Establishing expectations for pan-European diatom based ecological status assessments

(1)

RESEARCH OUTPUTS / RÉSULTATS DE RECHERCHE

Author(s) - Auteur(s) :

Publication date - Date de publication :

Permanent link - Permalien :

Rights / License - Licence de droit d’auteur :

Bibliothèque Universitaire Moretus Plantin

Institutional Repository - Research Portal

Dépôt Institutionnel - Portail de la Recherche

researchportal.unamur.be

University of Namur

Establishing expectations for pan-European diatom based ecological status

assessments

Kelly, M.G.; Gómez-Rodríguez, C.; Pardo, I.; Kahlert, M.; Almeida, S.F.P.; Bennett, C.; Bottin,

M.; Delmas, F.; Rosebery, J.; Descy, J.-P.; Dörflinger, G.; Kennedy, B.; Marvan, P.;

Opatrilova, L.; Pfister, P.; Schneider, S.; Vilbaste, S.

Published in:

Ecological Indicators

DOI:

10.1016/j.ecolind.2012.02.020

Publication date:

2012

Document Version

Early version, also known as pre-print

Link to publication

Citation for pulished version (HARVARD):

Kelly, MG, Gómez-Rodríguez, C, Pardo, I, Kahlert, M, Almeida, SFP, Bennett, C, Bottin, M, Delmas, F,

Rosebery, J, Descy, J-P, Dörflinger, G, Kennedy, B, Marvan, P, Opatrilova, L, Pfister, P, Schneider, S &

Vilbaste, S 2012, 'Establishing expectations for pan-European diatom based ecological status assessments',

Ecological Indicators, vol. 20, pp. 177-186. https://doi.org/10.1016/j.ecolind.2012.02.020

General rights

Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain

• You may freely distribute the URL identifying the publication in the public portal ?

Take down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

(2)

EcologicalIndicators20(2012)177–186

ContentslistsavailableatSciVerseScienceDirect

Ecological

Indicators

j ou rn a l h o m e pa 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

Establishing

expectations

for

pan-European

diatom

based

ecological

status

assessments

Martyn

G. Kelly

a,∗

_,

_Carola

_{Gómez-Rodríguez}

b

_,

_Maria

_Kahlert

c

_,

_Salomé

_F.P.

_Almeida

d

_,

_Cathy

_Bennett

e

_,

Marius

Bottin

f

_,

_Franc¸

_ois

_Delmas

f

_,

_Jean-Pierre

_Descy

g

_,

_Gerald

_Dörﬂinger

h

_,

_Bryan

_Kennedy

i

_,

_Petr

_Marvan

j

_,

Libuse

Opatrilova

k

_,

_Isabel

_Pardo

b

_,

_Peter

_Pﬁster

l

_,

_Juliette

_Rosebery

f

_,

_Susanne

_Schneider

m

_,

_Sirje

_Vilbaste

n

a_Bowburn_Consultancy,₁₁_Monteigne_Drive,_Bowburn,_Durham_DH6_5QB,_UK

b_Departamento_de_Ecología_y_Biología_Animal,_Facultad_de_Biología,_Campus_{Lagoas-Marcosende,}₃₆₂₀₀_Vigo,_Spain

c_Department_of_Aquatic_Sciences_and_Assessment,_Swedish_University_of_Agricultural_Sciences,_PO_Box_7050,_SE₇₅₀₀₇_Uppsala,_Sweden

d_{GeoBioSciences,}_{GeoTechnologies}_and_{GeoEngineering}_(GeoBioTec)_Research_Unit,_Department_of_Biology,_University_of_Aveiro,_3810-193_Aveiro,_Portugal e_Scottish_Environment_Protection_Agency,_Inverdee_House,_Baxter_Street,_Torry,_Aberdeen_AB11_9QA,_UK

f_REQUE_Research_Unit,_Cemagref_–_Bordeaux_Centre,_50,_Avenue_de_Verdun,_Gazinet,_France

g_Laboratoire_d’Ecologie_des_Eaux_Douces,_URBO,_Département_de_Biologie,_Université_de_Namur,_Rue_de_Bruxelles_61,_B-5000_Namur,_Belgium h_Division_of_Hydrometry,_Water_Development_Department,_Cyprus

i_EPA,_John_Moore_Road,_Castlebar,_Co._Mayo,_Ireland j_Limni_s.r.o.,_Kalvodova_13,₆₁₆₀₀_Brno,_Czech_Republic

k_Vyzkumny_ustav_{vodohospodarsky}_T.G._Masaryka,_{v.v.i.Podbabska}_30,₁₆₀₀₀_Praha_6,_Czech_Republic l_ARGE_Limnologie,_Hunoldstr._14,_A-6020_Innsbruck,_Austria

m_Norwegian_Institute_of_Water_Research,_{Gaustadalleen}_21,₀₃₄₉_Oslo,_Norway

n_Estonian_University_of_Life_Sciences_(former_Estonian_Agricultural_University),_Riia_181,₅₁₀₁₄_Tartu,_Estonia

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received3November2011

Receivedinrevisedform31January2012 Accepted7February2012

Keywords:

Waterframeworkdirective Diatoms

Phytobenthos Rivers Monitoring

a

b

s

t

r

a

c

t

TheEuropeanUnion(EU)WaterFrameworkDirectivedepends,foreffectiveimplementation,on Mem-berStates(MSs)agreeingtoaconceptoftheunimpacted“reference”state,whichwillthenprovide the“expected”valueinEcologicalQualityRatio(EQR)calculations.Referenceassemblagesoforganism groupswill,inturn,vary,duetogeological,hydrological,climatic,physicochemicalandbiological fac-tors.MemberStatestacklethisbyestablishing“types”whichsharecommoncharacteristics.However, forthepurposesofensuringconsistentapplication,broadtransboundarytypeswerealsoestablished withinfiveGeographicalIntercalibrationGroups(GIGs,referredtohereas“regions”)aspartoftheEU’s intercalibrationexercise.Inthispaper,weevaluatethesetypesusingriverdiatomassemblagesandalso providereferencethresholdvaluesfortwocommonmetricsusedinpan-Europeandiatomassessments. Adatabasewasassembled,representing14EUMemberStatesfromIrelandandPortugalintheWest,to EstoniaandCyprusintheEast,inordertoexplorebiogeographicalpatternsinassemblagesunaffected byanthropogenicpressures.Multivariateanalyseswereusedtoexaminethispatternanditsrelationship withgeographic,typologicalandabioticparameters.Aftertaxonomicharmonisation,NMDSordination ofsamplesindicatedweakdifferencesinassemblagecompositionamongregions.ANOSIManalyses,in turn,indicatedthatMSwasthebestfactortogroupsimilarsampleswhereasalkalinity,recognisedas theprimaryenvironmentalvariablestructuringdiatomcommunities,althoughsignificant,explained lessvariabilityinthedataset.This,webelieve,reflectstheimportanceofmethodologicalfactorsother thantaxonomy(e.g.countingprotocol,sampleseason)thatmaybeconstantwithinaMSbutwhich varybetweenMSs.Whentwodiatommetrics,theTIandIPS,wereappliedtothedata,differencesin thedistributionofthemetricscoresbetweenMSweregenerallynotstatisticallysignificanteventhough somedifferencesbetweenregionswereapparent.AtrendofincreasingvaluesofTI(decreasingvaluesof IPS)wasobservedinthesequence:Nordic<Alpine<Mediterranean<Central-Baltic<Eastern Continen-talregions.Additionally,somedifferenceswereobservedamongtypeswithintheMediterraneanand Nordicregions,thoughnotforotherregions.Thedatausedinthisexerciseprovidesuswitharegionand, insomecases,atypespecificbenchmarkdatasetagainstwhichnationalreferencedatacanbecompared.

∗ Correspondingauthor.

(3)

1. Introduction

TheprincipleunderlyingtheWaterFrameworkDirective(WFD: EuropeanUnion,2000)isthatallwaterbodies,regardlessoftypeor location,throughouttheEUshouldbemanagedinaconsistentand transparentmanner,withcommonunderstandingofhowmuch deviationfromthenaturalstatewaspermissiblebefore preventa-tiveorrestorativeactionwasneeded.

Althoughmemberstates(MSs)wereeachexpectedtodevelop nationalmethodstoevaluate“ecologicalstatus”,theWFDrequires that methods are expressed as Ecological Quality Ratio (EQR), deﬁnedasthe“observed”(O)valueofacandidatemetric,divided bythatvalue“expected”(E)atreferencecondition(O/E). Conse-quently,itisimportantthatallMSshaveacommonunderstanding ofthebiologicalstateat“reference condition”,asitisfromthis thattheexpectedvalueisderived.Inpractice,evenifallother vari-ableswereheldconstant,theexpectedvaluewillvarydepending uponthetypeofwaterbodyunderconsideration.Forthisreason, theEUestablishedﬁveGeographicalIntercalibrationGroups(GIGs; referredtointhispaperas“regions”)and,withineachregion, differ-entiatedanumberofwaterbodytypes(“types”)inordertoensure like-with-likecomparisonsofbiologicalcommunities.

TheWFDspecifiestheuseofabioticfactorssuchasaltitude, catchmentsizeandgeologytodefinetypeswithinMSandthese same principles were applied tothe development of intercali-brationtypes(ECOSTAT,2004).Thesetypesweredesignedtobe commontypesfor allbiological quality elements(BQEs) which couldbeeasilyappliedbyallMSs.Consequently, thetypesare notoptimisedforanyparticularBQEandfirstattemptsat inter-calibrationindicatedsignificantdifferencesinvaluesofEamong MSswithinaregion.Whilstthismayreflectgenuinedifferences inthenaturaldiatomassemblages,thepossibilitythatdifferent approachesand criteria had beenusedtoselect referencesites couldnotberuledout(Kellyetal.,2009)

Inpractice,thisissuedividesintotwocomponents:the neces-sity for an underlying concept of what “reference condition” actuallyrepresentsandappropriatescreeningstrategiestoselect samplesthatmatchthisconcept.Theconceptispartdetermined bytheWFD,whichrequiresMStocharacterisehydromorphological andphysicochemicalconditionsforwaterbodytypesthatsupport “highecologicalstatus”(HES),andpartbysubsequent documenta-tionfromECOSTAT(e.g.Wallinetal.,2005).Thisworkrecognised that“referencecondition”wasnotnecessarilythe“pristine”state whichis,acrossmuchofEurope,impossibletoﬁndduetohigh pop-ulationdensitiesandalonghistoryofanthropogenicalterationof landscapes.Thereis,inevitably,atensionbetweenecologicaland statisticalfactorsinderivingaviablereferenceconceptand“best attainablecondition”(sensuStoddardetal.,2006)willvaryfrom regiontoregion.Themiddlewayistopursue“minimallydisturbed conditions”,whichattemptstoderive,throughaniterativeprocess, explicitcriteriawhichestablishtheleastamountofambienthuman disturbanceinaregion(Stoddardetal.,2006).

Thecriteriaforselectingsitesunderreferenceconditionshould bebased initially and primarilyon non-biological measuresto avoidcircularity(Baileyetal.,2004;Stoddardetal.,2006).Primary screeninguseddataonlanduseinthecatchment,supplementedby valuesofchemicalvariablesassociatedwithpressures.However, fewsitessurvivedthisintensivescreeningprocessinMSdatasets, leadingtoarisk that,forsomecountries, Ewouldbebasedon limiteddata.AsErepresentsthevalueofbiologicalquality ele-mentsatreferenceconditions,thereferenceconceptneedstoyield sufﬁcientdatabeforerobustinferencestobedrawn.Thiscanbe overcome,tosomeextent,bypoolingdatafromseveralMStoform abenchmarkdataset.Analternativewouldbetouseinformation fromsitesthatfailthefullscreeningbutwhichmaystillbe suit-ableasreferencesitesforcertaincombinationsofbiologicalquality

elements(BQEs)andpressures(theseare“partial”referencesites” sensuECOSTAT,2010).Forexample,hydromorphologicalchanges whichmayimpactthebenthicinvertebrateormacrophyte assem-blagesareunlikelytoaffectmetricsusedtoassessphytobenthos responsetonutrients.Siteswheretherearehydromorphological changesmight,therefore,qualifyas“partial”referencesites.The ideaexploredinthisstudyisthatthesuitabilityofsuchsitescanbe evaluatedbycomparisonwithabenchmarkdatasetof“true” refer-encesites.Thisisapragmaticsolutionbalancingthestatisticaland conceptualneedsofareferencesitenetworkwithinaMS.

One of theBiological QualityElements (BQEs) that MSsare requiredtoassessis“MacrophytesandPhytobenthos”andmany MSsusediatomsasproxiesforphytobenthos.Europe-wide com-parisonsareaidedasmethodsforsamplingandanalysisofdiatoms (CEN, 2003, 2004) are consistent and yield datasets that are amenabletopan-Europeancomparisons.Theﬁrstintercalibration exercise(Kellyetal.,2009)indicatedvariabilityinreference val-ues,withconsequencesforthecalculationofEQRs.However,due tothevariablequalityofreferencescreeninginthisearlierphase, coupledwiththesmallsizeofdatasetsandthelackofaharmonised approachtotaxonomy,itwasunclearhowmuchofthisdifference representedtruedifferencesindiatomassemblagesatthe refer-encestate,orhowmuchwasmethodological.

Theaimofthispaperistoexaminetheissueofreference condi-tionsforphytobenthos,payingparticularattentiontotheextentto whichdiatomassemblagecompositionandthemetricsusedto cal-culatethephytobenthosIntercalibrationCommonMetric(pICM:

Kellyetal.,2009)varywithinandbetweenregions.Inparticular, wewillevaluatewhethertheintercalibrationtypologyisadequate forphytobenthosataEuropeanscale,whetheranalternative typol-ogyneedstobederivedorwhethercomparisonsshouldbemade onlyattheregionlevel.Weuseadatasetof“true”reference sam-plesandtheoutputswillthenprovidebiologicalcriteriaagainst which“partial”referencesamplesmaybeassessed.Anexample of the latter case would be a sample from an otherwise pris-tineriverthatisdownstreamofaphysicalbarrier,e.g.adam.As manyMSshave onlyalimited numberof referencesites,there arebeneﬁts(e.g.morestatisticallyrobustsamplesizes)inbeing abletousesuchsitestocomputeEfornationalassessment sys-tems.

2. Methods

2.1. Datasetandscreeningprotocols

ThedatausedinthispaperwascollectedbyEUMemberStates, eitheraspartofresearchstudiestodevelopmonitoringtoolsfor theWFD,oraspartofroutinemonitoringprogrammes.Diatom samplecollectionandanalysisfollowedCENguidance(CEN,2003, 2004).Biologicaldata(diatomassemblagecomposition,asrelative abundance–RA)andgeographiclocationandenvironmentaldata werestoredinAccess2000databases.MSsassignedeachsiteto theappropriateregion–Alpine(ALP),Central-Baltic(CB), Eastern-Continental(EC),Mediterranean(MED)orNordic(N)and,within eachregion,totheappropriatetype(Table1)althoughthiswas notalwaysstraightforwardastheintercalibrationtypologyrarely alignedwithnationaltypologies.

Allsitesweresubmittedtoascreeningprocessbasedon thresh-oldsagreedinCB-GIGforartiﬁcialandintensiveagriculturalland useaswellaswaterchemistryparameters,i.e.BOD5,O2,NH4,NO3,

NO2,PO4(Bennettetal.,2011).Rejectionthresholdswereapplied

forlanduseandtheleaststringentwater-chemistrythresholdfor CB-GIGtypeswasusedforscreeningpurposes.Allparameterswere evaluatedsimultaneouslyandallsamplesthatdidnotcomplywith, atleast,oneofthethresholdswereremoved.1096samplesfrom

(4)

M.G.Kellyetal./EcologicalIndicators20(2012)177–186 179

Table1

Briefsummaryoftheintercalibrationtypology,describedinmoredetailinECOSTAT (2004).

Type Rivercharacteristics

Nordicregion

R-N1 Smalllowlandcatchments,siliceousgeology, moderatealkalinity,clear

R-N2 Small-mediumlowlandcatchments,siliceous geology,lowalkalinity,clear

R-N3 Small-mediumlowlandcatchments,siliceous geology,lowalkalinity,humic

R-N4 Mediumlowlandcatchments,siliceous

geology,moderatealkalinity,clear R-N5 Smallmid-altitudecatchments,siliceous

geology,lowalkalinity,clear

R-N7 Smallhighlandcatchments,siliceousgeology, lowalkalinity,clear

R-N9 Small-mediummid-altitudecatchments,

siliceousgeology,lowalkalinity,humic Central-Balticregion

R-C1 Smalllowlandcatchments,siliceousgeology, sandysubstratum

R-C2 Smalllowlandcatchmentssiliceousgeology, rockysubstratum

R-C3 Smallmid-altitudecatchments,siliceous geology

R-C4 Mediumlowlandcatchments,mixedgeology

R-C5 Largelowlandcatchments,mixedgeology

R-C6 Small,lowlandcatchments,calcareousgeology Alpineregion

R-A1 Pre-alpine;smalltomedium–sizedhigh altitudecatchments,calcareousgeology

R-A2 Alpine;smalltomediumhighaltitude

catchments,siliceousgeology Mediterraneanregion

R-M1 Small,midaltitudecatchments,highly seasonal

R-M2 Medium,lowlandcatchments,highlyseasonal R-M3 Large,lowlandcatchments,highlyseasonal

R-M4 Small/mediumhighaltitudecatchments,

seasonal,highsedimenttransport

R-M5 Smallcatchments,temporary

EasternContinentalregion

R-E1 Carpathians:smalltomedium,mid-altitude catchments

R-E2 Plains:medium-sized,lowlandcatchments R-E3 Plains:largeandverylarge,lowland

catchments

R-E4 Plains:medium-sized,mid-altitude

catchments

R-E5 Balkans:medium-sized,mid-altitude

catchments

R-E6 DanubeRiver:middleanddownstream

stretches

Fig.1.Mapshowinglocationofsamplesites.

533referencesitesin14countriesdistributedacrossallﬁveregions passedallofthesecriteria(Fig.1;Table2).

In order toensure that thedataset wasnot biased towards particularregionsorstreamtypes,alimitof20samplesperMS pertypewasestablished;wherethislimitwasexceeded,20 sam-pleswereselectedatrandombutwithanoverridetoensurethat thissamplingmaximisedthenumberofsitesselected(i.e. multi-plesamplesfromthesamesitewereonlyallowedIFthiswasnot attheexpenseofahithertounrepresentedsite).Finally,alltaxa identiﬁedonlytogenuswereremoved,alongwithtaxawitha pre-dominatelyplanktonichabit,inordertoremovetheinﬂuenceof upstreamimpoundments.Ifthesecriteriaresultedintheremovalof ≥20%ofthetotalvalves,thesamplewasexcludedfromtheanalysis. Thesescreeningstagesreducedthenumberofreferencesamples availableforanalysesto409samples.

2.2. Diatomtaxonomyandharmonisation

Taxonomicharmonizationtookplace inseveralstages. First, synonymsweremerged,afterwhichweappliedtheconventionsof

Kahlertetal.(inpress)andKellyandEctor(inpress)andremoved anytaxawithamaximumRA≤2%andwithlessthan10records. However,evenafterthistherewerestillsometaxawhose differen-tialtreatmentbyMSmayhaveasigniﬁcanteffectontheoutcomes ofanordinationandoverridethetrueecologicalsignal.

Table2

Summaryofthephytobenthosreferencedatabasebyregion.ValuesshownumberofsitesthatfulﬁlallscreeningcriteriaforeachMemberState,withthenumberofsamples inbrackets.

MS ALP CB EC MED N Totalsites Totalsamples

AT 2(2) 2(2) 4 4 BERW 5(15) 5 15 CY 8(13) 8 13 EE 1(1) 1 1 ES 48(58) 48 58 FI 55(55) 55 55 FR 98(189) 5(5) 57(118) 160 312 IE 1(1) 15(23) 16 24 NO 10(10) 10 10 PL 30(30) 30 30 PT 34(36) 34 36 RO 11(20) 11 20 SE 19(19) 46(46) 65 65 SK 47(169) 48 169 UK 40(199) 20(85) 60 284 533 1096

(5)

In order to solve this problem, a second stage of taxa har-monization used Canonical Correspondence Analysis (CCA, Ter Braak1986,1994,1995;Palmer1993;softwarePASTversion1.65, Hammer &Harper March 2007,Hammer et al., 2001)to anal-ysethe dataset.Each MS wasentered as a dummy categorical variable. CCA outcomes were examined for taxa that resulted in the formation of distinct clusters of samples from particu-lar MS at the margins of the data “cloud”. These taxa were thenstudiedmore closelytoseeifthere maybereasons asso-ciatedwithidentificationconventionsfortheirseparation;ifso, apragmaticsolutionwasappliedtothedatasetandtheprocess wasrepeateduntilnoobviousnomenclaturalproblemsappeared after CCA. As a result of this, Diadesmis gallica and D. pusilla were merged,along with Encyonopsis cesatii var. cesatii and E. cesatiivar. geitleri,Navicula symmetricaand N.schroeteri, Tabel-lariaflocculosaandT.fenestrata.Somepracticalidentificationissues remainedbutwouldhaverequiredare-examinationofthe origi-nalmicroscopeslideswhichwasbeyondthescopeofthisparticular project.

2.3. Ordinationandclassiﬁcation

Diatomassemblage datawereconverted topercentages and then square-root transformed. Non-metric Multi-Dimensional Scaling(NMDS) using Bray-Curtis dissimilarityindex was used to explore the structure of the data (command: metaMDS in Vegan). Analyses of similarity (“ANOSIM”: Clarke, 1993) were used toevaluate whether there were signiﬁcantdifferences in assemblage composition among types, Member States (MSs), regionsoralkalinityclassesand,therefore,whetherthesefactors could be used to group similar samples. Alkalinity measure-mentsweredividedintoﬁveapproximatelyequallysizedclasses: very low (<0.1meq/l CaCO3), low (0.1–0.2meq/l), intermediate

(0.2–0.4meq/l),high(0.4–1.0meq/l)and veryhigh(>1.0meq/l). One ANOSIM analysis was conducted for each factor of inter-est. Additionally, independent analyses were carried out for each region, type, MSs or alkalinity classes as grouping factor. Bonferroni correction, to reduce the risk of wrongly rejecting thenull hypothesis, yielded a critical probability of 0.0036 for ˛=0.05 (14ANOSIM tests). “Type”was onlyused asa variable if it was represented in >1 MS. NMDS and ANOSIM analyses were performed using the Rsoftware package (R Project Core DevelopmentTeam,2005)withtheVeganpackage(Oksanenetal., 2007).

2.4. Calculationandanalysisofmetrics

Apreliminaryintercalibrationhasalreadybeenundertakenin Central-Balticregion,thelargestoftheregions, spanningmuch oflowlandEurope.Anintercalibrationmetric,pICM,was devel-oped,basedontwowell-establishedwaterqualitymetrics,Indice dePolluosensibilitéSpeciﬁque(IPS:CosteinCEMAGREF,1982)and Trophieindex(TI:Rottetal.,1999).ThepICMwasthenregressed against national metrics and national EQRs for High/Good and Good/Moderatestatusboundariesconvertedtoequivalentvalues ofpICM.This,then,providedabasisforcomparingthepositionof nationalboundaries(Kellyetal.,2009).

ThecomponentmetricsofthepICM(IPSand TI)were calcu-latedusing queries withintheAccess databaseonall screened samples(theeffectoftaxonomicharmonisationontheoutcomesis minor:seeKellyandEctor,inpress).Differencesinmedianvalues ofIPSandTIbetweenregionsandbetweentypeswithinregions wereexploredusingthenon-parametricKruskal–Wallistestand theMann–Whitneytest.Notallsitescouldbereliablyassignedto

typesbyMS,sotheseanalyseswereperformedonasubsetof319 samples.

2.5. Applicationofresultstointercalibration

Thisexerciseprovidesaframeworkforstatisticalcomparisons ofreferenceconceptsacrossEurope,eveniffullscreening infor-mationisnotavailable.Threescenariosforhowthesebenchmark datasetsmaybeusedtovalidatereferenceandHESdataare exam-ined.Ineachcase,weuseddatasuppliedbyMSsto:

• testthevalidityof“nationalreference”(NR)samplesprovidedby aMSwhichcontributeddatatothebenchmarkdataset; • testthevalidityofreferencedataprovidedbyaMSwhichdidnot

contributedatatothebenchmarkexercise;and,testthevalidity ofhighstatusboundaryprovidedbyaMSwhichdoesnothave referencesites.

3. Results

3.1. Compositionofdiatomassemblagesfromreferencesamples 159harmonisedtaxa wereidentified fromthe409reference samples.Overall,Achnanthidiumminutissimumwasthemost con-stantand,often,mostnumericallyabundanttaxon,foundin98% of samplesacrossall five regions, witha maximum RAof 94% (Table 3). Thiswas followed by Fragilariacapucina (60%, 62.5% respectively)andEncynonemaventricosum(59%,89%respectively). Overall,thedifferencesbetweenregionswererelativelysmall:the mostdistinctoftheregionswasALP,whichwascharacterisedby theabsenceorinfrequentoccurrenceofacid-sensitivetaxa(e.g. Achnanthesoblongella,Eunotiaspp.)reflectingthepredominately calcareousnatureoftheseuplandareas.

Achnanthidiumminutissimumwas alsothetaxonmost likely to predominate in samples, forming the maximum RA of any taxon in 50% of all samples, followed by Cocconeis placentula (8%),Tabellariaﬂocculosa(6%)andAchnanthidiumpyrenaicum(5%). However,thereweresomeintra-regional differences:inALP, A. pyrenaicumdominatedmorefrequently(37%)thanA.minutisimum (32%)whilst,in EC, Cocconeisplacentula wasthemostcommon dominantspecies(23%),followedbyA. minutissimum(15%)and Gomphonemapumilum(11%).

3.2. Ordinationandclassiﬁcation

TheordinationtechniqueNMDS(stress=20.9%withtwo dimen-sions)showedalimitedamountofaggregationforsomeregions. ALPregionisthemoststronglyaggregated,withlow valuesfor bothaxes1and2whilstECandMEDregionshaveoverlapping distributionswithlowAxis1scores.SamplesfromNregiontend tohavehighaxis1scoreswithsamplesfromCBregiondispersed throughoutthediagram(Fig.2).ANOSIManalysesdemonstrated signiﬁcantdifferencesinassemblagecompositionamongregions, MSsandtypes(Table4).TheR-statisticintheregionalANOSIMwas lowerthanthatofeitherMSortype(Table4).TheANOSIMtesting alkalinityasagroupingfactorexplainedlessvariancethanregion, typeorMS.

3.3. Variationinmetrics

Asigniﬁcanteffectofregionwasalsoobservedwhenvaluesof theconstituentmetricsoftheICMwerecompared(Fig.3;Table5) withALPandNregionshavinghighermeanvaluesofIPSandlower valuesforTIthantheothers. ECregion,in particular,had some sampleswithlowvaluesofIPS/highvaluesofTI.

(6)

Table3

Thethirtytaxamostcommonlyencounteredinallreferencesamplesinthedatasetalongwiththeirmaximumrelativeabundance(“Max”)andtheproportionofsamplesin whichthetaxawerefound(“constancy”).Inmostcases,taxanamesrefertospeciesasunderstoodbyKrammerandLange-Bertalot(1986,1997,2000,2004)andassociated varieties;exceptionsarelistedinKellyandEctor,inpressorSection2.

Taxon All ALP CB EC MED N

Max Constancy Max Constancy Max Constancy Max Constancy Max Constancy Max Constancy Achnanthidiumminutissimum (Kütz.)Czarnecki 95.5 69.0 88.5 38.7 95.5 90.7 93.7 31.9 93.6 52.8 92.4 86.8 Achnanthidiumpyrenaicum (Hustedt)Kobayasi 95.7 21.5 77.5 36.8 92.5 20.2 95.7 17.0 71.8 33.5 30.3 10.0

AchnanthesoblongellaOestrup 51.4 18.0 51.4 31.8 11.9 3.7 41.7 28.3

Amphorapediculus(Kützing) Grunow

60.5 21.9 23.0 18.9 22.3 33.6 21.3 21.5 60.5 24.8 1.3 4.1

Brachysiravitrea(Grunow) RossinHartley

31.5 14.4 8.6 14.0 0.8 2.5 31.5 39.7

CocconeisplacentulaEhrenb. 79.5 41.3 71.2 27.4 78.9 63.6 79.4 28.9 79.5 44.7 47.1 20.5

Eunotiabilunaris(Ehrenb.)F.W. Mills

32.7 16.0 32.7 17.8 1.2 0.7 1.2 5.0 19.8 38.8

EunotiaimplicataNörpel. Lange-Bertalot&Alles

36.9 15.5 10.4 15.0 36.9 44.7

Eunotiaminor(Kützing) GrunowinVanHeurck

15.8 14.6 15.8 16.2 0.3 0.7 8.9 8.1 13.9 32.9

EncyonemaneogracileKrammer 14.4 15.8 2.0 2.8 14.4 17.4 10.6 5.6 4.8 37.0

Encyonemaventricosum(Ag.) Grunow

89.2 41.7 89.2 34.0 36.2 56.1 29.3 20.7 9.7 43.8

Eolimnaminima(Grunow) Lange-Bertalot

61.6 17.6 3.8 3.8 18.6 29.3 14.8 9.6 53.7 19.9 61.6 10.5

Fragilariacapucina Desmazieres

62.5 48.6 3.3 16.0 58.8 70.1 6.1 10.4 49.4 27.3 62.5 72.1

Fragilariavaucheriae(Kutz.) J.B.Petersen

25.0 35.6 9.4 10.4 23.4 51.4 6.5 12.6 20.7 21.7 25.0 48.9

Gomphonemaolivaceum (Hornemann)P.Dawson

27.0 14.9 9.3 26.4 19.3 19.3 18.7 16.3 4.7 6.8 27.0 7.8

Gomphonemaparvulum(Kutz.) Kutz.

61.4 27.5 0.8 3.8 61.4 41.4 6.4 5.9 14.2 16.8 26.9 39.7

Gomphonemapumilum(Grun.) Reichardt&Lange-Bert

86.0 33.3 67.8 34.9 80.6 39.9 65.7 26.7 86.0 46.6 27.5 17.4

Hannaeaarcus(Ehr.)Patrick 76.3 23.4 9.6 22.6 76.3 30.8 36.4 5.2 5.1 11.8 17.9 32.4

Meridioncirculare(Greville) C.A.Agardh

22.6 15.5 1.5 8.5 5.4 21.8 10.9 15.6 13.9 5.6 22.6 16.9

NaviculacryptocephalaKützing 7.5 13.7 0.3 1.9 5.2 19.9 7.5 6.7 4.3 13.0 6.4 15.1

Naviculacryptotenella Lange-Bertalot

25.4 21.2 1.8 20.8 17.8 27.1 3.7 12.6 25.4 32.9 2.8 9.6

Nitzschiadissipata(Kützing) Grunow

15.0 22.6 6.5 19.8 11.4 36.1 15.0 16.3 8.7 16.8 3.8 12.3

NaviculagregariaDonkin 39.7 19.4 0.3 4.7 39.7 38.0 10.5 14.8 1.9 8.7 7.9 10.0

Naviculalanceolata(Agardh) Ehrenberg

64.3 17.2 1.5 2.8 59.0 34.0 64.3 20.7 0.9 2.5 16.3 8.2

Nitzschiapalea(Kutz.)W.Sm. 59.5 16.5 9.2 3.8 17.9 29.0 59.5 3.0 19.5 17.4 8.5 11.9

Naviculatripunctata (O.F.Müller)Bory

40.8 14.1 3.0 11.3 18.0 20.9 40.8 22.2 5.7 13.7 4.2 0.9

Planothidiumlanceolatum (Breb.)RoundetBukh.

42.3 16.6 2.5 11.3 9.4 21.5 42.3 21.5 10.6 14.3 3.6 10.5

Reimeriasinuata(Gregory) Kociolek&Stoermer

28.0 28.8 26.3 25.5 26.1 43.3 19.5 24.4 28.0 25.5 24.8 14.2

Synedraulna(Nitzsch)Ehr. 70.7 25.8 2.1 9.4 30.9 35.8 70.7 16.3 2.9 11.8 18.3 35.2

Tabellariaﬂocculosa(Roth) Kützing

88.2 31.2 88.2 36.1 18.2 1.5 8.8 4.3 86.2 77.2

N.taxa 192 95 186 108 145 169

N.samples 942 106 321 135 161 219

Table4

Anosimresultsforentiredatasetof409referencesamples.P-valueafterBonferronicorrectionis0.0036for˛=0.05.

Variable R Signiﬁcance Comment Numberofsamples

Region 0.338 *** ₄₀₉

MS 0.511 *** _Only_MS_with₅_or_more_samples_included._AT₍₃_samples)

andEE(1sample)excluded(toofewsamples)

405 Type 0.355 *** _RC1,_RE4,_RN4_removed_(too_few_samples)_RA2_excluded

(19samples,allinFR);RM3excluded(7samples,allinPT)

287

Alkalinity 0.285 *** _Data_missing_for_many_samples ₂₇₈

(7)

Fig.2.NMDSplotshowingtheordinationofsamples,classiﬁedbyregion,basedon speciesrelativeabundanceandBray-Curtisdissimilarity.

Therewerenosignificantdifferencesforeithermetricwithin ALP,ECandMEDregionsbutbothshowedsignificantdifferences forCBandNregions.Allregions,exceptALP,showedasignificant differencebetweenMSforTIwhilstCBandNregionsalsoshowed asignificanteffectforIPS.InthecaseofNregion,lowerIPSandTI valuesareassociatedwithhumicstreamtypesandtheremaybe acasefortreatingtheseseparately.However,moreworkwillbe neededtoensurethatthisisnotconfoundedbyaMSeffect. 3.4. Applicationoftheseresultstointercalibration

3.4.1. Scenario1:testingvalidityofNRsamplesprovidedbyaMS whichcontributeddatatothebenchmarkdataset

Thepopulationofreferencesamplesavailableforthistestwas greaterthanthatcontributedtothebenchmarkdataset,asefforts hadbeenmadewithintheMSinquestiontoexpandthepoolof referencesamplesavailable.Theseadditionalsamplesrepresent bothextrasitesandmoresamplesfromsitescontributingtothe benchmarkdataset.MS1-GR,thereforeisnotwhollyindependent ofthebenchmarkdataset;howeveritisclearfromFig.4aandb thatthetwodatasetshaveverysimilarvaluesforIPSandTI.By con-trast,MS1-NRissignificantlydifferentfromthebenchmarkdataset (p<0.001forbothIPSandTI).Hadtherebeennosignificant differ-ence,MS1-NRcouldqualifyas“partial”referencesites.However, asastatisticallysignificantdifferencewasfound,MS1nowneeds totakeextrastepstoensurethevalidityofpredictionsmadefrom

Table5

SigniﬁcanceofdifferencesinTIandIPSvaluesbetweenregionsandbetweentypes withinregions,assessedusingKruskal–Wallistest.

Betweentype BetweenMS Notes

TI IPS TI IPS Betweenregions <0.001 <0.001 <0.001 <0.001 Withinregions ALP 0.239 0.150 0.364 0.276 2MS: mostly FRdata CB 0.001 0.002 <0.001 <0.001 6MS EC 0.931 0.611 0.053 0.519 2MS MED 0.084 0.575 0.001 0.196 3MS N <0.001 <0.001 <0.001 <0.001 3MS

thesesamples(notethatsomeoftheMS1-NRsamplesrepresented the“bestavailable”sitesforstreamtypeswithno“true”reference sitesavailable).

3.4.2. Scenario2:testingvalidityofreferencedataprovidedbya MSwhichdidnotcontributedatatothebenchmarkexercise

Fig.4canddshowscleardifferencesinIPSandTI,bothofwhich arehighlysignificant(p<0.001)suggestingthatMS2ssamplesdo notfulfilthecriteria.MS2thenreconsidereditsreferencedataand adjusteditscriteria.Itcouldalsohavemadeareasonedcaseforthis discrepancy(itispossible,forexample,thattheirnationaltypes arenotrepresentedinthebenchmarkdataset).Hadtherebeenno significantdifference,therewouldhavebeen,attheveryleast, reas-surancethatMS2sEQRvalueswerecomparablewiththosefrom otherMS.

3.4.3. Scenario3:testingvalidityofhighstatusboundary providedbyaMSwhichdoesnothavereferencesites

Highecologicalstatus(HES)samplesfromMS3,whichset ref-erencevaluesbyexpertjudgement,arealsoplottedinFig.4candd alsoshowsasigniﬁcantdeviationinIPSvalues(p<0.001)forboth metrics,suggestingthatMS3sexpertjudgementofHESmayneedto bere-evaluated.Thebenchmarkdatasetprovidesaresourcewhich MS3’sexpertsmayusetocomparefuturedata.

4. Discussion

TheWFDembodiestheEU’sprincipleofsubsidarity,leavingthe detailsofimplementationtoindividualMSs.Theintercalibration exerciseexistsprimarilytoensureconsistentapplicationofthe ecologicalprinciplesoftheWFD;atthesametimethoseinvolved inintercalibrationneedtorecognisethatmultiplesolutionstothe challengesraisedbytheWFDarepossible.ByensuringthatMSs havecommonconceptsofreferenceconditionsandgood ecolog-icalstatus(“GES”),intercalibrationensuresconsistentapplication ofthereferenceconcept.

The harmonised ecological status class boundaries which emerge fromthe intercalibrationexercise are, however,legally bindingontheMS.TheGESboundary,inparticular,has signifi-cantregulatoryimplications,potentiallyleadingtohighercostsfor watertreatmentforbothindustryandconsumers.Thereare, there-fore,largefinancialconsequencesofthisworkandweneedtofind abalancebetweenanecologicallydefensibleapproachtoreference condition,andonethatispracticalforMSstoimplement. 4.1. Selectionofreferencesites

Theapproach todefining Europeanreferenceconditions has evolved in the decade since the WFD was first published. The definitionintheWFDitselfregardsthereferencestateasthe hydro-morphologicalandphysicochemicalconditionswhichsupportHES whilst,at thesame time, defining HESas thebiological condi-tionswhich“reflectthosenormallyassociatedwiththattypeunder undisturbedconditions”.Muchsubsequentworkhasinvolved try-ingtounpickthisingrainedcircularity,particularlyasthelackof referenceconditionsinmanyareasledtosubjectivedefinitionsof HESwhichweredifficulttovalidateinpractice.Theoutcomeof thishasbeenatrend,startingwithWallinetal.(2005),towards definingreferenceconditionsusingabioticcriteriaalone,andthen usingbiologicalassessmentsfromsiteswhichpassedthese crite-riatogenerate“expected”valuesofmetrics.Inthispaperwehave identifiedreferencesitesusingabioticcriteriaandthendefine bio-logicalreferencevaluesfortheintercalibrationmetricsfollowing AnnexIIIofthenewICguidance(Wallinetal.,2005;ECOSTAT, 2010).Thisdatasetcanthenbeusedasabenchmarkagainstwhich

(8)

Fig.3.VariationbetweenvaluesofIPSandTIofreferencesampleswithinandbetweentheﬁveregions.Horizontallinesshowmedianvalues;boxesshow25thand75th percentilesofthedata.(a,b)Comparisonbetweenregions;(c–f)comparisonswithinregions:(c,d)ALP;(e,f)CB;(g,h)EC;(i,j)MED;(k,l)N.

otherdatasetscanbeevaluated,evenifabioticpressuredataare notavailable.

However,a distinction needsto bedrawn betweena refer-encesiteandsamplescollectedfromthatsite.Inthis study,we notedoccasionalsamplesfromapparentlypristinesiteswhichhad diatomassemblagessuggestingimpactedconditions.However rig-orousthescreening,wecannotexcludeallshort-termincidents thatmayhave affectedthebiology andwe mustalsoallowfor thepossibilitythatentirelynaturaleventsmayleadtolocalised

enrichmentforaperiodoftime.Inotherwords,asinglesample whosecompositionsuggestsimpactis not,ofitself,groundsfor decidingthatasiteisnotinreferencecondition,solongasthesite fulﬁlstheabioticcriteria.Suchsituationsshould,however,prompt are-checkingofabioticcriteriaaswellasadditionalbiological sam-pling.Weregardtheabioticscreeningprocessandthebenchmark datasetswhichresultasthestartingpointfordebateand discus-sion,notasaﬁnalandunambiguousstatementonthebiological assemblagesassociatedwith“referencecondition.”

(9)

Fig.4.Useofthebenchmarkdatasettovalidatereferenceandhighecologicalstatusconcepts:examplesfromtheCentral-Balticregion.Benchmark;MS1-GR:subsetof samplesfromsiteswhichfulﬁlallcriteriaforreferenceconditions,MS1-NR:samplesfromaMemberStatewhichparticipatedinthereferencescreeningexercisebutwhich alsohasapopulationof“nationalreference”samples;MS2:samplesfromaMemberStatewhichdidnotparticipateinthescreeningexercise;MS3:“highecologicalstatus” samplesfromaMemberStatewithnoreferencesites.Horizontallinesrepresentthe10thpercentileofIPSand90thpercentileforTI.Seetextformoredetails.

4.2. Testingtheintercalibrationtypology

Theintercalibrationtypologywasestablishedearlyinthe inter-calibrationprocess,based onknownpredictors of streambiota (ECOSTAT,2004);however,neitherthisstudynorearlierstudies onphytobenthos(Kellyetal.,2009)indicatethatitisa particu-larlyusefulapproachfordifferentiatingbetweendifferentstream types.Inthisprojectitwasanticipatedthat acombinationof a largerdatabasethanwasavailableinKellyetal.(2009)andthat ahighstandard ofreferencescreeningwould allowaworkable typology,perhapsstraddlingexistingGIGboundaries,toemerge. Thisprovednot tobethecase: althoughsigniﬁcantdifferences betweenregionswereobservedusingbothNMDS(Table4)and diatommetrics(Table6),itwasMSratherthantypeoralkalinity thatemergedasthemostpowerfulfactorstructuringthedataset. Thelackofeffectforalkalinitywasparticularlynoteworthy,asmost MSsrecognisethisasthekeyvariablestructuringdiatom assem-blagesintheabsenceofpressurewithintheirnationalassessment systems(Tisonetal.,2005;Kellyetal.,2008).

Itispossiblethatsomeofthecompromisesnecessarytoproduce aharmoniseddatasetdecreasedthesensitivityofthe multivari-ateanalysesbutalternatively,methodologicaldifferencesbetween MSmaskedtheexpectedalkalinity-drivensignal.AlthoughallMSs adheredtostandardmethods(CEN,2003,2004),thesearebroadly writtenandthereisscopeforinterpretationtosuitlocal circum-stances.Kahlertetal.(inpress)notedthatharmonisingtaxonomy followingsimilarconventionstothoseadoptedinthispaperhad

Table6

Medianplus10th(forIPS)and90th(forTI)percentilesofmetricsvaluesrecorded fromreferencesamplesforeachoftheregions.Valuesfortheentireregionarealso given(X-GIG).

ALP CB EC MED N X-GIG

IPS Median 18.8 17.7 16.7 17.7 19.3 18.3

10th%ile 15.5 15.0 14.9 14.3 18.0 15.0

TI Median 1.32 2.00 2.10 1.55 1.14 1.52

90th%ile 1.96 2.66 2.71 2.52 1.69 2.56

littleeffectonbetween-operatorvariationinapan-European ring-test,andsuggestedotheraspectsofprocedure(e.g.treatmentof brokenvalves)maycontributesystematicerrorstotheanalytical process.Inthepresentstudy,differentsamplingpracticesmayalso haveledtodifferencesinspeciescompositionforthesamples pro-videdbyindividualMS.Suchpracticeswillvaryforgoodreasons (seeKellyetal.,1998forafullerdiscussion)but,asstreamalgae displaydistinctperiodicity(e.g.Marker,1976),adjacentMSswhich adoptdifferentsamplingregimesfor otherwise similarstreams mayhavequitedifferentdiatomassemblages.TheICexercisehas toworkwithintheconstraintsimposedbysuchlimitations.We didattempttoaccountfor“season”inouranalyses;however,crisp delimitationofseasonsisdifﬁcultina datasetspanninganarea from60◦Nto35◦Sand10◦Wto37◦E.

Althoughnocaseforatypologybasedonthediatom assem-blagewithinregionscanbemadefromthesedata,thisislikelydue tothenatureoftheintercalibrationexerciseratherthanreﬂecting thetrueconditionofphytobenthosassemblages.Akeyquestion forthepresentexercise,therefore,iswhetherornotfurtherwork wouldimproveintercalibrationofboundaries:eachMSmayhave aninternallyconsistentapproachwhichallowsspatialand tempo-ralcomparisonswithinapoliticalunit;differencesbetweenthese approachesmaythenhavetoberegardedas“noise”forthe pur-posesofintercalibration.

4.3. Implicationsfordiatombiogeographyandecology

Apositiveviewoftheresultspresentedhereisthatitis pos-sibletogotoastreamalmostanywhereinEuropewithafairly goodideaofwhatdiatomswemightexpecttoﬁndintheabsence ofanthropogenicpressure.Table3showshowmanydiatomtaxa arecosmopolitanacrossthestudyareaandwhilstsome biogeo-graphicaldifferencesareapparent(lowRAofAchnanthesoblongella andEunotiaspp.coupledwithrelativelyhighRAof Achnanthid-iumpyrenaicuminALP,forexample),itistheconsistencyofthese assemblagesthatismoststriking.Thecounterview,however,is thatthisconsistencywasonlyobtainedbymergingseverallarge

(10)

M.G.Kellyetal./EcologicalIndicators20(2012)177–186 185 speciescomplexesandopinionamongsttheauthorsisdividedon

theextenttowhichthisrepresentsalossofecological informa-tion.Manynewspecieshavebeendescribedinrecentyears(e.g.

Krammer,1997a,b,Lange-Bertalot,2001;Mannetal.,2008); how-ever,theecologicalimplicationsofthisknowledgearenotyetclear. Thereissomeevidenceforspecieswithincomplexeshaving dis-tinctecologicalpreferences(Poulíˇckováetal.,2008,forSellaphora pupulaag.,inlakes;PotapovaandHamilton,2007,for Achnanthid-iumminutissimum)buttwoissuesarise:

• Can these ﬁne-level differences be resolved reliably by ana-lysts(Kahlertetal.,2008,inpress)?Ifnot,additional“signal” isswampedbyanalytical“noise”;

• Dothesedifferencesactuallyrepresent differencesin ecologi-calfunctioning?Thisdecomposes“ecologicalstatus”intotwo components:theeffectofapressureonthephytobenthosand theconsequenteffect ofthe changestothe phytobenthoson othertrophiclevels.Fine-scaletaxonomymightprovideinsights intotheformerbutifgrazers,forexample,donotdifferentiate betweencloselyrelatedspeciesofdiatom,thendifferentiation willaddnothingtoourholisticviewofecologicalstatus. Devel-opmentandselectionofWFDtoolshasfocussedalmostentirely ontheformer(e.g.Heringetal.,2006)whereasgreaterfocuson thelatterwouldseemtoofferanumberofusefulinsightstowater managers.

Aninterpretationofourresults,bearingthesepointsinmindis thatthetaxonomyofKrammerandLange-Bertalot(1986,1997, 2000,2004) maybeadequate torepresentthe mainecological gradientsacrossEurope.Usingthistaxonomy,onlyweak biogeo-graphicpatternsemergesuggestingthattherearefew limitsto dispersionandthemainclimaticdriverssuchaslatitudinal tem-peraturegradientsarelessimportantthanmethodologyandlocal environmentalfactors(Bennettetal.,2010).Ifmorereﬁned tax-onomyisused,biogeographicaldifferenceswithinthesegradients mayemergebutinterpretationiscomplicatedwithinandbetween regionsbythelackoftaxonomicprecision.Onceagain,consistent taxonomyisbetterachievablewithinthanbetweenMSsandthe intercalibrationexerciseneedstoworkwithintheseconstraints. 4.4. Implicationsforintercalibration

Thisprojectallowedalargedatasettobecompiledalthoughstill only14ofthe27MSarerepresented.Althoughthegeographical spreadoftheseissufﬁcientlywideforthistoberepresentativeof theEUasawhole,westillneedtointerpretresultswithcaution asthepositioningofecologicalstatusboundariesmayhave signif-icantﬁnancialimplicationsforMS.Resultspresentedheresuggest thatworkingatregionallevelismostappropriate;theremaybea fewtypeswithacharacteristicecologicalresponsebutwebelieve thattheseshouldonlybedistinguishedfromtheregionasawhole whereseveralMScontributedata.Wehave alsoemphasisedin thispaperthatgreatersensitivityshouldbeachievablewithinMS (although,evenhere,dataprecisionshouldnotbeautomatically assumed).

Theresultsalsopointtotheadvantagesofusingmetricssuch as the IPS and TI. Although potentially useful information on assemblagecompositionislostormergedintobroad“sensitivity” categories,suchmetricshelptoiron-outvariabilityduetofactors suchassubstratum,season,positionofthephytobenthos assem-blagewithinamicrosuccession,etc.Theoutcome,foreachregion, isa“benchmarkdataset”,integratingthediatomassemblagesof allsamplesthatsurvivethescreeningprocess.Nonetheless,some typesofstream(e.g.veryhighalkalinitystreamsinCBregion)will beunder-representedasveryfewsitesmeetallscreeningcriteria. Theimplicationofthisisthatthebenchmarkdatasetsshouldguide

theintercalibrationprocessbutnotbetreatedassettingabsolute limits.

ThevaluesinTable6arerecommendedaslimitsforthe con-stituentmetricsofthepICMthatcanbeappliedacrossaregion, forthepurposesofintercalibration.Thesevaluesalsoprovidea meansofvalidatingEforthoseMSwhodonothavereference val-uesoftheirown(e.g.BE-FL,NL).Wealsorecommendthemedian valuesofthebenchmarkdatasetsasinterimexpectedvaluesfor thoseMS whohavenot yetadoptedanationalmethod, believ-ingthattheintercalibrationprocesshasprovidedasimplemetric (pICM)that is effective acrossthe EUplus, now, validated ref-erencevalues. Incombination withthesamplingand analytical methods(CEN,2003,2004),theseprovidean“off-the-shelf” solu-tion for fulﬁlling obligations for assessing phytobenthos. Some furtherresearchis requiredtoexplorevariationwithinregions, to evaluateif within-region (oreven cross-region)super-types are valid, and whether this improves the sensitivity of inter-calibration. The principal limitation is the lack of comparable environmentaldatafrom MS.MostMSshave internally consis-tentenvironmentaldatasetsthathaveenabledthemtodevelop robusttypologiesormodelsthataremoresensitivetotheirown conditionsthanvaluesproducedhereandwemustemphasisethat thisexercisevalidatesbutdoesnotreplacenationalreference val-ues.

The expectation, therefore, is that metric values for a MS’s referencesitesshouldfallwithinthestatisticallimitsofthe bench-markdatasetfortheregioninquestion.Statisticaltestingofthe hypothesisthatsamplesaredrawnfromthesamepopulationis complicated bythenon-parametric distributionand interpreta-tionisnotstraightforwardasunderlyingtypologicalfactorsmay exertaninﬂuencewhichmayclustertheMS’sdatatowardsone endofthebenchmarkdataset.Finally,althoughwehavescreened sitesrigorouslyusingabioticcriteria,therewillbeafewsamples thatsuggestedenrichment,eitherduetounknownpollutionevents priortothesampling/screeningexercise,orduetonaturalcauses (deathand decayofananimalcarcassinthestream,for exam-ple).Aswedonotalwayshaveperfectknowledgeofasitepriorto samplingandastheremaybenaturalprocessesthatcould poten-tiallycontributetoanepisodicincreaseinnutrientconcentration, itwouldbeunwisetorejectthesesamplesoutright.Wetherefore suggestthatthe90thpercentileforTI(10thpercentileforIPS)as apragmaticcriterionforMStousetochecktheirownreference condition.ThoseNRsitesthatfalloutsidetheselimitswillneed tobere-evaluatedandtheexpectationshouldbethatmostofthe referencesiteschosenbyaMSnotinvolvedinthisexercisewill fallwithinthe90th(or10th)percentileunlesstheMScan other-wiseprovideanecologicaljustiﬁcationforthedeviation.Overall, ourambitionshouldbetominimisetheroleofexpertjudgement and“gardening”ofdatasetstopurgesitesthatdonotconformtoa subjective“ideal”.

4.5. Concludingcomments

Thenatureoftheintercalibrationexercise,followingaperiod ofmethoddevelopmentbyindividualMS,meansthattheprocess haslittleopportunitytoinﬂuencecollectionofthedatawithwhich ithastowork.Eventhoughcommonstandardsfordatacollection andanalysisareapparentlyinuse,therewasconsiderablevariation duetoMS,whichlimitsthesensitivityoftheanalysesperformed. Intercalibrationis,however,the“artofthepossible”,recognising thattheprimarypurposeofdatacollectedbyindividualMSistoﬁt intoongoingwatermanagementprograms.Yet,atthesametime, theintercalibrationexercisehaspresentedanexcellent opportu-nitytogeta continent-wideviewofthecomposition ofdiatom assemblagesinnear-pristinecondition.

(11)

Acknowledgements

Contributorstothispaperacknowledgethesupportofnational environmentalagenciesforprovisionofdataandﬁnancialsupport. M.G.K.: Scottish and Northern Ireland Forum for Environmen-tal Research; M.K.: Swedish Environmental Protection Agency; S.F.P.A.:PortugueseWaterInstitute(INAG);L.O.andP.M.:Ministry oftheEnvironmentoftheCzechRepublicwithintheInstitutional researchplanMZP0002071101–ResearchandProtectionofthe Hydrosphere;S.S.:NorwegianDirectorateforNatureManagement. Inaddition,wearegratefultotheSwedishEnvironmental Protec-tionAgencyforsupportfortheworkshopattheErkenFieldStation inSwedenandtoRichardJohnson.(UniversityofAgricultural Sci-ences,Uppsala,Sweden)andEckhardCoring(EcoRing,Hardegsen, Germany)forhelpfulcommentsonadraftmanuscript.

References

Bailey,R.C.,Norris,R.H.,Reynoldson,T.B.,2004.BioassessmentofFreshwater Ecosys-tems:UsingtheReferenceConditionApproach.KluwerAcademicPublishers, NewYork.

Bennett,C.,Owen,R.,Birk,S.,Buffagni,A.,Erba,S.,Mengin,N.,Murray-Bligh,J., Ofenböck,G.,Pardo,I.,vandeBund,W.,Wagner,F.,Wasson,J.-G.,2011.Bringing Europeanriverqualityintoline:anexercisetointercalibratemacro-invertebrate classiﬁcationmethods.Hydrobiologia667,31–48.

Bennett,J.R.,Cumming,B.F.,Ginn,B.K.,Smol,J.P.,2010.Broad-scale environmen-talresponseandnicheconservatisminlacustrinediatomcommunities.Global EcologyandBiogeography19,724–732.

CEN,2003.Waterquality-Guidancestandardfortheroutinesamplingand pre-treatmentofbenthicdiatomsfromrivers.EN13946:2003.ComitéEuropeande Normalisation,Geneva.

CEN,2004.Waterquality–Guidancestandardfortheidentiﬁcation, enumera-tionandinterpretationofbenthicdiatomsamplesfromrunningwaters.EN 14407:2004.ComitéEuropeandeNormalisation,Geneva.

Clarke,K.R.,1993.Non-parametricmultivariateanalysisofchangesincommunity structure.AustralianJournalofEcology18,117–143.

CEMAGREF,1982.EtudedeMéthodesBiologiquesQuantitativesd’Appreciationde laQualitédesEaux.RapportQ.E.Lyon-A.F.B.Rhône-Mediterrannée-Corse. ECOSTAT,2004.OverviewofCommonIntercalibrationTypes.WorkingGroup2.

AEcologicalStatus,Final Version5.1. Availablefrom: http://ec.europa.eu/ environment/water/water-framework/objectives/indexen.htm.

ECOSTAT,2010.GuidanceDocumentontheIntercalibrationProcess2008–2011.102 pp.CommonImplementationStrategyfortheWaterFrameworkDirective. Hammer,O.,Harper,D.A.T.,Ryan,P.D.,2001.PAST:palaeontologicalstatistics

soft-warepackageforeducationanddataanalysis.PalaeontologicaElectronica4(1), 9pp.

Hering,D.,Johnson,R.K.,Kramm,S.,Schmutz,S.,Szoszkiewicz,K.,Verdonschot, P.F.M.,2006.AssessmentofEuropeanstreamswith diatoms,macrophytes, macroinvertebratesandﬁsh:acomparativemetric-basedanalysisoforganism responseduetostress.FreshwaterBiology51,1757–1785.

Kahlert,M.,Albert,R.-L.,Anttila,E.-L.,Bengtsson,R.,Bigler,C.,Eskola,T.,Gälman,V., Gottschalk,S.,Herlitz,E.,Jarlman,A.,Kasperoviciene,J.,Kokoci ński,M.,Luup,H., Miettinen,J.,Paunksnyte,I.,Piirsoo,K.,Quintana,I.,Raunio,J.,Sandell,B.,Simola, H.,Sundberg,I.,Vilbaste,S.,Weckström,J.,2008.Harmonizationismore impor-tantthanexperience–resultsofthefirstNordic-Balticdiatomintercalibration exercise2007(streammonitoring).JournalofAppliedPhycology21,471–482. Kahlert,M.,Kelly,M.G.,Albert,R-L.,Almeida,S.F.P.,Beˇsta,T.,Blanco,S.,Coste,M., Denys,L.,Ector,L.,Fránková,M.,Hlúbiková,D.,·Ivanov,P.,Kennedy,B., Mar-van,P.,Mertens,A.,·Miettinen,J.,·Picinska-Fałtynowicz,J.Rosebery,J.,Tornés, E.,·Vilbaste,S.,Vogel,A.Identificationisaminorsourceofuncertaintyin diatom-basedecologicalstatusassessmentsonacontinent-widescale:resultsofa Europeanring-test.Hydrobiologia,inpress.

Kelly,M.G.,Ector,L.Effectofstreamliningtaxalistsondiatom-basedindices: impli-cationsforintercalibratingecologicalstatus.Hydrobiologia,inpress. Kelly,M.G.,Cazaubon,A.,Coring,E.,Dell’Uomo,A.,Ector,L.,Goldsmith,B.,Guasch,H.,

Hürlimann,J.,Jarlman,A.,Kawecka,B.,Kwandrans,J.,Laugaste,R.,Lindstrøm, E.-A.,Leitao,M.,Marvan,P.,Padisák,J.,Pipp,E.,Prygiel,J.,Rott,E.,Sabater,S., vanDam,H.,Vizinet,J.,1998.Recommendationsfortheroutinesamplingof diatomsforwaterqualityassessmentsinEurope.JournalofAppliedPhycology 10,215–224.

Kelly,M.,Bennett,C.,Coste,M.,Delgado,C.,Delmas,F.,Denys,L.,Ector,L.,Fauville,C., Ferréol,M.,Golub,M.,Jarlman,A.,Kahlert,M.,Lucey,J.,NíChatháin,B.,Pardo,I., Pﬁster,P.,Picinska-Faltynowicz,J.,Rosebery,J.,Schranz,C.,Schaumburg,J.,van Dam,H.,Vilbaste,S.,2009.Acomparisonofnationalapproachestosetting eco-logicalstatusboundariesinphytobenthosassessmentfortheEuropeanWater FrameworkDirective:resultsofanintercalibrationexercise.Hydrobiologia621, 169–182.

Kelly,M.,Juggins,S.,Guthrie,R.,Pritchard,S.,Jamieson,J.,Rippey,B.,Hirst,H.,Yallop, M.,2008.AssessmentofecologicalstatusinUKriversusingdiatoms.Freshwater Biology53,403–422.

Krammer,K.,1997a.DieCymbelloidenDiatomeenEineMonographiederweltweit bekanntenTaxa.Teil1:AllgemeinesundEncyonemaPart.Bibliotheca Diatomo-logica36,1–382.

Krammer,K.,1997b.DieCymbelloidenDiatomeenEineMonographiederweltweit bekanntenTaxa.Teil2:Encyonemapart,EncyonopsisandCymbellopsis. Biblio-thecaDiatomologica37,1–469.

Krammer,K.,Lange-Bertalot,H.,1986.DieSüßwasserﬂoravonMitteleuropa2: Bacil-lariophyceae.Teil1:Naviculaceae.GustavFischerVerlag,Stuttgart.

Krammer,K.,Lange-Bertalot,H.,1997.DieSüßwasserﬂoravonMitteleuropa2: Bacil-lariophyceae.Teil2:Bacillariaceae,Epithemiaceae,Surirellaceae.2teAuﬂage. GustavFischerVerlag,Stuttgart.

Krammer,K.,Lange-Bertalot,H.,2000.DieSüßwasserﬂoravonMitteleuropa2: Bacil-lariophyceae.Teil3:Centrales,Fragilariaceae,Eunotiaceae.2teAuﬂage.Gustav FischerVerlag,Stuttgart.

Krammer,K.,Lange-Bertalot,H.,2004.SüßwasserﬂoravonMitteleuropa2: Bacil-lariophyceae.Teil4:Achnanthaceae.KritischeErgänzungenzuAchnanthess.l., Naviculas.str.,Gomphonema.SpektrumAkademischerVerlag/GustavFischer, Heidelberg.

Lange-Bertalot,H.,2001.DiatomsofEurope,vol.2.Naviculasensustricto,10Genera SeparatedfromNaviculasensulato,FrustuliaA.R.G.GantnerVerlagK.G.,Ruggell. Mann,D.G.,Thomas,S.J.,Evans,K.M.,2008.RevisionofthediatomgenusSellaphora: aﬁrstaccountofthelargerspeciesintheBritishIsles.Fottea,Olomouc8,15–78. Marker,A.F.H.,1976.ThebenthicalgaeofsomestreamsinsouthernEnglandI Biomassoftheepilithoninsomesmallstreams.JournalofEcology64,343–358. Oksanen,J.,Kindt,R.,Legendre,P.,O’Hara,R.B.,2007.Vegan:CommunityEcology

Packageversion1.8-6.http://cran.r-project.org/.

Potapova,M.,Hamilton,P.B.,2007.Morphologicalandecologicalvariationwithin theAchnanthidiumminutissimum(Bacillariophyceae)speciescomplex.Journal ofPhycology45,561–575.

Poulíˇcková,A., ˇSpaˇcková,J.,Kelly,M.G.,Duchoslav,M.,Mann,D.G.,2008.Ecological variationwithinSellaphoraspeciescomplexes(Bacillariophyceae):specialists orgeneralists?Hydrobiologia614,373–386.

RProjectCoreDevelopmentTeam,2005.R:Alanguageandenvironmentfor sta-tisticalcomputing.RFoundationforStatisticalComputing,Vienna,Austria.

http://www.R-project.org.

Rott,E.,Pipp,E.,Pﬁster,P., vanDam,H., Ortler,K.,Binder,N.,Pall, K.,1999. IndikationslistenfürAufwuchsalgeninösterreichischenFließgewassern.Teil 2: Trophieindikation(sowiegeochemische Präferenzen,taxonomische und toxikologischeAnmerkungen).Wasserwirtschaftskatasterherausgegebenvom Bundesministeriumf.Land-u.Forstwirtschaft,Wien.

Stoddard,J.L.,Larsen,D.P.,Hawkins,C.P.,Johnson,R.K.,Norris,R.H.,2006.Setting expectationsfortheecologicalconditionofstreams:theconceptofreference condition.EcologyApplied16,1267–1276.

Tison,J.,Park,Y.-S.,Coste,M.,Wasson,J.G.,Ector,L.,Rimet,F.,Delmas,F.,2005. Typologyofdiatomcommunitiesandtheinﬂuenceofhydro-ecoregions:astudy ontheFrenchhydrosystemscale.WaterResearch39,3177–3188.

Wallin, M., Wiederholm, T., Johnson, R.K., 2005. Guidance on establishing reference conditions and ecological status class boundaries for inland surface waters. 93pp. CIS Working Group 2.3– REFCOND. http://www-nrciws.slu.se/REFCOND/7thREFCONDﬁnal.pdf.