most datafromthePechoraSeahavebeencollectedduringsummerand autumn,and withthelackof seasonaldata,wehad toassume year-roundvalidityforthebiomassdata.

For the fishgroups, data onbiomass were calculated based onProkhorova(2013).ForAtlanticcod(Gadusmorhua),haddock (Melanogrammus aeglefinus) and long rough dab (Hippoglos- soidesplatessoides)biomassestimatesweresuppliedbythePolar Research Instituteof MarineFisheriesand Oceanography, Mur- mansk,Russia(PINRO)(pers.comm.,DmitriProzorkevich).Forthe benthicgroups,datafromDenisenkoetal.(2003)wereusedand thecalculated P/Band Q/Bvalues fromUllsfjord (69–70N,20

E)in NorthernNorway(Nilsenet al.,2006).Snowcrab(Chio- noecetesopilio)isaninvasivespecieswithanincreasingbiomass intheBarentsSea(Sundet,2015).Thebiomassofsnowcrabwas estimatedbythemodel,butdiet(Koltsetal.,2013),production andconsumptionareestimatedbasedonvaluesfromtheEastern BeringStraitwerethespeciesoccursnaturally(Aydinetal.,2007).

Forzooplankton,datafromDvoretskyandDvoretsky(2009,2015) wereusedforbiomasscalculations.Detritusgroupswereadjusted tosustainthelargebenthicproductioninthePechorawithfood.A detaileddescriptionoftheassemblyoftheEcopathmodelisgiven inAppendixA.

2.1.1. Qualityofmodelandecologicalrobustness

Link(2010)outlinedasetofcomparisonsofinputdata,ratios andinformationtobeperformedinadvanceofanyfittingofthe model(PREBALanalysis).Thesearedescribedasasetof“rulesof thumb”toapplyinanearlysearchforoutliers(unrealisticallyhigh orlowvalues),andidentifyneedstoreevaluateanyofthedata attachedtothefunctionalgroupsinthemodel.ForourEcopath model,aPREBALdiagnosticidentifiedarelativelyfairsetofbiomass inputvalues.Bothinabsolutevaluesandrelatedtoprimarypro- duction(AppendixB).

2.2. Ecospace

EcospaceisthespatialmoduleofEcopath,whichdynamically allocatesbiomassacrossuserdefinedgridcells(Christensenand Walters,2004).Weuseda20×30gridcellmap(Fig.2)ofthemodel areawith23km×23kmcells.Fourhabitatsweredefined,<20m, 20–50m,50–100m,>100mwaterdepth.Thespatialmodelistwo- dimensional,sothedepthismerelyanameforthetypeofhabitat.

Allgroupsinthemodelwereassignedinproportionofthepopula- tiontothedifferenthabitats.Thegroup“nearshorebivalves”was fullyassignedtothehabitat<20m,whileoffshorebivalveswere assignedwith0.7(70%)ofthepopulationin20–50mand0.3(30%) in50–100m.Seaweedwereassignedasbeingallin<20m,andcod andhaddockboth50%to50–100mand50%to>100m.Allother functionalgroupsarebydefault“everywhere”(25%perhabitat).

2.3. Ecotracer

ThecontaminantmoduleofEcopath,Ecotracer,usestheflow ofbiomassbetweenthefunctionalgroupsandtheenvironment andpredictsconcentrationsofcontaminantsthatflowpassively withthebiomassinthefood-web.Thecontaminantsareassumed tofollowthebiomasspassivelyandinstantaneously.Themodel alsoallowsfordirectuptakefromtheenvironment,forexample acrossthegills.Decayrateofthepollutantisalsospecifiedforeach functionalgroupinthemodelandinthesurroundingenvironment (water).

We investigated how different inputs would influence the resultsinEcotracerbytestingthreedifferentinputcombinations (Table1).

1.Case1:Thelevelof PAHsinmusselsand redkingcrabafter exposureinourlaboratoryexperimentswereusedastheini- tialconcentrationassumingsimilaruptakeratesinthefieldas inthelaboratory.Lookingatthespreadonlythroughdietbut usingnodirectuptake,e.g.,overgills.

2.Case2:ThesameinitialconcentrationofPAHsasincase1but including direct uptake from the environment as calculated fromtheconcentrationsinwater.

3.Case3:ThescenarioofasuddenreleaseofMDO.Zeroinitial concentrationof PAHs. PAHsuptake fromwater and food.A maximumconcentrationofPAHswasdefinedforthegroupsof nearshoreandoffshorebivalvesandsnowcrabsothattheoutput fromEcotracerwassimilartotheinputvaluesincases1and2.

Thefollowingparameterswereenteredasabasisforthecalcu- lations:

a.Initial environmental concentration Co: we used 0.0104 tonnes/km²,avaluerecordedintheharbouratSkjervøy5days afterthespill.

b.Decayrate(/year):weused10tonnes/year,agenerichighvalue asweexpecttheMDOwiththePAHstodisperse,evaporateand degraderapidlyunlesstakenupbybiota.

80 L.-H.Larsenetal./EcologicalModelling331(2016)77–85

Fig.2.Above:HabitatslayerinEcospaceforthePechoraSea.Darkblueareaare>100mwaterdepth,lighterblue,100–50m,blue/grey50–20,lightgrey<20mwaterdepth.

Darkgrey:land.Below:ThecontaminantslayerusedinEcotracer,redishighconcentrationofPAHs,andgraduallydecreasinguntilthewhitearea,containingnoPAHs.Grey island.(Forinterpretationofthereferencestocolourinthisfigurelegend,thereaderisreferredtothewebversionofthisarticle.)

c.Baseinflowrateofthecontaminanttotheecosystem(tonnes/

km²/year):weused0aswearelookingataone-timeevent.

d.Basevolumeexchangelossofcontaminant(/year):weused0, andthisisnotrelevantwhenthereisnoinflow.

Foreachfunctionalgroup,wespecified:

e.Initialconcentrationintonnespollutantpertonnebiomass(t/t):

incase1and2weusedobservedconcentrationsfromtheexper- imentsandincase3weused0.

f.Concentrationinimmigratingbiomass(t/t):Weused0forevery groupaswehaven’testimatedanyimmigratingbiomass.

g.Directabsorptionrate(tonnes/tonnes/year),forexampleuptake overthegills:weused0foreverygroupinthefirstrun,inthe secondrunweuseduptakeratescalculatedfromtheexposure experimentsforthebenthicgroupsincludingcrustaceans.Inthe thirdrunweusedtheratesthatgavetheobservedconcentra- tionsintheexperiments.

h.Decayrate(tonnes/year):weused10tonnesperyearforwarm- blooded groups, mammals and seabirds, 1 tonne/yearfor all other groups except the detritus and phytoplankton groups whereweused0.

L.-H.Larsenetal./EcologicalModelling331(2016)77–85 81

Table1

MeasuredvaluesusedasinputtoEcotracer.Accumulationratiosobtainedafteraoneweekstudy.Thepolycyclicaromatichydrocarbons(PAHs)weremeasuredinsofttissues ofbluemusselsandIslandicscallopsandthehepatopancreasoftheredkingcrab.Alltissueconcentrationin␮g/kgwetweight(wetwt)andwaterconcentration(␮g/L).

Group/environment Ecopathgroup Data Waterconcentration Accumulationratio

Bluemussel (Mytilusedulis)

Nearshorebivalvia 4466(field)4482(lab) 18 249(oneweek)

Icelandicscallop (Chlamysislandica)

Offshorebivalvia 2957(lab) 18 169(oneweek)

Redkingcrab,(Paralithodescamtschaticus) Snowcrab 22254(lab) 18 1236(oneweek,

accumulationfrom feedandwater) Skjervøyharbour

seawater5daysafterspill

Initialconcentration, Environment

10.4^a

aAssumedtostaymostlyattop1mofthewatercolumn.

i.Proportionofcontaminantassimilated[0–1]:weused0.7forall groups.

Allthe equations aredescribed in the Ecopath usermanual (Christensenetal.,2008)anddescribedinAppendixA.

OurinitiallymeasuredPAHsconcentrationinwater(

PAH₁₆) fromtheharbourofSkjervøy,followingtheaccidentwas10.4␮g/L.

1000Lperm³gives0.0104g/m²ofPAHs,assumingthatthediesel staysmostlyatthesurface,top1m.Thiswasusedastheinitial environmentalconcentration.

Fromtheexperimentsinthelaboratorywestarted byusing 0.000004466tonnescontaminant/tonnebiomass(4466␮g/kgwet weight(wet wt)) fornearshore bivalves.Fromtheexperiments withbluemussels,0.000002957tonnes/tonne(2957␮g/kgwetwt) foroffshorebivalves,takenfromtheexperimentswithIcelandic scallop.Finallyweused0.0000222548tonnes/tonne(22258␮g/kg wetwt)fromredkingcrab,usedforthesnowcrab.Thesewere usedastheinitialconcentrationsforthecorrespondinggroupsin cases1and2andtheupperlimitforthesegroupsincase3.

Thedirectuptakeratiowascalculatedfromtheexperiments.

Inthehighexposuregroup,thelevelofPAHswas17.94␮g/Land afteroneweekofexposure,thebluemusselshadaconcentrationof 4482␮g/kgwetwt.Thismeansthereisahighratioofaccumulation fromwatertomussels.TheIcelandicscallopalsohadahighratio ofuptakefromthewaterwiththesamelevelofPAHs,reaching asofttissue concentrationof2957␮g/kgwetwt.Redkingcrab wereexposedboththroughfeedandwaterandaccumulatedmuch highertissueconcentrationsthanmussels(22254␮g/kgwetwt).

Fromthisweassumedahighdegreeoftrophictransferandused 0.7asassimilationefficiencyinEcotracer.

Wecompared themodelusing zeroasthedirectuptake for everygroupinthefirstcasestudytoeliminatethisasavariable, whileinthesecondcaseweranwiththehighuptakeratesmea- suredinthelaboratory:aratioof249fornearshorebivalvesand 150foroffshorebivalves.Forsnowcrab,weused250aswelleven thoughtheobservedconcentrationwasmuchhigherthaninthe bivalves.

Thethirdcasestudyusedzeroinitialconcentrationforallthe groupsandthedirectuptakewasadjustedtogetcomparablelevels ofPAHs.By usinga directuptake ratioofone andanassimila- tionrateof0.7asimilarinitialconcentration,asmeasuredinthe experiments(Sagerupetal.,unpublisheddata),wasachievedfor nearshoreandoffshorebivalves.Forthesnowcrab,wereduced thedirectuptakeratioto0.2astheyarenotfilterfeeders.

Ourthreemodelrunsweredesignedtoinvestigateamomentary releaseofMDOfromthewreckageofaship.ThePAHswasaddedas alayerinEcospacewithhighconcentrationnearthewrecksiteat Vaygach,andwithrapidlydecreasingconcentrationswithdistance tothespill(Fig.2).EcotracerwasrunwithEcospaceasaspatial modelwithPAHsasacontaminantlayer.

3. Results 3.1. Ecopath

BuildinganEcopathmodelisaninformativeexercisethatgen- eratesaknowledgebaseonbiologicalcomponentsofanareaand identifiesanyexistingknowledgegaps.Eventhoughmusselsare thepreferredpreyoftheestimatedalmost4000walruses,24000 eiders,snowcrabandmanyspeciesoffish,theecotrophicefficiency (EE)fortheoffshorebivalvewasestimatedbythemodeltobe0.126 andnearshorebivalvestobe0.150.TheEEisestimatedonascale from0to1,where1wouldmeanallproductionisconsumed.The EEislowforseveralofthebenthicgroups.Thismeansthereisalot ofbenthicproductionnotbeingconsumedbypredators.

3.2. Ecotracer

PAHslevelsspreadfastinthePechorafoodwebandespecially totoppredatorssuchasseals,eidersandwalruses(Table2).Table2 alsoshowsresultingvaluesafter0.5yearsand5yearsforallgroups.

Fig.3summarizestheconcentrationperbiomassforthefirst5years afterthespillfornearshorebivalvesandwalrusesforallthreecases.

AlltheresultsfromtheEcotracerrunsareprovidedinAppendixA.

3.2.1. Ecospace

Sincethecontaminantscamefromonepointsourceandwere notdistributedthroughoutthewholeecosystem,weranEcotracer asaspatio-temporalmodel(Ecospace).Thecontaminantconcen- trationgraduallydecreaseswithdistancefromthespillsite(Fig.2).

Thespatialoverlapbetweenthefunctionalgroupsofanimals andthespilllayermeansthatmanyanimalsarenotexposedatall tothepollutant.Thespatio-temporalmodelestimatedthebiomass concentrationstobereachedafterabout0.5–3years(AppendixA).

For case 1, thePAHs levels of3713␮g/kgwet wt in eiders, 1290␮g/kg wet wt in seals and walrus were estimated to be 753␮g/kgwetwt(AppendixA).

Incase2,theconcentrationlevelsperbiomassweresimilarto case1forthetoppredatorgroups(mammals,seabirdsandfish).

However,thebenthosandcrustaceangroupsaccumulatedslightly higherlevelsasaresultofdirectuptakefromthewater.Thiswas theonlymodelrunwhereconcentrationinbivalvescontinuedto increasefor2years(AppendixA).

Therewasnocorrelation betweenmaximumlevels ofPAHs according to Ecotracer and the variables production/biomass, consumption/biomassor theconsumption of bivalves(Pairwise correlationtest,Pearsons,p>0.05).

4. Discussion

Ecological modelling systems are valuable support tools for managing human influence on the marine ecosystems. Using

82 L.-H.Larsenetal./EcologicalModelling331(2016)77–85

Table2

ResultsfromEcotracerforallthreecases.Concentrationsin␮g/kgwetweight.

Case1 Case2 Case3

Time(years) 0.5 5 0.5 5 0.5 5

Water 1294.86 44.28 0.41 0.08 0.38 0.09

Toothedwhales 288.87 73.27 289.00 81.60 0.05 0.01

Baleenwhales 8.27 15.97 8.63 25.90 0.01 0.01

Seals 984.65 228.16 985.00 244.00 0.05 0.02

Walruses 553.40 173.86 553.00 215.00 0.69 0.07

Seabirds-pelagic 6.12 4.27 8.98 38.40 0.23 0.02

Ducksandeiders 2685.53 722.63 2690.00 1000.00 1.59 0.36

Cod 24.15 1.23 24.30 3.18 0.05 0.02

Haddock 168.31 9.62 169.00 12.90 0.18 0.05

Longroughdab 0.91 0.51 1.16 5.94 0.17 0.09

Polarcod 0.16 0.10 0.63 1.65 0.00 0.00

Otherdemersalfish 0.68 0.40 0.90 3.95 0.12 0.07

Pelagicfish 0.13 0.10 0.62 1.93 0.00 0.00

Sandeel(Ammodytidae) 0.28 0.11 0.71 1.87 0.00 0.01

Snowcrab 6813.47 38.67 6830.00 54.00 0.13 0.07

Echinoderms 0.78 0.57 6.96 12.90 0.10 0.17

Polychaetes 1.14 0.21 6.25 4.99 0.05 0.05

Nearshorebivalves 7.32 1.18 23.40 43.60 1.60 0.33

OffshoreBivalvia 15.60 1.37 21.00 18.60 2.90 0.24

Otherbenthos 0.94 0.27 6.42 6.46 0.05 0.07

Shrimp 2.57 1.15 12.20 26.40 0.41 0.14

Krill 1.12 – 8.93 – 0.02 10.14

Zooplankton 2.34 0.10 6.93 2.07 0.03 0.01

Jellyfish 0.24 0.03 0.30 0.25 0.00 0.00

Detritus 9.37 0.23 10.30 2.15 0.10 0.02

ecosystemmodellingcombinedwithecotoxicologyisinteresting and combining the two methodologies toquantitatively assess expectedimpactsthroughoutthefoodwebisavaluabletoolfor environmentalmanagement.Ecologicalimpactsofpollutants,such asPAHs,areonlysignificantiftheimpairsurvival,growth,repro- duction,causegeneticdisruptionorseriouslyaffectenergyflow throughtheecosystem.

TheEcotracermoduleinEcopathhasnotbeenwidelyapplied before, but ourcurrent applicationindicates goodpotential for beingatoolforcombiningtoxicologyandfood-webmodelling.A keychallengeformodellingecotoxicologyistheavailabilityofdata onthesamespeciesandtoxicologicalcompoundsonacomparative scaleastheecosystemcomponents.MDOdissolvesanddisperses rapidlyinseawaterandfromfieldmeasurementsatSkjervøyone monthafterofthespill,thewaterintheimmediatesurroundings hadnon-detectablelevelsofPAHs.However,cleanbluemusselsset out2.5monthsafterthespillaccumulatedPAHsfromthesurround- ings(SagerupandGeraudie,2014).Thisindicatethatbluemussels areextremelyefficientinaccumulatingPAHsfromseawater,and supportsouruseofhighaccumulationanddirectuptakeratesas inputtoEcotracer.Musselsaresuspension-feedingorganismsthat retainparticlesontheirgills,includingoildroplets.Thebluemus- selsaccumulatePAHsfrombothfoodandwaterindiscriminately (Baussantetal.,2001),indicatingthattheaccumulationofPAHsin bluemusselsareindependentofhydrocarbonswatersolubility.

ThebioavailabilityofPAHsafteraspilldependsonevaporation, dissolution,dispersionofoildropletsintothewatercolumn,water- in-oilemulsification, sinking and sedimentation (Fingas, 2011).

MDOisrelativelyeasilydispersedinwater.Therefore,a poolof hydrocarbon may quickly beavailable for accumulation in the organismsafteraspill.Thesolubilitydependsonthestructureand decreasesapproximatelylog-linearwithmolecularweight(Miller etal.,1985).Theheavymoleculesareboundinthedispersedoil droplets,butasthesedropletsarefilteredbythemussels,accumu- lationoccurs(Baussantetal.,2001).

The assimilation efficiencies for PAHs (concentration in prey/concentrationinpredator)arepoorlyknownfortheArctic.

Fromourexperimentontrophictransferusingredkingcraband

mussels,wecanconcludethattherearehighassimilationefficien- ciesastheconcentrationperbiomassinredkingcrabwasmuch higherthaninmussels(Table1).Theredkingcrabefficientlyaccu- mulatesPAHsfrommusselsaswellasdirectlyfromthewater.PAHs haveadifferentmolecularstructure,stabilityandbioavailability andtheassimilationefficienciesmayvarygreatly.Ourgenericvalue forassimilationefficienciesof0.7waschosentoreflectthevaria- tioninaccumulationofthedifferentPAHs.Baussantetal.(2001) showthatthebioconcentrationfactorinfish,calculatedastheratio betweenuptakeandelimination,variedfrom22to1495fordif- ferentPAHs.Ourassimilationrateof0.7willnotbethesamefor allgroupsinthemodel.Asnoexperimentallyverifiedassimila- tionratesexistforindividualPAHsweappliedavalueof0.7forall groups,exceptforcasethreewere0.2wasusedforsnowcrab.

Using3cases,ormodelrunswithdifferentinputcombinations, letusexplorehowthemodelrespondstodifferentinputvariables.

In case1, using0for theuptake ratefromwaterfor biological groups,theresultingvaluesarereachedonlythroughconsump- tionandthelinksofthefood-web.Incases1and2,themodel predictedveryhighassimilationefficiencyintoppredators.Inves- tigatingthedietproportionsofeachfunctionalgroupwillhelptrack thePAHs.Weused0.114asdietproportionofsnowcrabinseals andthisprobablyexplainswhysealswerepredictedtoaccumu- latehighlevelsofPAHs.Forwalrusesweapplieddietproportions of0.5nearshoreand0.35offshorebivalvesand0.05snowcrab, whileducksandeidershaveadietproportionof0.5nearshoreand 0.06offshorebivalves.Toothedwhaleshavea0.1dietproportionof snowcrab,indicatingthereasonswhyallthesegroupsaccumulate highmaximumlevels(Table2andAppendixA).Therefore,contri- butionofPAHsinfishgroupsdoesnotcontributetoPAHsinthe toppredators.

Thetwo processesofadvectionandspreadingdeterminethe movementbehaviourofanoilspill.MDOisalowviscousoilforming athinfilmonthesurfaceofthewater(Fingas,2011).Asthedissolu- tionanddispersionofMDOtheconcentrationgraduallydecreases withincreasingdistancefromthespillsite.Thereforethespatial model(Ecospace)oftheinitialconcentrationintheenvironmentis neededandwasintegratedinourcases.Further,PAHsarerelatively

L.-H.Larsenetal./EcologicalModelling331(2016)77–85 83

Fig.3. Ecotraceroutputforthegroups“nearshorebivalves”(left)and“walruses”(right)forcase1(upper),case2(middle)andcase3(lower).y-AxisshowsPAHsconcentration perbiomassin␮g/kgwetweight(noteverydifferentscales)andx-axisshowsyearsafterspill.Maximumvaluegivenintextoneachfigure.

biodegradableandthemodelmaybeoverestimatingthedegree ofbioaccumulation,asPAHsdonotbiomagnify(Neff,2002).PAHs alsodecomposebyphotochemicaloxidation(DuttaandHarayama, 2000;King et al.,2014)and microbial degradationin seawater (Harayama etal.,2004).Theseprocessesareofhighimportance forthedegradationofoilspilledatsea,butforsimplicityofthe model,thepollutantsareonlyhandledastracermoleculeswitha setdecayrate.

When running Ecotracer, one needs a balanced time series modelaschangesinbiomassinfluencetheconcentrationofpol- lutantsperbiomass.Forthegroupsandeel(Ammodytidae)wesee theresultofthegroupdyingout,therebyproducingahighconcen- trationasanartefactoflowbiomass(seefigureofcases1and2in AppendixA).WecanobservesimilarpeaksinPAHsconcentrations perbiomassforkrill.Howeverkrillisnotpredictedtodieout,so thismaybearesultoffluctuatingpopulationsizes.

Anaccidental,momentaryoilspillusuallyspreadsfromthedis- chargesite,andhasaspatialandtemporal componentdifferent fromthatofacontinuouspollutionsituation,e.g.,mercurylevels

intheoceans(BoothandZeller,2005).Tomodelthespatialcompo- nentwithEcospace,isuniqueasitgivesthepossibilitytomodela rangeofconcentrationsfromonespillsite.AfterthePrestigeoilspill inGalicia,Spain,November2002,PAHsweremeasuredinbloodof yellow-leggedgulls(Larusmichaelis).ThevalueoftotalPAHsin bloodingullsfromtheoilexposedcolonyatLobeiras,wasamax- imumof228␮g/kgwetwt17monthsafterthespill,whilegulls fromunexposedcolonieshadatotalPAHsconcentrationsinthe bloodofabout100␮g/kgwetwt(Pérezetal.,2008).InMay2009a cargovessel,MV“Petrozavodsk”ranagroundatBjørnøya(74^◦N19

◦E)intheBarentsSea,andleakedMDO.InJune2009,onemonth afterthegrounding,PAHslevelsinbloodfromglaucousgull(Larus hyperboreus)reached214␮g/kg,buttheaverageof28birdssam- pledwas42.7␮g/kg(Strømetal.,2011).In2010,only3of14gulls fromBjørnøyahaddetectablelevelsofPAHsintheirblood.This agreeswiththespatialmodelpredictionwheremaximumvalue forpelagicseabirdswas24␮g/kgwetwtafterabout20months.

Withinecotoxicology,measuredandcalculatedconcentrations ofcontaminantsareusuallyverylow(␮g/kgorevenng/kg).Butthe

84 L.-H.Larsenetal./EcologicalModelling331(2016)77–85

inputintotheEcotracerisontonnes/tonneslevelandvisualrepre- sentationsoftheconcentrationsperbiomassarethereforedifficult tointerpret.Researchonthelinkbetweentoxicologyandecology applyingmodellingonanecosystemlevelisclosetonon-existent.

Carreretal.(2000)statethat“mosttoxicsubstancemodelsfocuson thedynamicsofthechemical,andthereforesimplifytheproblem ofassessingtherateofconsumptionofcontaminatedfoodusing empiricalequationsbasedonthedimensionsoforganisms.”Pre- dictingtheactualoutcomeofanoilspillisvirtuallyimpossible,as therewillbeunforeseenconsequencesandinteractions.Ingeneral, mussels,suchasbluemussel,aresensitivetoincreasedlevelsof hydrocarbons.However,musselshavethepossibilitytoclosetheir shellsforlongperiodsandevenmonthsofzerofoodintake.This abilitymayproveadvantageoustothemusselsinaspillsituation.

O’ClairandRice(1985)foundthatmusselswerelesssensitiveto hydrocarbonsinthewaterthantheirpredator,thestarfishEvaste- riastroschenii,andtheysuggestthatchronicexposurefromanoil spillwouldincreasemortalityinthestarfishandthusgivethemus- selsthepossibilitytoexpandduetoreducedpredationpressure.

TheaimofthispaperwastotestwhetherEcotracercanbeused tosimulatecontaminantspreadinanArcticfoodweb.Usingone assimilationefficiencyforallgroupsmadeiteasiertocomparethe spreadinsteadofaddinguncertaintybyusingdifferentassimila- tionefficiencies.ItmaybeclaimedthatEcotraceroversimplifiesthe kineticsofspreadofpollutantsbyonlytakingafewvariablesinto consideration.However,wearguethatsimplificationisnecessary astheproblemisinfinitelycomplexjustlikeanecosystem.

ApplyingtheEcotracermoduleinaseaareawithlimitedback- grounddatahasbeenachallengingexercise.ThePechoraSeaholds limitedfisheriesresources,andthus dataonhumanremovalof biomassfromthemodelareaarepoor.TheEcopathwithEcosim modellingsystemhasbroadlybeendevelopedandappliedinareas wheretimeseriesofrecordingsoflandingsareavailableasinput data.Howeverthelackofsuchdatamadeitimpossibletosatisfac- torilyapplytheEcosimmodule.

ThePechoraSeahasstrongseasonality,andbyonlyconsider- ingthelimiteddata,mostlycollectedduringsummer,andusing ittorepresent afull yearonlyadded uncertainty tothemodel predictions.Internallyinthemodelarea,migrationsoccur.Also immigrationtoandemigrationfromthemodelareatakeplaceas partofthelifecycleofseveralofthefunctionalgroups(e.g.,whales andbirds).Immigrationandemigrationwerenotaddressedinour study.

5. Conclusions

Ecotracerisavaluabletooltocombinefood-webmodellingand ecotoxicology.Bridgingthesetwobranchesofbiologyisofimpor- tancetoliftthefocusofenvironmentalpollutiontoanecosystem level.Themodelledconcentrationsseemedunrealisticallyhighin somefunctionalgroups,especiallytoppredators.Providingdata thatcan beusedas input,fromthesamespecies or functional groups,preytypesandpollutantswasamajorchallenge.Aswell, thereareelementsofphysiologicalcharacterandkineticsthatare nottakenintoconsiderationinEcotracer.However,tobeableto modelthespreadofpollutantsatecosystemlevel,usingmanyfunc- tionalgroups,simplificationisalsoveryimportantandEcotracer provestobeacomprehensivemodellingtool.

EcopathwithEcosimandEcotracerisacomprehensivemodel tostudyecosystemcomplexity.Attemptstolinkmodelsforthe spreadofpollutantsinthefoodwebareessentialtoidentifygaps ofknowledge.However,inthecurrentapplicationofthemodel package,wedidnotmanagetogetholdofsufficientlyreliabletime seriesexceptfornomorethanafewofthe27functionalgroups,and wewerethusunabletomakeuseofthetimeintegratingproperties oftheEcosimmodule.Thescarcityofdataovertimeforcombined

ecotoxicology –ecology modellingin Arcticseas thus becomes evidenthere.Despitetheseshortcomingsandpotentialsourcesof error,ourexercisehasshownthatafood-webinfluencedbyasin- gleaccidentaleventcanbemodelled,andspreadofcontaminants addressedinasatisfactorywayapplyingtheEcotracermodule.

Wesuggestfurtherworktoincludedataontropictransfertotop predatorsandspreadofPAHsintheecosystem.Also,comparisons ofmodellingresultsfromexperimentswithheavyfueloils(HFO) areencouraged.

Acknowledgements

This study is conducted as part of a large interdisciplinary researchprogramme,A-lexaco-operationbetweenUiT-TheArctic UniversityofNorway(FacultyofLawandFacultyofSocialScience), Akvaplan-niva(environmentalstudies)andMarintek(technology forthefutureofArcticShipping).FundingcomesfromtheNorwe- gianMinistryofForeignAffairsthroughtheBarents2020program, refno12/00900.WewouldliketothankDmitriProzorkevichat PINROfor supplying stockestimates fromthe PechoraSeaand theparticipantsatthe“30yearsofEcopath”conferenceforvalu- ablefeed-backanddiscussionsduringtheconference,andChris Emblow,Akvaplan-niva, for preparingFig.1 and correctingthe authorsEnglish.

AppendixA. Supplementarydata

Supplementarydataassociatedwiththisarticlecanbefound,in theonlineversion,athttp://dx.doi.org/10.1016/j.ecolmodel.2015.

10.011.

References

AMAP/CAFF/SDWG,2013.IdentificationofArcticMarineAreasofHeightenedEco- logicalandCulturalSignificance:ArcticMarineShippingAssessment(AMSA) IIc,Oslo.

Aydin,K.,Gaichas,S.,Ortiz,I.,Kinzey,D.,Friday,N.,2007.AComparisonoftheBering Sea,GulfofAlaska,andAleutianIslandsLargeMarineEcosystemsThroughFood WebModeling.U.S.DepartmentofCommerce.

Bakke,T.,Sørensen,K.,1985.Experimentallong-termoilexposureonrockyshore mesocosms.Int.OilSpillConf.Proc.1985,649.

Baussant,T.,Bechmann,R.K.,Taban,I.C.,Larsen,B.K.,Tandberg,A.H.,Bjornstad,A., Torgrimsen,S.,Naevdal,A.,Oysaed,K.B.,Jonsson,G.,Sanni,S.,2009.Enzymatic andcellularresponsesinrelationtobodyburdenofPAHsinbivalvemolluscs:a casestudywithchroniclevelsofNorthSeaandBarentsSeadispersedoil.Mar.

Pollut.Bull.58,1796–1807.

Baussant,T.,Sanni,S.,Skadsheim,A.,Jonsson,G.,Borseth,J.F.,Gaudebert,B.,2001.

Bioaccumulationofpolycyclicaromaticcompounds:2.Modelingbioaccumula- tioninmarineorganismschronicallyexposedtodispersedoil.Environ.Toxicol.

Chem.20,1185–1195.

Bechmann,R.K.,Larsen,B.K.,Taban,I.C.,Hellgren,L.I.,Moller,P.,Sanni,S.,2010.

ChronicexposureofadultsandembryosofPandalusborealistooilcausesPAHs accumulation,initiationofbiomarkerresponsesandanincreaseinlarvalmor- tality.Mar.Pollut.Bull.60,2087–2098.

Boltunov,A.N.,Belikov,S.E.,Gorbunov,Y.A.,Mentis,D.T.,Semenova,V.S.,2010.The AtlanticWalrusoftheSoutheasternBarentsSeaandAdjacentRegions:Review ofPresentDayStatus.WWF-Russia,Moscow.

Boltunov,A.N.,Belikov,S.E.,Nikiforov,V.,Semenova,V.S.,Stishov,M.S.,Pukhova, M.A.,2014.AerialsurveyofthePechoraSeaandtheareaofVaigachIslandin spring2014.In:MarineMammalsoftheHolarctic(VIII).,pp.82,SaintPetersburg.

Booth,S.,Zeller,D.,2005.Mercury,foodwebs,andmarinemammals:implications ofdietandclimatechangeforhumanhealth.Environ.HealthPerspect.113, 521–526.

Carrer,S.,Halling-Sorensen,B.,Bendoricchio,G.,2000.Modellingthefateofdioxins inatrophicnetworkbycouplinganecotoxicologicalandanEcopathmodel.Ecol.

Model.126,201–223.

Christensen,V.,Walters,C.,Pauly,D.,2008.EcopathwithEcosim–AUser’sGuide.

FisheriesCentre.UniversityofBritishColombia,Vancouver.

Christensen,V.,Walters,C.J.,2004.EcopathwithEcosim:methods,capabilitiesand limitations.Ecol.Model.172,109–139.

Culbertson,J.B.,Valiela,I.,Pickart,M.,Peacock,E.E.,Reddy,C.M.,2008.Long-term consequencesofresidualpetroleumonsaltmarshgrass.J.Appl.Ecol.45, 1284–1292.

Denisenko,S.G.,Denisenko,N.V.,Lehtonen,K.K.,Andersin,A.-B.,Laine,A.O.,2003.

MacrozoobenthosofthePechoraSea(SEBarentsSea):communitystructureand

L.-H.Larsenetal./EcologicalModelling331(2016)77–85 85

spatialdistributioninrelationtoenvironmentalconditions.Mar.Ecol.Prog.Ser.

258,109–123.

Dommasnes,A.,Christensen,W.,Ellertsen,B.,Kvamme,C.,Melle,W.,Nøttestad,L., Pedersen,T.,Tjelmeland,S.,Zeller,D.,2002.AnEcopathModelfortheNorwegian SeaandBarentsSea.,pp.213–240.

Dutta,T.K.,Harayama,S.,2000.Fateofcrudeoilbythecombinationofphotooxida- tionandbiodegradation.Environ.Sci.Technol.34,5.

Dvoretsky,V.G.,Dvoretsky,A.G.,2009. Summermesozooplanktonstructurein the PechoraSea(south-eastern BarentsSea).Estuar. Coast. ShelfSci. 84, 11–20.

Dvoretsky,V.G.,Dvoretsky,A.G.,2015.Earlywintermesozooplanktonofthecoastal south-easternBarentsSea.Estuar.Coast.ShelfSci.152,116–123.

Ferriss,B.E.,Essington,T.E.,2014.Doestrophicstructuredictatemercuryconcen- trationsintoppredators?Acomparativeanalysisofpelagicfoodwebsinthe PacificOcean.Ecol.Model.278,18–28.

Fingas,M.,2011. OilSpillScience andTechnology:Prevention, Response,and Cleanup.GulfProfessionalPublishing,Burlington,MA.

Gilfillan,E.S.,Vandermeulen,J.H.,1978.AlterationsinGrowthandPhysiologyofSoft- ShellClams,Myaarenaria,ChronicallyOiledwithBunkerCfromChedabuctoBay, NovaScotia,1970–76.J.Fish.Res.BoardCan.35,630–636.

Harayama,S.,Kasai,Y.,Hara,A.,2004.Microbialcommunitiesinoil-contaminated seawater.Curr.Opin.Biotechnol.15,205–214.

IMO,2006.RevisedGuidelinesfortheIdentificationandDesignationofParticularily SensitiveSeaAreas.I.M.Organization.

King,S.M.,Leaf,P.A.,Olson,A.C.,Ray,P.Z.,Tarr,M.A.,2014.Photolyticandphotocat- alyticdegradationofsurfaceoilfromtheDeepwaterHorizonspill.Chemosphere 95,415–422.

Kolts,J.,Lovvorn,J.,North,C.,Grebmeier,J.,Cooper,L.,2013.Effectsofbodysize, gender,andpreyavailabilityondietsofsnowcrabsinthenorthernBeringSea.

Mar.Ecol.Prog.Ser.483,209–220.

Larsen,L.-H.,Kvamstad-Lervold,B.,Sagerup,K.,Gribkovskaia,V.,Bambulyak,A., Rautio,R.,Berg,T.E.,2015.TechnologicalandenvironmentalchallengesofArctic shipping:acasestudyofafictionalvoyageintheArctic.PolarRes.(Inpress).

Link, J.S., 2010. Adding rigor to ecological network models by evaluating a set of pre-balance disgnostics: a plea for PREBAL. Ecol. Model. 221, 1580–1591.

Lydersen,C.,Chernook,V.,Glazov,D.,Trukhanova,I.,Kovacs,K.,2012.Aerialsurvey ofAtlanticwalruses(Odobenusrosmarusrosmarus)inthePechoraSea,August 2011.PolarBiol.35,1555–1562.

Miller,M.M.,Wasik,S.P.,Huang,G.L.,Shiu,W.Y.,Mackay,D.,1985.Relationships betweenoctanol–waterpartitioncoefficientandaqueoussolubility.Environ.

Sci.Technol.19,522–529.

Moore,M.N.,Livingstone,D.R.,Widdows,J.,1989.Hydrocarbonsinmarinemollusks:

Biologicaleffectsandecologicalconsequences.In:Varanasi,U.(Ed.),Metabolism

ofPolycyclicAromaticHydrocarbonsintheAquaticEnvironment.CRCPressInc., Florida,USA.

Nahrgang,J.,Camus,L.,Carls,M.G.,Gonzalez,P.,Jonsson,M.,Taban,I.C.,Bechmann, R.K.,Christiansen,J.S.,Hop,H.,2010.Biomarkerresponsesinpolarcod(Bore- ogadussaida)exposedtothewatersolublefractionofcrudeoil.Aquat.Toxicol.

97,234–242.

Neff,J.M.,2002.BioaccumulationinMarineOrganisms,EffectofContaminantsfrom OilWellProducedWater.Elsevier.

Nilsen,M.,Pedersen,T.,Nilssen,E.M.,2006.Macrobenthicbiomass,productivity (P/(B)over-bar)andproductioninahigh-latitudeecosystem,NorthNorway.

Mar.EcolProg.Ser.321,67–77.

O’Clair,C.E.,Rice,S.D.,1985.Depressionoffeedingandgrowthratesoftheseastar Evasteriastroscheliiduringlong-termexposuretothewater-solublefractionof crudeoil.Mar.Biol.84,331–340.

Pérez,C.,Velando,A.,Munilla,I.,López-Alonso,M.,Oro,D.,2008.Monitoringpoly- cyclicaromatichydrocarbonpollutioninthemarineenvironmentafterthe prestigeoilspillbymeansofseabirdbloodanalysis.Environ.Sci.Technol.42, 707–713.

Peterson,C.H.,Rice,S.D.,Short,J.W.,Esler,D.,Bodkin,J.L.,Ballachey,B.E.,Irons,D.B., 2003.Long-termecosystemresponsetotheExxonValdezoilspill.Science302, 2082–2086.

Polovina,J.,1984.Modelofacoralreefecosystem.CoralReefs3,1–11.

Prokhorova,T.,2013.SurveyReportfromtheJointNorwegian/RussianEcosys- temSurveyintheBarentsSeaandAdjacentWaters,August–October2013.

IMR/PINRO.

Sagerup,K.,Geraudie,P.,2014.MiljøundersøkelseetterdieselutslipptilSkjervøy havn(inNorwegian).APN-6821.01.Akvaplan-niva.

Sagerup,K.,Larsen,L.H.,Nahrgang,J.,Frantzen,M.,Geraudie,P.,2015.Biological effectsofmarinedieselinRedKingCrab(Paralithodescamtschaticus)through water-andfoodborneexposure,unpublisheddata.

Spiridonov, M.,Gavrilo,E., Krasnova,E., Nikolaeva,N., 2011. Atlasof Marine andCoastalBiologicalDiversityoftheRussianArctic.FacultyofGeography, LomonosovMoscowStateUniversity.

Strøm,H.,Isaksen,K.,Golovkin,A.,2000.SeabirdandWildfowlSurveysinthe PechoraSeaDuringAugust1998.NorwegianOrnithologicalSociety.

Strøm, H.,Verboven, N.,Hallanger,I.,2011. MiljøundersøkelserpåBjørnøyai forbindelsemedgrunnstøtingenavM/VPetrozavodsk2009–2011.Rapportfra NorskPolarinstitutttilKystverket(inNorwegian).

Sundet,J.,2015.Snøkrabbe(SnowCrab).SnowCrabBiology,SpreadandCommercial FisheryintheBarentsSea.InstituteofMarineResearch.

Thomann,R.V.,1989.Bioaccumulationmodeloforganicchemicaldistributionin aquaticfoodchains.Environ.Sci.Technol.23,699–707.

Widdows,J.,Donkin,P.,Evans,S.V.,1985.RecoveryofMytilusedulisL.fromchronic oilexposure.Mar.Environ.Res.17,250–253.

Appendix A

This is appendix to “The Mussel Path - Using the contaminant tracer, Ecotracer, in Ecopath to model the spread of pollutants in an Arctic marine food web”.

Appendix A and includes detailed description of input data. Appendix B presents the PREBAL procedure applied to assess the quality of the input data.

The model include 27 functional groups, and input data and major assumptions are summarised. The Pechora Sea modelling area is approximately 125 000 km

(Figure 1 in the article). The biomass, production and consumption is summarized in Table 1, the diet matrix in Table 2.

1. Toothed whales

Boltunov et al. (2014) reported sighting of 50 beluga whales (Delphinapterus leucas) during an aerial survey of parts of the Pechora Sea. The survey covered 3000 km

. Assuming that the density of whales is the same all over the Pechora Sea, this gives a biomass of 0.01667 tons/km

, assuming an individual weight of 1000 kg.

2. Baleen whales

PINRO and IMR perform joint ecosystem surveys of the Barents Sea in August – September each year. In 2013, 4 minke whales (Balaenoptera acutorostrata) were observed in the Pechora Sea (Prokhorova 2013), while 3 individuals were recorded during the 2014 survey (Klepikovsky 2014). No other baleen whale species were recorded in the Pechora Sea during these surveys. Both Baleen and toothed whales can only be present in the Pechora Sea when there is no ice, and their migrations into the area is assumed to be for a limited time period each year. An average individual body weight of 5252 kg was used as an estimate for the minke whale biomass (Sigurjónsson and Víkingsson 1997), providing a biomass estimate within the modelling area of 0,000168 tons/km

.

3. Seals

The east-ice stock of harp seal (Pagophilus groenlandicus) whelp on the drift ice, mainly west of the Pechora Sea, in March and April, and feed in the Pechora Sea parts of the year Boltunov et al. (2014).

An aerial survey of the Pechora Sea and the sea area off Vaigach Island in spring 2014 observed 82 seals, consisting of ringed- (Pusa hispida), harp,- and only one bearded seal (Erignathus barbatus) over transects that covered an area of 3000 km

Boltunov et al. (2014). Assuming an average weight of 100 kg per individual, the seals have a biomass of 0.0273 tons/km

. This estimate is somewhat lower than the 0.087 estimate used for the Barents Sea Dommasnes et al. (2002). Therefore, this estimate was adjusted to 0.0332 tons/km

when fitting the model.

According to Lindstrøm et al. (1998) Seals of the Barents- and Pechora Sea mainly feed on herring

(Clupea harengus) and polar cod (Boreogadus saida). Samples taken in the western part of the model

area, west of Kolguyev Island (Figure 1 in the article) had a frequency of about 70 % herring in the

diet, while closer to Pechora Bay; the polar cod was more important (47 %) and the occurrence of the

herring was reduced (26 %).