Using ecological models to assess ecosystem status in support of the European Marine Strategy Framework Directive

(1)

HAL Id: hal-02154205

https://hal.archives-ouvertes.fr/hal-02154205

Submitted on 16 May 2020

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Using ecological models to assess ecosystem status in support of the European Marine Strategy Framework

Directive

Chiara Piroddi, Heliana Teixeira, Christopher Lynam, Chris Smith, María Alvarez, Krysia Mazik, Eider Andonegi, Tanya Churilova, Letizia Tedesco,

Marina Chifflet, et al.

To cite this version:

Chiara Piroddi, Heliana Teixeira, Christopher Lynam, Chris Smith, María Alvarez, et al.. Using

ecological models to assess ecosystem status in support of the European Marine Strategy Framework

Directive. Ecological Indicators, Elsevier, 2015, 58, pp.175-191. �10.1016/j.ecolind.2015.05.037�. �hal-

02154205�

(2)

EcologicalIndicators58(2015)175–191

ContentslistsavailableatScienceDirect

Ecological Indicators

jou rn al h om ep a g 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 i n d

Review

Using ecological models to assess ecosystem status in support of the European Marine Strategy Framework Directive

Chiara Piroddi

^a,∗

, Heliana Teixeira

^a

, Christopher P. Lynam

^b

, Chris Smith

^c

,

Maria C. Alvarez

^d,l

, Krysia Mazik

^d

, Eider Andonegi

^e

, Tanya Churilova

^f,k

, Letizia Tedesco

^g

, Marina Chifﬂet

^e

, Guillem Chust

^e

, Ibon Galparsoro

^e

, Ana Carla Garcia

^h

, Maria Kämäri

^g

, Olga Kryvenko

^f,k

, Geraldine Lassalle

^i,j

, Suzanna Neville

^b

, Nathalie Niquil

^j

,

Nadia Papadopoulou

^c

, Axel G. Rossberg

^b

, Vjacheslav Suslin

^k

, Maria C. Uyarra

^e

aEuropeanCommission,JointResearchCentre(JRC),InstituteforEnvironmentandSustainability(IES),WaterResourcesUnit,21027Ispra(VA),Italy

bCentreforEnvironment,Fisheries&AquacultureScience(Cefas),PakeﬁeldRoad,LowestoftNR330HT,UK

cHellenicCentreforMarineResearch,P.O.Box214,71003Heraklion,Crete,Greece

dInstituteofEstuarine&CoastalStudies,UniversityofHull,CottinghamRoad,HullHU67RX,UK

eAZTI,MarineResearchDivision,Herrerakaiaportualdeaz/g,20110Pasaia,Spain

fInstituteofBiologyoftheSouthernSeas,2NakhimovAve,299011Sevastopol,RussianFederation

gFinnishEnvironmentInstitute,MarineResearchCentre,Helsinki,Finland

hIMAR,InstitutodoMar,LargoMarquesdePombal,3004-517Coimbra,Portugal

iIRSTEA,UREABX,AquaticEcosystemsandGlobalChanges,50avenuedeVerdun,33612Cestascedex,France

jCNRS,UMR7208BOREA,NormandieUniversité,UniversitédeCaenBasse-Normandie,14032Caencedex5,France

kMarineHydrophysicalInstitute,2KapitanskayaStr.,299011Sevastopol,RussianFederation

lNaturalEngland,SustainableDevelopment,TempleQuayHouse,BristolBS16DG,UK

a r t i c l e i n f o

Articlehistory:

Received9July2014

Receivedinrevisedform14April2015 Accepted19May2015

Keywords:

MSFD

Marineecosystems Ecologicalmodels Model-derivedindicators Pressures

Habitats

Biodiversitydescriptors

a b s t r a c t

TheEuropeanUnion’sMarineStrategyFrameworkDirective(MSFD)seekstoachieve,forallEuropean seas,“GoodEnvironmentalStatus”(GEnS),by2020.Ecologicalmodelsarecurrentlyoneofthestrongest approachesusedtopredictingandunderstandingtheconsequencesofanthropogenicandclimate-driven changesinthenaturalenvironment.Weassessthemostcommonlyusedcapabilitiesofthemodelling communitytoprovideinformationaboutindicatorsoutlinedintheMSFD,particularlyonbiodiversity, foodwebs,non-indigenousspeciesandseaﬂoorintegritydescriptors.Webuiltacatalogueofmodelsand theirderivedindicatorstoassesswhichmodelswereabletodemonstrate:(1)thelinkagesbetweenindi- catorsandecosystemstructureandfunctionand(2)theimpactofpressuresonecosystemstatethrough indicators.Oursurveyidentiﬁed44ecologicalmodelsbeingimplementedinEurope,withahighpreva- lenceofthosethatfocusonlinksbetweenhydrodynamicsandbiogeochemistry,followedbyend-to-end, speciesdistribution/habitatsuitability,bio-optical(remotesensing)andmultispeciesmodels.Approx- imately200indicatorscouldbederivedfromthesemodels,themajorityofwhichwerebiomassand physical/hydrological/chemicalindicators.Biodiversityandfoodwebsdescriptors,with∼49%and∼43%

respectively,werebetteraddressedinthereviewedmodellingapproachesthanthenon-indigenous species(0.3%)andseaﬂoorintegrity(∼8%)descriptors.Outof12criteriaand21MSFDindicatorsrele- vanttotheabovementioneddescriptors,currentlyonlythreeindicatorswerenotaddressedbythe44 modelsreviewed.Modellingapproachesshowedalsothepotentialtoinformonthecomplex,integra- tiveecosystemdimensionswhileaddressingecosystemfundamentalproperties,suchasinteractions betweenstructuralcomponentsandecosystemsservicesprovided,despitethefactthattheyarenotpart oftheMSFDindicatorsset.Thecataloguingofmodelsandtheirderivedindicatorspresentedinthisstudy, aimathelpingtheplanningandintegrationofpoliciesliketheMSFDwhichrequiretheassessmentof allEuropeanSeasinrelationtotheirecosystemstatusandpressuresassociatedandtheestablishment ofenvironmentaltargets(throughtheuseofindicators)toachieveGEnSby2020.

∗Correspondingauthorat:InstituteofMarineScience,SpanishResearchCouncil,Barcelona,Spain.

E-mailaddress:cpiroddi@hotmail.com(C.Piroddi).

http://dx.doi.org/10.1016/j.ecolind.2015.05.037

(3)

176 C.Piroddietal./EcologicalIndicators58(2015)175–191 Contents

1. Introduction...176

2. Cataloguestructure...176

3. Modelcharacteristics...177

3.1. Biogeochemicalmodels...177

3.2. Multispeciesmodels...177

3.3. SpeciesDistributionModels(SDM)/HabitatSuitabilityModels(HSM)...177

3.4. Meta-communitymodels...177

3.5. Bio-opticalmodels...180

3.6. Hydrodynamic–biogeochemicalModels...180

3.7. End-to-endmodels...180

4. Modelpotentialtoaddressdescriptorsandindicatorsforbiologicaldescriptors...180

4.1. Biodiversitycomponentsandhabitats...183

5. Modelsgeographicalcoverage...185

6. Addressingpressureswithmodels...185

7. Gapsanddevelopmentneeds...188

Acknowledgements...190

AppendixA. Supplementarydata...190

References...190

1. Introduction

Theuseof robustandappropriate indicatorsthat canassess whetheranecosystemanditsservicesarewellmaintainedandsus- tainablyused(Layke,2009;Walpoleetal.,2009;TEEB,2010)has beenrecognisedasanessentialstepforthepracticalimplemen- tationofconservationandmanagementpolicies(Romboutsetal., 2013).SeveraleffortshavebeenundertakenataEuropeanscaleto evaluatemarineecosystemstructureandtheirresponsetohuman activities,usingkeyindicatorstoassessandsustain“GoodEnvi- ronmentalStatus”(GEnS;Borjaetal.,2011).Theseinitiativeshave beencarriedouttoassisttheMarineStrategyFrameworkDirective (MSFD,2008/56/EC;EuropeanCommission,2008),themainEuro- peanDirectivethatfocusesonmarinewatersandaimsatassessing thestatusofanecosystemunderanthropogenicpressuresandthe requiredinterventionstobringthesystembacktoitsdesiredgood status,makinghumanactivitiessustainable,sincethisisoneofthe objectivesoftheMSFD.ToachieveGEnS,11descriptors,29associ- atedcriteriaand56indicators(frombiological,physico-chemical indicatorsaswellaspressureindicators—includinghazardoussub- stances,hydrological alterations,litter andnoise, and biological disturbancesuchasintroductionofnon-indigenousspecies)have beenidentiﬁed(Cardosoetal.,2010;EuropeanCommission,2010) (Tables2and4).

Despitethefactthatseveralattemptshavebeenmadetoassess theenvironmentalstatusofmarinewatersinanintegrativemanner (Borjaetal.,2011;Halpernetal.,2012;Tettetal.,2013),signiﬁ- cantgapsstillexistonunderstandingmarineecosystemstructures andfunctionsandtheirresponsetohumanpressures(Katsanevakis etal.,2014;Borjaetal.,2013).Currently,ecologicalmodelshave beenrecognisedaspowerfultoolstoevaluateecosystemstructure andfunctionandpredicttheimpactsofhumanactivities(Fulton andSmith,2004;Shinetal.,2004;ChristensenandWalters,2005;

Plagányi,2007;Fulton,2010)andclimatechange(Tomczaketal., 2013;Chustetal.,2014)onmarinesystems.

Thus,thisstudyaimstoassessthemostcommonlyusedcapa- bilityofthemodellingcommunitytoinformonindicatorsoutlined intheEUMSFD(2008/56/EC),focusingparticularlyonbiodiver- sityrelateddescriptors:biologicaldiversity(D1),non-indigenous species(D2),foodwebs(D4),andseaﬂoorintegrity(D6).Todate, therehasbeennothoroughevaluationofthecapabilitiesofeco- logicalmodelstoprovideinformationasexplicitlyoutlinedbythe MSFDindicatorstructure,thistaskhasbeenonlypartiallyunder- taken(e.g.,Reiss etal.,2014).Withthis work,weaimtoﬁll in thisknowledge gapbyproviding aninventoryofmodelsin EU regionalseas thatcouldassess MSFDindicatorsassociatedwith

biodiversity, non-indigenous species, food webs and seaﬂoor integrity.Forthisreason,wehavebuiltamodelcatalogueranging from lower to higher trophic levels, including those that suc- cessfullycouplethetwocompartmentsandassociatedecosystem processes.Thisinventory,developedaspartoftheDEVOTESFP7 Project(http://www.devotes-project.eu/),servestohighlightthe vastpotentialofmodel-derivedindicatorsthatcanbeassociated withMSFD descriptorsand aims toprovidea thorough assess- mentoftheirrelevanceanddegreeof“operationality.”Adetailed descriptionofmodelsandassociatedreferencestogetherwiththe fullcatalogueareprovidedassupplementarymaterials(S1andS2).

Yet,we acknowledgethat this study doesnot aimto serve as review of allthe existing modelsavailable in theliterature, butinsteadhighlightaprocessofexploringmodellingpotential tosupport speciﬁcEuropean policies.Because of thenature of theseissues,though,similarcasestudiesconductedelsewhereare likely tolead tosimilaroutcomes,conclusions,andrecommen- dations (e.g.,because ofsimilar/same model availability and/or processunderstanding).Thus,thisworkemphasisesseveraltypes ofecologicalmodellingandderivedindicatorsthatexistatEUlevel stressinghowsuchdiversityofmodellingapproachescouldbeuse- fultosupportmanagementpoliciesandthelimitationsthatstill occurtoachievethistask.

Inparticular,thisstudyisdividedintosixsections,comprising (1)cataloguestructure;(2)ageneraloverviewofmodelcharac- teristics;(3)modelpotentialtoaddressMSFDGEnSdescriptors andindicators(includingtheabilitytoaddressbiodiversitycompo- nentsandhabitattypes);(4)geographicalcoverageofmodels;(5) abilitytoaddresspressures;and(6)gapsinmodelstype/modelling capabilityandneedsforfurtherdevelopment.

2. Cataloguestructure

Thecataloguehasbeenbuiltprimarilywithmodels/areastar- geted by the DEVOTES partners (which represent 23 research institutionsfromEUandnonEUcountries),yetwithaneffortto integrateavailablemodels/areasfromotherinventories(e.g.,the MEECEprojecthttp://www.meece.eu/Library.aspx)andscientiﬁc literature(seeS1).

The catalogue has been structured with several ﬁelds fol- lowing theMSFDCommissionDecision 2010/477/EU(European Commission,2010)andgroupedintosixmaincategories:

i.Model/Indicatorpropertieswiththefollowingsub-categories:

a.MSFDdescriptor/indicator,descriptor/indicatoroutlinedinthe directive

(4)

C.Piroddietal./EcologicalIndicators58(2015)175–191 177 b.Modelderivedindicator(MDI),indicatorresultantfrommodel

output

c.MDItypedeﬁnedas1.Static(e.g.,snapshotoftheindicator atapreciseperiodoftime),2.Dynamic(e.g.,indicatorwhich changesintime)or3.Spatialdynamic(e.g.,indicatorwhich changesintimeandspace)

d.MDI statusof developmentdeﬁned as 1. Operational, when theindicatorisdeveloped,testedandvalidated(e.g.,itcould beeitheranindicatorusedbytheMemberStates(MS)for national environmentalmonitoring;or in EU/International Conventions’ monitoring programmes; or validated with observed/surveydataalthoughnotnecessarilyapprovedby anynational/internationallaworconvention);2.Underdevel- opment,anindicatorproposalexists,butnotyetvalidatedin ﬁeld/realdata(e.g.,indicatornot yetusedfor MSnational environmentalmonitoringorfor EU/InternationalConven- tions’ monitoring programmes; or not yet validated with surveydata);3.Conceptual,anindicatoridea,supportedby theoreticalgrounds,althoughnopracticalmeasure/metricis yetavailable(e.g.,indicatornotyettested)

e. MDI target/reference values and unit deﬁned as thresh- olds/limitsrepresentingboundariesbetweenanacceptable andunacceptablestatus

f.Modelnamereferringtothelabelusedtoidentifyaparticular model

g.Model type referring to model characteristics/properties and/ortothetechniqueusedtoassessspeciﬁcecosystems h. Datarequirementsreferringtodataneededtorunacertain

model

i.Conﬁdence/uncertaintyreferring to theability ofmodelsto assessuncertaintyfortheinput/outputdataanditisdeﬁned asthetypeofstatisticalanalysisusedtoevaluateit

j.SourceScientiﬁcliteratureandorInstitutionalreportsuppor- tingselectedMDI/modelsentries

ii.Model/MDIinrelationtoMSFDDescriptors:referringtomodels andMDIbroadcapabilitytoaddressthe11descriptorsofthe directive(D1–D11).

iii.Model/MDIcorrespondencewithMSFDBiodiversityIndicators:

referringtomodelsandMDIassessedinrelationtotheircapabil- itytoprovideinformationforthespeciﬁcindicatorslistedunder thecriteriaofthefourdescriptors(D1/D2/D4/D6)asofﬁcially outlinedintheEuropeanCommission(2010).

iv.Model/MDI correspondence with biodiversity components referring to which biodiversity components (e.g., microbes, phytoplanktonand ﬁsh) theindicatorwasrelated toorwas evaluatedwith.Categoriesadoptedforbiodiversitycomponents followed those of theEuropean Commission(2010) and EU CommissionStaffWorkingPaper(CSWP,2012).

v.Model/MDI coverageofspeciﬁc habitattypesandgeograph- ical range/scale referred towhether an MDI was related to certain habitats and geographical areas. Categories adopted forHabitatTypesfollowedthoseoftheEuropeanCommission (2010)andEUCommissionStaffWorkingPapers(CSWP,2011, 2012). Concerning geographical coverage, we have adopted well-establishedinternationalcriteriaforsmallerscalesubdi- visionsorecologicalassessmentareasinordertoincreasethe spatialdetailontheinformationcollected(e.g.,theInternational CouncilfortheExplorationoftheSea(ICES)andGeneralFish- eriesCommissionfortheMediterranean(GFCM)subdivisions;

seemapsunderS1).

vi.Model/MDIrelationtospecificpressures:referringtowhether therewasscientificevidenceofarelationshipbetweenapres- sureandaspecificindicator.Indicatorswererelatedtopressures eitherasresponsive/sensitiveto,oraffectedbyagivenpres- sure(stateindicators,e.g.,mainlythroughchangesintrends) orindicatorswereactuallypressureindicatorsthemselves.The

consideredpressuresfollowthelistofpressuresandimpactsof Annex3oftheMSFD(seeS3).

3. Modelcharacteristics

Themodelcataloguerevealedthatcurrently44 modelshave beenappliedwithoutputsrelevanttoMSFDdescriptors(Table1).

These ecological models being used to describe or understand ecosystemprocessescanbecategorisedunderseventypesofmod- ellingapproachesdescribedbelow:

3.1. Biogeochemicalmodels

Thebulkpropertiesofbiogeochemicalﬂuxesinmarineecosys- temsarecombinedwithinformationonphysicalforcing,chemical cyclingandecologicalstructuretosimulatetheresponseoflower trophic level groups (phytoplankton and zooplankton) to envi- ronmental conditions, including climate variability and change (Gnanadesikanetal.,2011;JørgensenandFath,2011).Suchmodels typicallyhaveverysimpliﬁedrepresentationsofbiologicalorga- nisms,andassociatedtrophicstructure(Anderson,2005).

3.2. Multispeciesmodels

Thesemodelsrepresentpopulationsofdynamicallyinteracting speciesorfunctionalgroups.Somemodelsalsoresolvemultiple stagesorsize-classeswithinpopulations(ChristensenandWalters, 2004;Hollowedetal.,2000;ShinandCury,2001).Focusofthese modelsisonunderstandingtheimplicationoftheindirectinterac- tionsinecosystemsthatresultfromthecomplexnetworksofdirect predator–preyinteractions in marinecommunities. The models aimtorepresent,forexample,top-downorbottom-upeffectsalong marinefoodchain rangingfromprimaryproducers(e.g.phytoplankton)totoppredators(e.g.,marinemammals),ortheroleof indirectcompetitiveinteractionsamongspecies(Fungetal.,2015).

Effectsofexploitationbyﬁsheriesandenvironmentalchangeare alsofrequentlydescribedbythesemodels.

3.3. SpeciesDistributionModels(SDM)/HabitatSuitability Models(HSM)

SDMcombineobservationsofspeciesoccurrenceorabundance withenvironmentalexplanatoryvariables todevelopecological andevolutionaryunderstandingandtopredictdistributionacross selectedhabitats(Elithand Leathwick,2009;Reissetal.,2014).

HSMrelateﬁeldobservationstoasetofenvironmentalvariables (e.g., reﬂecting key factors of theecological niche like climate, topography,geology)toproducespatialpredictionsonthesuit- abilityoflocationsforatargetspecies,communityorbiodiversity (Hirzeletal.,2006).AnewgenerationofSDM/HSM–i.e.dynamic bioclimaticenvelopemodels– nowprovidegreaterlinkstothe mechanisticunderstandingofnicheecology.Suchmodelstypically includeadditionalmodelcomponentsthatdescribephysiological responsesofspeciestotheenvironment,populationdynamicsand dispersal,tofurtherconstrainthedistributionofsuitablehabitat andprovidemorerealisticspeciesdistributionprojections(Cheung etal.,2011).

3.4. Meta-communitymodels

Meta-communityisasetofinteractingcommunitieswhichare linkedbythedispersalofmultiple,potentiallyinteractingspecies.

In this context,meta-community modelsaretheoretical frame- worksdescribingspeciﬁcmechanisticprocessesinordertopredict empirical community patterns. They deal mainly with species

(5)

178C.Piroddietal./EcologicalIndicators58(2015)175–191

Table1

Summarytableofmodelslibraryshowingmodels’name,acronym,datatype(SP:spatial;DY:dynamic;ST:static),numberofmodelderivedindicatorsanduncertainty(VOD:validatedwithobserveddata;VOD*:someofthe indicatorsstillneedtobevalidatedwithobserveddata;NA:notavailable;STAT:statisticalanalysis;BOOT:bootstrap;PE:pedigree).

# Modelname Modelacronym Typeofthemodel Coupled Datatype Modelderived

indicators

Uncertainty

1 EuropeanRegionalSeasEcosystemModel(ERSEM) ERSEM Biogeochemical No SP-DY 2 VOD

2 BlackSeachlorophyllandcoloureddissolved/detrital matter(Chl&CDM)model

BS-Chl&CDM Bio-opticalmodels(remotesensing) No SP-DY 4 VOD*

3 BlackSeamodelofdownwellingradiance(BS-PARModel) BS-PAR Bio-opticalmodels(remotesensing) No SP-DY 1 VOD

4 BlackSeaParticleSizeDistribution(PSD)model BS-PSD(PSC) Bio-opticalmodels(remotesensing) No SP-DY 3 VOD

5 BlackSeaspectralPrimaryProduction(SPP)model BS–SPP Bio-opticalmodels(remotesensing) No SP-DY 1 VOD*

6 BlackSealInherentOpticalPropertiesmodel(IOPs) BS-IOPs Bio-opticalmodels(remotesensing) No SP-DY 3 VOD

7 NorthSeaOpticalProperties(NSOP) NSOP Bio-opticalmodels(remotesensing) No DY 1 STAT

8 1DGeneralOceanTurbulenceModel(GOTM)and EuropeanRegionalSeasEcosystemModel(ERSEM)and EcopathwithEcosim(EwE)

GOTM-ERSEM-EwE Endtoend Yes DY 6 NA

9 PrincetonOceanModel(POM)andBlackSeaIntegrated ModellingSystem-Ecosystem(BIMS-ECO)andEcopath withEcosim(EwE)

POM-BIMS-ECO-EwE Endtoend Yes DY 3 NA

10 RegionalOceanModelSystem(ROMS)andEastern BoundaryUpwellingSystems(BiOEBUS)and Object-orientedSimulatorofMarineecOSystems Exploitationmodel(OSMOSE)

ROMS-BioEBUS-OSMOSE Endtoend Yes SP-DY 5 NA

11 RegionalOceanModelSystem(ROMS)andN2P2Z2D2

biogeochemicalmodelandObject-orientedSimulatorof MarineecOSystemsExploitationmodel(OSMOSE)

ROMS-N2P2Z2D2-OSMOSE Endtoend Yes SP-DY 12 NA

12 NorwegianSeaEcosystem,End-to-End NORWECOM.E2E Endtoend Yes SP-DY 6 NA

13 EcologicalReGionalOceanModel(ERGOM)andModular OceanModel(MOM)andFishModel

ERGOM+MOM+Fish Endtoend Yes DY 2 VOD

14 ECOSystemModel(ECOSMO)andStochasticMulti-Species model(SMS)

ECOSMO-SMS Endtoend Yes SP-DY 2 NA

15 EuropeanRegionalSeasEcosystemModel(ERSEM)and PrincetonOceanModel(POM)andObject-oriented SimulatorofMarineecOSystemsExploitationmodel (OSMOSE)

ERSEM-POM-OSMOSE Endtoend Yes SP-DY 10 NA

16 Hubbell’sneutralmodelofbiodiversity(HNM) HNM Meta-community No ST 1 NA

17 EcopathwithEcosim(EwE) EwE Multispecies No ST-DY-SP 136 PE-VOD*

18 NorthSeaThresholdgeneraladditivemodels(NStGAM) NStGAM Multispecies No DY 4 BOOT

19 Population-DynamicalMatchingModel(PDMM) PDMM Multispecies No DY 1 VOD

20 BayofBiscayQualitativetrophicmodel BoBQualit Multispecies No ST 1 NA

21 Length-basedmultispeciesmodel(LeMANS) LeMANS Multispecies No DY 2 VOD

22 StochasticMulti-Speciesmodel(SMS) SMS Multispecies No DY 2 VOD

23 ProudmanOceanographicLaboratoryCoastalOcean ModellingSystem(POLCOMS)andEuropeanRegionalSeas EcosystemModel(ERSEM)

POLCOMS-ERSEM Physical

(hydrodynamic)–biogeochemical

Yes SP-DY 6 NA

24 3DGeneralEstuarineTransportModel(GETM)and EuropeanRegionalSeasEcosystemModel(ERSEM)

GETM-ERSEM Physical

Yes SP-DY 16 VOD*

25 PrincetonOceanModel(POM)andBlackSeaIntegrated ModellingSystem-Ecosystem(BIMS-ECO)

POM-BIMS-ECO Physical

Yes DY 4 NA

(6)

C.Piroddietal./EcologicalIndicators58(2015)175–191179

26 St.PetersburgEutrophicationModel(SPBEM) SPBEM Physical

Yes SP-DY 7 VOD

27 EuropeanRegionalSeasEcosystemModel(ERSEM)and PrincetonOceanModel(POM)

ERSEM-POM Physical

Yes SP-DY 11 NA

28 3DGeneralEstuarineTransportModel(GETM)and EcologicalRegionalOceanModel(ERGOM)

GETM-ERGOM Physical

Yes SP-DY 8 VOD*

29 BAlticSeaLong-Termlarge-ScaleEutrophicationModel (BALTSEM)

BALTSEM Physical

Yes SP-DY 7 VOD

30 BiogeochemicalFluxModel(BFM)andPrincetonOcean Model(POM)

BFM-POM Physical

Yes SP-DY 5 NA

31 BlackSeaEcosystemModel BSEM Physical

Yes SP-DY 13 VOD*-STAT

32 EcologicalReGionalOceanModel(ERGOM)andModular OceanModel(MOM)

ERGOM+MOM Physical

Yes SP-DY 7 VOD

33 ECOSystemModel(ECOSMO) ECOSMO Physical

Yes SP-DY 6 NA

34 MOHIDandPelagicBiogeochemicalModel(LIFE) MOHID-LIFE Physical

Yes SP-DY 4 VOD*

35 NucleusforEuropeanModellingoftheOceans(NEMO)and BiogeochemicalFluxModel(BFM)

NEMO-BFM Physical

Yes SP-DY 10 NA

36 RegionalOceanModelSystem(ROMS)andEastern BoundaryUpwellingSystems(BiOEBUS)

ROMS-BioEBUS Physical

Yes SP-DY 6 NA

37 RegionalOceanModelSystem(ROMS)andN2P2Z2D2

biogeochemicalmodel

ROMS-N2P2Z2D2 Physical

Yes SP-DY 12 NA

38 SwedishCoastalandOceanBiogeochemicalmodel(SCOBI) andRossbyCenterOceancirculationmodel(RCO)

RCO-SCOBI Physical

Yes SP-DY 7 VOD

39 EcologicalNicheFactorAnalysis(ENFA) ENFA SDM/HabitatSuitabilityModels No ST 1 NA

40 BayofBiscayHabitatsuitabilitybasedonGeneralised AdditiveModels(GAM)

BoBGAM SDM/HabitatSuitabilityModels No ST 1 NA

41 BayofBiscayHabitatsuitabilitybasedonGeneralised LinearModels(GLM)

BoBGLM SDM/HabitatSuitabilityModels No ST 1 NA

42 HabitatsuitabilitybasedonMaxEnt(MaximumEntropy) MaxEnt SDM/HabitatSuitabilityModels No ST 2 NA

43 Niche-TraitModel(NTM) NTM SDM/HabitatSuitabilityModels No ST 1 NA

44 Process-drivenhabitatmodel PDH SDM/HabitatSuitabilityModels No ST 1 NA

(7)

180 C.Piroddietal./EcologicalIndicators58(2015)175–191 compositionand abundanceand theirvariation withina meta-

community(Huguenyetal.,2007).

3.5. Bio-opticalmodels

Theopticalpropertiesofbiologicalmaterials,suchasphyto- planktonicor heterotrophicunicellular organisms, are analysed andthenmodelledtopredictdistributionsofbiologicalcommuni- tiesoverwidespatialareas(withremotesensingdata)orinterms ofexpecteddepthlimitationsthatcanbeinferredfrommodelling studies.Bio-opticalmodelsarebasedonvariousfundamentalthe- oriesofopticswhichapplytoasingleparticlemakinguseofasetof equations/algorithms(MorelandMaritorena,2001;IOCCG,2006).

3.6. Hydrodynamic–biogeochemicalModels

Thesearemainlycoupledhydrodynamicandbiogeochemical modelstocaptureglobalscalepatternsinphysical–chemicalcom- ponentsaffectinglowertrophiclevelgroups(e.g.,phytoplankton andzooplankton)(Gnanadesikanetal.,2011;JørgensenandFath, 2011).

3.7. End-to-endmodels

Inrecentyears,hydrodynamic-biogeochemicalmodels(orjust biogeochemical models) have been coupled with multispecies models.Thesesocalledend-to-end(E2E)modelscombinephysi- cochemicaloceanographicprocesseswithorganismsrangingfrom lowtrophiclevel(LTL)tohighertrophiclevelorganisms(HTL)into asinglemodellingframework(Traversetal.,2009).

Ofthemodelsreportedinthisstudy,morethanhalfwerecou- pledecologicalmodels(Table1).Themostcommontypeofmodels currently in the catalogue were hydrodynamic-biogeochemical models (36%) followed by end-to-end (18%), species distribu- tion/habitatsuitability,bio-opticaland multispecies (14%each), biogeochemicalandmeta-community(2%each)models(Table1).

In the framework of ecological studies, physical–biological interactionsarethemainfactorsthatcanbetterdescribeecosystem propertiesandthespatialand/ortemporalevolutioninfunction ofrelevantpressuresidentiﬁed,climatechangeoranthropogenic impacts.Thisis reﬂectedinthechoiceof modellingapproaches andinthegrowingneedtocoupledifferenttypesofmodelswithin a singlemodellingframework (Traverset al., 2009;Roseet al., 2010).Thisisparticularlytrueifthemodelsareintendedtopredict changesandprovideguidanceinaframeworkofbiodiversitycon- servationandecosystem-basedmanagement(Traversetal.,2009;

Kaplanetal.,2012).

Recentsoftwaredevelopments,withinthecurrent(DEVOTES) andformerEUprojects(e.g.,MEECEhttp://www.meece.eu/),have shownthatthesemodels(hydrodynamic-biogeochemicalandmul- tispeciesmodels)canbecoupledtoruntogether.Thisrepresentsa powerfultoolforscenariotestingofclimatechangeandanthro- pogenic impacts simultaneously. There is a growing trend for E2Emodelling,whichincludesanthropogenicandphysicaldrivers behind observed changes, identifying both direct and indirect causes(Fulton,2010;Shinetal.,2010b;Travers-Troletetal.,2014), andsobetterfacilitatesthesettingoftargetsandimplementation ofmanagementmeasures(Curyetal.,2008;Kaplanetal.,2012).

Fig.1illustratesthecapacityofthesevenmodeltypestorepre- sentthedifferentcomponentsofmarineecosystems,includingor excluding,humancomponentsand/orclimateimpacts.

Coupled(bothE2Eandhydrodynamic-biogeochemicalmodels) andbio-optical(remotesensing)modelsincludedinthiscatalogue wereprimarilyspatiallydynamicand5outof30modelswerealso

dynamic.Theremainingmodelsweremainlystaticwithonly5 outof14modelspresentingdynamicandspatialmodulesaswell (Table1).Thisisanimportantandinterestingresultsincespatial- dynamicmodelsareabletoprovidegreatercapacityforforecasting ofecosystemdynamics,althoughtheyrequireamoredatainten- sivecalibration(e.g.,theinitialtestingandtuningofamodel)and validation(e.g.,thecomparison/ﬁttingofmodelwithadataset representing“local”ﬁelddata)approaches(Jørgensen,2008).

Atotalof201model-derivedindicators(seeS1ofsupplemen- tarymaterials)wereincludedinthiscatalogue,ofwhichmorethan halfwereconsideredtobe“operational”(64%),whilethemajor- ityoftheremainderwerestill“underdevelopment”(33%),with onlyafew“conceptual”approaches(3%)presented(Table2).We acknowledgethatsomeindicatorsmighthavechangedtheirstatus sincethetimeofthissurvey(e.g.,someindicators“underdevelopment”mayhavebeenassessedandnowclassiﬁedas“operational”) butforthepurposeofthisworkwedecidedtokeeptheminthe statusofdevelopmentthattheywerereportedduringthesurvey.

EcopathwithEcosim(EwE)wasnotablyassociated withthe largestnumberofmodel-derivedbiodiversityindicators(Table2).

However, the majority of these biodiversity indicators were biomassesofspeciesorgroupsofspeciesatdifferenttrophiclevels ofthefoodweb.Foreaseofcharacterisation/evaluation,model- derivedindicatorsweregroupedintosevenmajorcategories(see Table3forthedetailedlist).Notsurprisingly,biomassindicators constitutedthe largestgroupwithapproximately57% followed bydiversityindices(13%)andphysical,hydrologicalandchemical indicators(12%).Regardingtargetsand/orreferencevaluesassoci- atedwithmodel-derivedindicators,thecataloguehighlightsthat only fewmodels in fewareas had assigned target orreference values,despitethefactthatthemajoritywereconsidered“operational”(i.e.developed,testedandvalidated).Thisisthecaseoffully developedmodelsforwhichvalidatedoutputsexist(e.g.,BSEMby Dorofeevetal.,2012),butunderpolicycontextssuchastheMSFD, lacktestedandvalidatedreferencevaluesortargetscompliantwith speciﬁclegalrequirements.

Also,veryfewofthereportedmodelshavebeenusedtoclearly assesstheeffectsofmeasurestomeetthetargetsthatwilleven- tuallybeestablished.Forinstance,multispeciesmodelshavebeen appliedintheIonianSeaandintheNorthSeaecosystemstoassess thereductioninﬁshingeffortasameasureto(a)bouncebackcom- mondolphinpopulations(e.g.,EwEmodelbyPiroddietal.,2011);

(b) assess theresponse of selected biodiversity indicators (e.g., PDMMbyShephardetal.,2013;Fungetal.,2013,orEwEmodelby LynamandMackinson,inpress);(c)testtheeffectofselectivefish- ingoncommunitybiodiversityconservation(e.g.,LeMANSmodel byRochetetal.,2011)andimplementedintheBayofBiscay(e.g., OSMOSEmodelbyChiffletetal.,2014)toevaluatetheeffectof differentfishingscenariosonsmallpelagicfishstocks.

Inaddition,notallthemodelswereabletoaddressuncertainty;

themajority(61%)lackedanapproach todetermineconfidence intervals/rangeofuncertaintyorrequiredfurthervalidationwork forindicators.Thisisareflection,asmentionedabove,ofthetypeof datapresentinthecataloguewhicharemorespatial-dynamicthan staticandforwhichvalidationismoredifficulttoobtain.Fromthe modelsthatreportedaddressinguncertainty(39%),datacompari- sonanddatavalidation(e.g.,modeloutputsfittedtosurveyeddata) wasthemostcommonmethodreported(Table1).

4. Modelpotentialtoaddressdescriptorsandindicatorsfor biologicaldescriptors

IntermsofsupportingtheMSFD,ecologicalmodelscanbethe mosteffectivemeanstomodelrelationshipsbetweenactivities, pressures,stateand thusindicators(Jørgensen, 2008;Jørgensen

(8)

C.Piroddietal./EcologicalIndicators58(2015)175–191 181

Fig.1. Illustrationofmodelscapacitytodescribetheecosystem,fromspeciﬁcprocessesintegratingbiologicalcompartmentsandtheassociatedabioticenvironmenttothe entireecosystemincluding,ornot,humancomponentsorclimateimpacts.Inparticular,1and7–refertobiogeochemicalandcoupledphysical–biogeochemicalmodels;

2and3–refertomultispeciesmodels(eitheratspeciesoratfoodweblevel);4–Speciesdistribution/HabitatSuitability;5–meta-communitymodelsand6–bio-optical models.E2Emodelsencompassallofthem.

and Fath,2011).Thisis because ofthe integrative characterof thesemodellingapproachesthatoftenconsidermanyecosystem componentsfromabioticfactorstobiotic interactions andpro- cesses.The 44 modelsavailable in the catalogue werecapable ofaddressingindicators in8of the11descriptors oftheMSFD (Table2)although,duetothefocusofthissurveywhichprimarily dealtwiththefourbiodiversityrelateddescriptors,theirmodelling potentialwasstrongerfortwo ofthesebiodiversitydescriptors:

biologicaldiversity(D1)andfoodwebs(D4).Nevertheless,human inducedeutrophication(D5),hydrographicalconditions(D7)and commercialﬁshandshellﬁsh(D3)werewelladdressedbythemod- elsinthiscatalogue.

Within the biodiversity related descriptors, non-indigenous species (D2) and seafloor Integrity (D6) were the most poorly addressedbythemodelscurrentlyinthecatalogue(Table2).How- ever,Pinnegaretal.(2014)showshowEwEmodelscanbeuseful inassessingtheresponseofanecosystemtotheintroductionof invasivespecies(D2).Similarly,increasingthespatialresolution ofmanyofthecurrentmodelswouldfurtherimproveourunder- standingofthedirecteffectoffishingandotheractivities(such asdecommissioningofoilrigsordevelopmentofa windfarm) onseafloorintegrity(D6).Inseveralcases,modelshavebeenused toinvestigatetheimpactsoftrawlingandtestfisheriesscenarios (e.g.,highresolutionERSEM-POMmodel,Petihakisetal.(2007)).

However,mostofthemodelsconsideredinthiscataloguedonot explicitlyincludedescriptionsofthesetypesofpressuresonthe marineenvironment,theydonot linktobenthichabitatlayers,

andtheirunderstandingofpressuresandimpactsisinmanycases stilllimitedbyscarceempiricalinformation(HooperandAusten, 2014).

Typically,asinglemodelwascapableofaddressingmorethan oneMSFDdescriptorandsometimesuptosix,asisthecaseofEwE (Table2).Asaresult,thesamemodelmaybenotedforhavingindi- catorsinmultiplestagesofdevelopment(e.g.,operational,under developedorconceptual)eitheracrossdescriptorsorwithinthe samedescriptor.Thisis becausethereportedstatusofdevelop- mentrelatesnottothemodelitselfbuttothedifferentindicators thatcan bederived fromthemodel.Thepotentialoftheavail- ablemodelstoaddressMSFDindicatorsspeciﬁcallythosewithin biologicaldescriptorswasevaluatedbyextractingthenumberof indicators(outlinedintheEuropeanCommission(2010))thateach modelcaninformon(Table2).Allmodelscouldaddressmulti- pleindicators,fromthesetof21MSFDindicatorsunderthese4 descriptors.Infact,20modelsinthecataloguehadthepotentialto addressatleasthalfoftheseindicators.Despitethehighpotential ofthemodelstoaddressMSFDindicators,notalloftheavailable model-derivedindicatorswerefullyoperational(seeSection2for deﬁnitionandTable4).Themeanpercentageofoperationalmodel- derivedindicatorsacrossallMSFDindicatorswas64%.Ouranalysis alsorevealedthattherewerethreeindicatorsrequiredunderthe biodiversity descriptors for which nomodel-derived indicators were available in thecatalogue (Table 4): D1C3-I2: population geneticstructure;D2C2-I1:Ratiobetweeninvasivenon-indigenous speciesandnativespeciesandD2C2-I2:Impactsofnon-indigenous

(9)

182C.Piroddietal./EcologicalIndicators58(2015)175–191

Table2

Models’capabilityperthe11MarineStrategyFrameworkDirectivedescriptors(D)assessedbythenumberofindicatorsprovidedbyeachmodel(fornames,seeTable1).Thedevelopmentstatusoftheindicatorsisindicated(op:

operational,ud:underdevelopment,co:conceptual).ThelastcolumnsummarisesthenumberofMSFDofﬁcialindicators(EuropeanCommission,2010)ofD1,D2,D4andD6(checkTable4)thatthemodel-derivedindicators caninformon.

D1 D2 D3 D4 D5 D6 D7 D8 D9 D10 D11 #MSFDindicators

addressedunder D1,D2,D4,D6 Biological

diversity

Non- indigenous species

Commercial ﬁsh

Food webs

Human- induced eutrophication

Seaﬂoor integrity

Hydrological alterations

Contaminants Contaminants infood

Marine litter

Energy/

noise

1 BALTSEM 7op 5op 3op 2op 16

2 BFM-POM 5op 3op 2op 2op 14

3 BSEM 6op/7ud 1op/1ud 1op/7ud 4ud 3op 9

4 EwE 82op/82ud/7co 1ud 53op/57ud/4co 82op/82ud/7co 13op/14ud/2co 17op/25ud/4co 13(+1^a)

5 ECOSMO 6op 3op 2op 3op 14

6 ECOSMO-SMS 2ud 2ud 2ud 8

7 ENFA 1op 1op 1op 14

8 ERGOM+MOM 7op 5op 3op 2op 16

9 ERGOM+MOM+ﬁsh 2op 2op 2op 7

10 ERSEM 2ud 2ud 1ud 12

11 ERSEM-POM 11op 6op 3op 5op 14

12 ERSEM-POM-OSMOSE 10ud 10ud 10ud 9

13 BoBGAM 1op 1op 1op 16

14 GETM-ERGOM 8ud 2ud 4ud 6ud 14

15 GETM-ERSEM 16ud 5ud 8ud 2ud 11ud 19

16 BoBGLM 1op 1op 1op 16

17 GOTM-ERSEM-EWE 6ud 4ud 6ud 3ud 8

18 HNM 1co 1co 1co 1co 16

19 BS-IOPs 3ud 2ud 3ud 8

20 LeMANS 2op 2op 2op 7

21 MaxEnt 2op 1op 1op 2op 17

22 MOHID–LIFE 4op 3op 3op 1op 10

23 NEMO-BFM 10ud 7ud 4ud 3ud 17

24 NSOP 1ud 1ud 1ud 8

25 NStGAM 4ud 2ud 4ud 1ud 10

26 NORWECOM.E2E 6op 3op 2op 3op 14

27 NTM 1ud 1ud 1ud 9

28 PDMM 1op 1op 1op 7

29 POLCOMS-ERSEM 6op 3op 2op 3op 14

30 POM-BIMS-ECO 4op 3op 2op 1op 14

31 POM-BIMS-ECO-EWE 3ud 3ud 3ud 9

32 PDH 1ud 1ud 1ud 11

33 BS-PSD(PSC) 3ud 3ud 3ud 5

34 BoBQualit 1co 1co 1co 8(+1^a)

35 RCO-SCOBI 7op 5op 3op 2op 16

36 BS-Chl&CDM 4ud 4ud 4ud 6

37 BS-PAR 1ud 3

38 BS-SPP 1ud 1ud 1ud 3

39 ROMS-BioEBUS 6op 3op 2op 3op 14

40 ROMS-BioEBUS-OSMOSE 5ud 5ud 5ud 9

41 ROMS-N2P2Z2D2 12op 8op 5op 4op 13

42 ROMS-N2P2Z2D2-OSMOSE 12op 12op 12op 11

43 SMS 2op 2op 2op 7

44 SPBEM 7op 5op 3op 2op 16

Numberofmodelsperdescriptor 44 3 17 43 26 5 17 0 1 0 0

aNewproposalsforDescriptor4FoodWebs,notyetconsideredunderthesetofIndicatorsoutlinedintheEUCommissionDecision(EuropeanCommission,2010).

(10)

C.Piroddietal./EcologicalIndicators58(2015)175–191 183

Table3

Themodel-derivedindicatorsgroupedinto7majorcategories,basedonwhatthe indicatorsinformon,withtheiroverallpercentagesintheDEVOTESCatalogueof model-derivedindicators.

Typeofindicators %

1 Biomass 57

2 Diversityindicators Biodiversityindices(e.g., Kemptondiversityindex, trophiclevelofthe community)and species/habitatdiversity, proportionsincommunity

13

3 Primaryorsecondary production

9 4 Spatialdistribution

indicators

Speciesspatialdistribution 6 5 Specieslife-history Traitssuchasfore.g.,

length,weightorlifespan 1 6 EcologicalNetwork

Analysis(ENA)indicators

Flows,energiesand efﬁciencies

2 7 Physical,hydrologicaland

chemical

Describingeitherhabitat integrityorpressures

12

invasivespecies at thelevelof (1) species, (2)habitats and (3) ecosystem.

Additionally,itisnoteworthythatthepotentialofmodelling approachestoaddressecosystemfundamentalpropertiessuchas D1C8I1“Interactionsbetweenstructuralcomponents”andD1C8I2

“Servicesprovided”(Table4)washigh.Theseaspects,despitebeing clearlymentionedintheEuropeanCommission(2010),werenot partoftheMSFDindicatorsset,mostprobablydue tothediffi- cultyindefiningthem throughspecificindicators.Nevertheless, themajorityofthemodel-derivedindicatorsincludedinthiscat- alogue(189outofthe201)havethepotentialtoinformonthese complex,integrativeecosystemdimensions.Inanycase,although thecatalogueshowsthepotentialofmodelstoaddressEcosystem Services(ES,sensuLiqueteetal.,2013),thesurveyperformedcan- notinformadequatelyonthecapacityoftheindicatorstosupport policy-makers’useoftheseESconcepts.Thisisacurrentlimita- tionoftheMSFDsetofindicators(Table4)whichdoesnotclearly requiretheassessmentofecosystemsservices,despitethefactthat in2011,asapartyoftheConventiononBiologicalDiversity(CBD), theEuropeanUnion(EU)adoptedanewstrategy(theBiodiversity Strategyto2020),whichintegratesESaskeyelementsforthecon- servationapproachtobiodiversity(Maesetal.,2012).Theroleof ESinsupportingconservationinitiativesandsocio-economicactiv- itiescallsforactiontomonitor,quantifyandvaluetrendsinthese services,soastoensurethattheyareadequatelyconsideredin decisionmakingprocesses.Todoso,aclearlinkageneedstobe establishedbetweenbiodiversityandecosystemfunctioningand thediversityandcomplexityofthebenefitstheyprovide,i.e.the ecosystemsservices(beitprovisioning,regulatingorcultural),in ordertoallowthedevelopmentofoperationalindicators.Yet,the indicatorsavailablearenotcomprehensiveandareofteninade- quatetocharacteriseES;dataareofteneitherinsufficientorthe linkagesarepoorlyunderstoodtosupporttheuseoftheseindica- tors(Liqueteetal.,2013).

4.1. Biodiversitycomponentsandhabitats

Habitatsand speciesarekeyattributesofbiologicaldiversity andtheiroccurrence,distributionandabundanceisusedascriteria toassesstheecosystemstatus(Table5).ToattainGEnSforD1,as statedintheMSFD,“nofurtherlossofbiodiversityatecologically relevantscaleshouldoccur,and,ifitdoes,restorationmeasures shouldbeputinplace”.ThedeﬁnitionofGEnSisdependenton theecologicalrelevanceandisapproachedatdifferentscalesof

complexity,fromspeciestohabitats,communitiesandecosystem (seeBorjaetal.,2013).

Biodiversity components indicated in the MSFD include microbes, phytoplankton, zooplankton, angiosperms, macroal- gae,benthicinvertebrates,fishes,cephalopods,marinemammals, reptilesandbirds,withspecificsubgroupswithinthelastfourcat- egories.Theirinclusioninecologicalmodelslistedinthecatalogue washighlyheterogeneous.Operationalmodel-derivedindicators concernedmainlyfish,phytoplankton,zooplankton,benthicand pelagicinvertebratesandmarinemammals(total64,45,31,23,and 17,respectively)(Fig.3),whiletheremainingbiodiversitycompo- nentswerecoveredwithlessthan10indicatorseach.Thisreflects thetraditionalfocusofmarineecosystemmodelling,drivenmainly bythewide-spreaduseoflowtrophiclevelmodelsrelatedtothe bottom-up forcing of production, and in parallel, motivatedby fisheriesorientedpoliciesandconservationinterestsinparticular species(Roseetal.,2010;Shinetal.,2010b).

Asexpected,thevariousmodelshaveusedsimilarcomponents differently and, dependingontheirﬁnal goal,theresolution of thebiodiversitycomponentsdifferedgreatly:fromsingletomulti- speciesmodels,inclusionofsingleormultiplefunctionalgroups andintegratingbothLTLandHTLkeyorganisms(e.g.,Oguzetal., 1999;LewyandVinther,2004;Schrumetal.,2006;Colletal.,2008;

Rossberg etal., 2010;Lassalle etal.,2011; Mateuset al.,2012;

Tsiarasetal.,2012).Ofthemodelscatalogued,onlyHubbell’sneu- tralmodelandthePopulation-DynamicalMatchingModel(PDMM) resolvebiodiversityatspecieslevel,andonlythePDMMdoesso through the entire marine foodchain (Fung et al., 2013). EwE model-derivedindicators, either operational,conceptualor still underdevelopment,havebeenusedtomodelalltypesofbiodi- versitycomponents(excludingmicrobes),withﬁshbeingthemost frequentlyassessedgroup(25%)followedbybenthicinvertebrates (15%),marinemammals(12%)andcephalopods(11%).Themicro- bialcomponent,asreportedinthecatalogue,wasonlyevaluatedby ERSEM-POMintheAegeanSeaandunderdevelopmentbyNEMO- BFM in theBaltic Sea. Whenmodelswere organisedaccording tomodeltype,multispeciesmodelsassessedthemajorityofbio- diversitycomponentswiththeexceptionofmicrobesthat were mostlyevaluatedbycoupledhydrodynamic–biogeochemicalmodels(Fig.3).

Thepredominanthabitattypesthatshouldbeassessedwithin theevaluationofthestatusundertheMSFD arewater-column, seabedand icehabitats,withecologicalmodelsreferringtoone orseveralofthesehabitats.Inourcatalogue,ofallpredominant habitats, water-column was the most comprehensively evaluated habitat,eitheronits own,or in relationtothe othertwo habitats. There were only two instances where seabed habitats were evaluatedontheir own.Ice-associated habitats were assessedbyhydrodynamic–biogeochemicalandmultispeciesmod- els while seabed habitats were evaluated in multispecies and SDM/Habitatsuitability/Communitymodels.Multispeciesaswell ascoupled(bothhydrodynamic–biogeochemicalandE2E)mod- elsweremainlyusedfortheassessmentofspeciesorgroupsof species/organismsthatcanbelinkedtowater-columnhabitats.

Examiningtheintersectionbetweenmodel-derivedindicators and habitats, the water column was the most widely covered habitat,speciﬁcallythecontinentalshelfwhereallcomponentsof biodiversitywerecovered(Table5).Themarineoceanicwatercol- umnwasalsowidelycovered;however,inthiscasemicrobeswere notevaluated.Inestuaries,onlyphytoplanktonandzooplankton wereassessed,whichwerealsothemaincomponentsmodelled inice-associatedhabitats.In theseabedhabitat,shallowsublit- toralmixedsedimentswerethemostcommonlyevaluatedwith model-derivedindicators assessing7 out ofthe11 biodiversity components.Invertebratesweremainlystudiedinrelationtothe water columnover thecontinental shelfalthough theyare also