MODELS – PLOTTING THE COURSE IN UNCHARTED WATERS · MODELS – PLOTTING THE COURSE IN UNCHARTED WATERS Ecological models to support beCer management and policy implementaon Chair:

MODELS–PLOTTINGTHECOURSEINUNCHARTEDWATERS

EcologicalmodelstosupportbeCermanagementandpolicyimplementaFon

Chair:ChrisFanWilson,OceanDTM

DEVOTESFinalConference,17–19October2016,Brussels

PRESENTORSINBLOCK3(14:00–15:15)LauraUusitalo(SYKEFinnishEnvironmentInsFtute)Next:GuillemChust,AZTIChristopherLynam(Cefas)ChristopherLynam(Cefas)

BiologicalconnecFvityatgeneFcandcommunitylevelsAuthors:GuillemChust,ErnestoVillarino,AnneChenuil,XabierIrigoien,NihayetBizsel,AntonioBode,CecilieBroms,SimonClaus,MaríaL.FernándezdePuelles,SerenaFonda-Umani,GaliceHoarau,MariaG.Mazzocchi,PatricijaMozeFč,LeenVandepiCe,HelenaVeríssimo,SoultanaZervoudaki,AngelBorjaLeader:AZTI,Partners:CNRS,PML,HCMR,MHI-NASU,DEU,CONISMA

DEVOTESFinalconference,17-19Oct2016,Brussels

ProxiesofconnecFvityandNeutraltheories

Natural selection / Adaptative process (Darwin 1859) versus Neutral theory of molecular evolution (Kimura 1983)

• “most of evolutionary changes at the molecular level is the result of randomly genetic drift acting on neutral alleles (not affecting fitness)”

• ‘stepping stone model’ of dispersal: populations tend to exchange migrants (or propagules) with nearest neighbours along a coastline

Species - Evolution

Niche theory adaptive processes (Hutchinson 1957) versus Neutral theory of biodiversity (Hubbell 1997, 2001) • “In an ecological community of trophically similar species (i.e. neutral), diversity arises at random (ecological drift), as each species follows a random walk”

• when migration rate is low (i.e. species are dispersal limited), species similarity declines with geographical distance

• neutral: all individuals are assumed to have the same prospects for reproduction and death

Communities - Ecology

3

www.azti.es 22/10/16

Genetic level

4

0

1

Pairw

isecommun

itysimilarity

Pairwisegeographicdistance

mhigh

mlow

0

1

Pairw

isegene

ticsimilarity(1-F

ST)

Pairwisegeographicdistance

mhigh

mlow

Community level

Proxies of connectivity and Neutral theories


Aim:Tocomparecommunity-basedanalysiswithpopula:ongene:canalysisin3differentgroupsacrossregional(European)seas[Task4.3.5]:• Phytoplankton• Zooplankton• BenthicMacroinvertebrates(Sok-boCom)Hypothesis:planktonspecieshaveahigherdispersaldistancecomparedtobenthicmacroinvertebratespeciesatthetwolevelsoforganizaFon,basedupontheconstraintsinmovementwithinthebenthichabitatwithashortplanktoniclarvalphaseofmacroinvertebratesRa$onale:toproposevulnerabilityindicatorsbasedonconnec$vityproxies

5

Modelling of marine connectivity and biodiversity across regional seas

www.azti.es 22/10/16 6

Dispersal types in benthic macroinvertebrates

Direct developers: a larval stage has very low dispersal potential and usually looks like the adult form of the animal. Lecithotrophic larvae: generally have greater dispersal potential than direct developers. Many fish species and some benthic invertebrates have lecithotrophic larvae, which are provided with a source of nutrition to use during their dispersal, usually a yolk sac. Short pelagic larval durations and do not disperse long distances. Planktotrophic species: generally have fairly long pelagic larval durations and feed while in the water column. Consequentially, they have the potential to disperse long distances. This ability to disperse is one of the key adaptations of benthic marine invertebrates. During their time in the water column, planktotrophic larvae feed on phytoplankton and small zooplankton, including other larvae.

Dendrasterexcentricus

Cucumariaminiata

Heliocidariserythrogramma

Dispe

rsinglarvae(P

elagicpha

se)

NDL

www.azti.es 22/10/16 7

Community analysis

Datasets(community):Phytoplankton• 36fixedstaFons• 555species• 7environ.variables:SST,Salinity,Oxy,Secchid.,nutrientsZooplankton:• 27fixedstaFons• 179species• 6environ.variables:SST,Salinity,Oxy,chla,depth,Diffusion

Benthicmacroinvertebrates• 593(aggregated)staFons• 2276species• 7environ.variables:SST,Salinity,Depth,Secchid.,nutrients,diffuseaCenuaFoncoeff.


Datasets(community):Phytoplankton(diatoms,dinoflagellates,andcocolithophorids)

•  36sta:onsofphytoplanktoncommunitydatafrommostofEuropeanSeas.•  555speciesrecorded.•  Regionsandreferences:

–  AtlanFcMarginandNorthSea(Saletal.,2013,Ecologydatapaper),–  South-easternBayofBiscay(ref.fromÁngelBorjaBasquequalitynetwork),–  KaCegatstraitandsouth-westernBalFcSea(reffromJesperandJacob),–  BlackSea:SinopBay(Monchevaetal.1995,TurkogluandKoray2002,Turkogluand

Koray2004,Bircan2005,BüyükhaFpoğlu2002,Feyzioğlu2010,Seyhan2005,GomezandBoicenco2004,Batetal.2011)

–  MediterraneanSea:GulfofTrieste(staFon99,NorthAdriaFc,MozeFcetal.,2012ECSS),SeaofMarmara(Uysal1996,Polat2000,Balkıs2003,Balkısetal.2004,Balkıs2004,DenizandTaş2009,Tufekcietal.2010,Altug2007,Polat(2011),Tas2011),IzmirBay(easternAegeanSea,Koray,1985,Koray,1995,MeFn,1995,Çolak,2000,Bizsel,2000,Göçer2001,Sabancıetal.2005,Inanan2007,BizselandNezan(2007),ÇolakandKoray(2007),Tümer2012).

8

Community analysis


Datasets(community):Sok-boComMacroinvertebrates3datasources:1) thepan-EuropeanMacroBendatabase(VandenBergheetal.2009;SeeotherMARBEFMEPSpapers,sentbyLeenVandepiCe;availableatEMODnetBiologyportal(hCp://bio.emodnet.eu/portal)),coveringtheIrishSea,NorthSea,NorwegianSea,BarentsSea,GulfofLion(easternMediterranean),with1814samplingsthatwerespaFallyandtemporallyaggregatedinto305staFons.2) Basquewaterqualitynetwork(19fixedcoastalstaFons,sampledduring2003-2008,andspaFallyandtemporallyaggregatedinto17staFons;seeBorjaetal.,2009;Pascualetal.,2011)coveringtheBasquecoast(south-easternBayofBiscay),withenvironmentaldataassociated.3) DanishdatasetcoveringKaCegatstraitandsouth-easternBalFcSea(ref.JacobCarstensenandJesperHarboAndersen;1415samplings,sampledduring1990-2013,thatwerespaFallyandtemporallyaggregatedinto271staFons).• Thesamplingconsideredallmacroinvertebratestaxasampledwithagrabof0.04to1m2surfaceofsok-boComsediment,withmostofthemused0.1m2.• StaFonsrestrictedtosamplesbetween0and450mdepth(removingallstaFonswithoutdepthinformaFon),between1990and2013(toreduceheterogeneityintemporalchanges),andataminimumdistancebetweensamplesof10km(otherwisetheyareaggregated).• 593(aggregated)sta:onsand2276species.

9

Community analysis

www.azti.es 22/10/16 10

•  EnvironmentaldatahasbeenobtainedfrominformaFonassociatedtothebiologicalstaFonsandaveragedincaseofmulFplesamplings,andifnotavailable,fromBio-Oracle(Tybergheinetal.2012).

•  Forphytoplankton,7environmentalvariableshavebeenanalysed:SST(extractedfromdata-associatedsamplingandfromBio-Oracle),seasurfacesalinity(samplingandBio-Oracle),dissolvedoxygen(Bio-Oracle),Secchidepth(sampling),ammoniacal-nitrogen(NH4-N)(sampling),totalNitrogen(sampling),andtotalPhosphorus(sampling).

•  Forzooplankton,6environmentalvariables:depth,SST,salinity,dissolvedoxygen,chlorophyll-a,andthediffuseaCenuaFoncoefficient.

•  Formacroinvertebrates,7environmentalvariableshavebeenanalysed:depth(sampling),SST(Bio-Oracle),seasurfacesalinity(Bio-Oracle),dissolvedoxygen(Bio-Oracle),nitrate([NO3]and[NO3+NO2],fromBio-Oracle),phosphate(ortho-phosphateconcentraFon[HPO4-2],fromBio-Oracle),anddiffuseaCenuaFoncoefficientat490nm(m-1)(Bio-Oracle).

Environmental data

Community analysis

Distanceindices

•  Oceanographic distance: minimum path distance between two stations, i.e. circumventing the terrestrial zone

A

BdA-B

•  Pairwise community similarity (βsim) which express the proportion of shared species with respect to the minimum number of species of the two sites (adapted for non-equal sampling areas)

b c a acba

sim +−=

),min(1β

(Lennon et al. 2001. J Animal Ecology 70:966-979)

Communityanalysis

12

Forallgroups,theMantelcorrelaFonbetweenspeciessimilarityandthelogarithmicgeographicdistancewashigherthanthatbetweenspeciessimilarityandenvironment.

Communityanalysis:Results

MantelandparFalManteltests

Dispersalscalecanbeinferred.


Definition: similarity is half its initial value

13

Halving distance as surrogates of dispersal scale

Soininenetal.2007Ecography

Logarithmicdecaymodel,1-S=cln(d)

whereSissimilarityatdistanced,cistherateofdistancedecay,assumingS=1whend=0;

thecorrespondinghalvingdistance,atwhichthesimilarityishalfitsiniFalvalue,is

dH=e0.5/c

www.azti.es 22/10/16 14

Logarithmic decay 1-S = cln(d)

Exponential decay S=S0e-cd

Slope (c) Halving distance (km) dHD=e(0.5/c)

Slope (c) S0 Halving distance (km)

dHD=-(ln(0.5))/c Zooplankton 0.0687 1444.3 9.520e-05 0.54 7280.9 Phytoplankton 0.0744 826.1 1.711e-04 0.65 4051.1 Macrozoobenthos 0.1111 90.1 4.334e-04 0.25 1599.3 Dispersal larvae macro. 0.1084 100.7 4.325e-04 0.25 1602.6 Non-dispersal larvae macro. 0.1202 64.1 5.150e-04 0.06 1345.9 1

HalvingdistancesfromlogarithmicandexponenFaldecaymodels

250 1750 3250 4750 6250

0.0

0.2

0.4

0.6

0.8

1.0

Geographic distance surrounding land (km)

Com

mun

ity s

imila

rity

(Bet

a-si

m)

250 1750 3750 5750 7750

0.0

0.2

0.4

0.6

0.8

1.0


Com

mun

ity s

imila

rity

(Bet

a-si

m)

250 1750 3250 4750 6250

0.0

0.2

0.4

0.6

0.8

1.0


Com

mun

ity s

imila

rity

(Bet

a-si

m)

250 1750 3250 4750 6250

0.0

0.2

0.4

0.6

0.8

1.0


Com

mun

ity s

imila

rity

(Bet

a-si

m)

NDLmacrozoobenthic(64km)<DLmacrozoobenthic(101km)<Phytoplank.(826km)<Zooplank.(1444km)

DispersalScale


Network graphs

15

Hierarchical clustering based on the βsim index for (a) phytoplanktonic, (b) zooplanktonic, and (c) macrozoobenthic communities

(colours of stations indicate different cluster groups). Size of stations indicates the number of connections (i.e., the similarity

between sites). Width of connections indicates the degree of similarity (thicker or thinner for more or less similar, respectively).

Connections with similarities lower than 0.6 were removed. Previous to the analysis, some stations were aggregated according

to their proximity for clarity.

1-20 -10 0 10 20 30

3040

5060

0510152025

IzmirMarmaraSinop

TriesteBasque

DanishBaltic

Bergen

l4

Norwegian

Prime

IzmirMarmaraSinop

TriesteBasque

DanishBaltic

Bergen

l4

Norwegian

Prime

-10 -5 0 5 10 15 20

3540

4550

5560

65

5

10

15

20

BermeoCOE2VIE3

SAE4

L4

Trieste

Mallorca

CPR1CPR2

CPR3CPR4

CPR5

CPR6

Naples

Saronikos

DanishBMPK2

Norway

BermeoCOE2VIE3

SAE4

L4

Trieste

Mallorca

CPR1CPR2

CPR3CPR4

CPR5

CPR6

Naples

Saronikos

DanishBMPK2

Norway

-10 0 10 20 30 40 50

4050

6070

80

0

510

1520

1

2

3

4

5

6

7

8 9 1011

12

13

14

1

2

3

4

5

6

7

8 9 1011

12

13

14(a) (b) (c)

SST(ºC)


Datasets(PopulaFongeneFcs):•  Meta-analysesusingdatafrompapersinrelaFontoDispersalesFmatesandIBDstaFsFcs(IBDslope:Fst/km,seeKinlanandGaines2003)

•  CollaboraFonswithUniversityofNordland,…Meta-analysis:•  Macroinvertebratesandplanktondatafrom148papers•  36studiesforzooplankton(addressing22species,6withIBDslopesig.)•  8studiesforphytoplankton(13speciesanalysed,3withIBDslopesig.)•  104studiesforbenthicmacroinvertebrates(addressing77species,65withIBDslopesig.)

•  397pairwiseobservaFons(FSTvsgeographicdistance)forphytoplankton,•  1952pairwiseobservaFonsforzooplankton,•  1423pairwiseobservaFonsformacrobenthos

16

Genetic Meta-analysis

www.azti.es 22/10/16 17

Coordinates,Geneinfo IBDslopesignificant Dispersalscale

Casteleynetal.2010PNAS

dispersaldistance=0.0016(IBDslope)-10001

IBDslopeNonsignificant

Speciesremovedfromtheanalysisof“species

byspecies”

Palumbietal.2003


www.azti.es 22/10/16 18

Ecological Archives E084-045-A2 Brian P. Kinlan and Steven D. Gaines. 2003. Propagule dispersal in marine and terrestrial environments: a community perspective. Ecology 84:2007–2020.Appendix B. Estimates of dispersal scale for marine taxa.†

Acanthaster planci I Echinodermata:AsteroideaAllozyme 0.00000805 0.00000805 199 Nishida and Lucas 1988Acanthaster planci (non-outbreaking)I Echinodermata:AsteroideaAllozyme 0.0000119 0.0000119 135 Benzie 1994Acanthochromis polyacanthus F Perciformes:PomacentridaeAllozyme 0.000506 0.000506 3.2 Doherty et al. 1994Acanthurus triostegus F Perciformes:AcanthuridaeAllozyme 0.0000475 0.0000475 33.7 Planes 1993Acanthurus triostegus F Perciformes:AcanthuridaeAllozyme 0.0000376 0.0000376 42.5 Planes et al. 1994Acropora cuneata I Cnidaria Allozyme 0.030 | | 0.03 0.053 Ayre and Hughes 2000Acropora cytherea I Cnidaria Allozyme 0.016 d 0.016 0.1 Ayre and Hughes 2000Acropora hyacinthus I Cnidaria Allozyme 0.014 | | 0.014 0.114 Ayre and Hughes 2000Acropora palifera I Cnidaria Allozyme 0.018 | | 0.018 0.089 Ayre and Hughes 2000Adalaria proxima I Mollusca:OpisthobranchiaAllozyme 0.024 0.024 0.067 Todd et al. 1998Alaria marginata M Phaeophyta AFLP 0.000381 0.000381 4.2 Kusumo and Druehl 2000

Genetic Estimates

Species Type‡ Taxonomy Method Dispersal scale (km) ReferenceIBD slope§ (Fst/km) IBD slope§ (Fst/km)

…

Coordinates,Geneinfo IBDslopesignificant Dispersalscale


19

GeneFcMeta-analysis

* * *NDLmacrozoobenthic(0.31km)<DLmacrozoobenthic(1.92km)<Phytoplank.(19.5km)<Zooplank.(88.9km)

Biologicalgroup(p<0.001) Markertype(p=0.009)

InteracFon(p=0.012)

DispersalScale

www.azti.es 22/10/16 20

Comparison genetic population and community levels

DispersalScale

•  Dispersaldistancesrankedthebiologicalgroupsinthesameorderatbothgene$candcommunitylevels,aspredictedbyorganismdispersalabilityandseascapeconnecFvity.

•  Thisisassociatedwithconstraintstothemovementofmacrozoobenthoswithintheseabedcomparedwiththepelagichabitat.

Macrozoobenthoswithoutdispersinglarvae

Macrozoobenthoswithdispersinglarvae

Phyto-andzooplankton< <

Conclusions

•  DispersallimitaFonsimilarlydeterminestheconnecFvitydegreeofcommuniFesandpopulaFons,supporFngthepredicFonsofneutraltheoriesinmarinebiodiversitypaCerns.

•  PracFcalconsequenceforbiodiversityconservaFon(~MSFD):populaFongeneFcsdatafromonlyafewspeciesmayhelptopredictcommunityconnecFvitypaCerns,andconversely,β–diversitymayprovideusefulinformaFontoinfersingle-speciesconnecFvity.

www.devotes-project.eu

Grantagreementnº308392

ThisprojecthasreceivedfundingfromtheEuropeanUnion’sSeventhFrameworkProgrammeforresearch,technologicaldevelopmentanddemonstra:onunderthegrantagreementnº308392

DEVOTESFinalconference,17-19Oct2016,Brussels

MODELS – PLOTTING THE COURSE IN UNCHARTED WATERS · MODELS – PLOTTING THE COURSE IN UNCHARTED WATERS Ecological models to support beCer management and policy implementaon Chair:

Documents