RESEARCHARTICLE Invasion strategy and abiotic activity triggers for non-native gobiids of the River Rhine JanBaer1*,FrankHartmann2,AlexanderBrinker1 1 FisheriesResearchStationBaden-Wu¨rttemberg,LAZBW,Langenargen,Germany, 2 Regierungspra¨sidiumKarlsruhe,FisheriesAdministration,Karlsruhe,Germany *[email protected] a1111111111 a1111111111 Abstract a1111111111 a1111111111 The24houractivitypatternsofthreenon-nativegobiids(roundgobyNeogobiusmelanosto- a1111111111 mus,WesterntubenosegobyProterorhinussemilunarisandbigheadgobyPonticolakes- sleri)wereassessedover46consecutivemonthsbetween2011and2014fromtheir occurrenceinthecoolingwaterintakeofanuclearpowerplantontheRiverRhine,Ger- many.Intotal,117717gobiidswereidentifiedandclassified.Theoccurrenceofallthree OPENACCESS speciesvariedstronglybetweensamplingyears,andspecies-specificactivitytriggerswere Citation:BaerJ,HartmannF,BrinkerA(2017) identified.Theactivityofjuvenilesofallthreegobiidsspecieswaspositivelytemperature Invasionstrategyandabioticactivitytriggersfor dependentwhileadulttubenosegobyactivityappearedtobenegativelytemperaturedepen- non-nativegobiidsoftheRiverRhine.PLoSONE dent.Increasingfluvialdischargeintheadjoiningmainriverstimulatedtheactivityofjuvenile 12(9):e0183769.https://doi.org/10.1371/journal. pone.0183769 roundgobybutinhibitedactivityofadulttubenosegoby.Exceptforadultbigheadgoby, activitywasalsostructuredbytimeofday,butwithnouniformmean.Meteorologicalfactors Editor:UlrikeGertrudMunderloh,Universityof Minnesota,UNITEDSTATES suchasprecipitation,airpressureanddurationofsunshinehourshadlittleornoinfluence ongobiidactivity.Onselectedrareoccasions,mainlyatnight,allthreespeciesexhibited Received:May5,2017 pulsedswarmingbehaviour,withthousandsofindividualsrecordedintheintakewater. Accepted:August10,2017 Roundgobyswarmsexhibitedboththehighestintensityandthelargestswarmingindividu- Published:September15,2017 als,suggestingapotentialcompetitiveadvantageovertubenoseandbigheadgoby.Electric Copyright:©2017Baeretal.Thisisanopen fishingsurveysinnaturalriverstretchescorroboratedthisobservation.Negativeeffectson accessarticledistributedunderthetermsofthe thenativefishfaunawereapparentonlyforthebullhead,Cottusgobio. CreativeCommonsAttributionLicense,which Theactivitytriggersidentifiedofferauniqueinsightintotheinvasionmechanismsof permitsunrestricteduse,distribution,and reproductioninanymedium,providedtheoriginal theseecosystem-changingnon-nativegobiids. authorandsourcearecredited. DataAvailabilityStatement:Thecollectionand countingofthefishwasdoneunderthepermission andpaymentofthenuclearpowerplantcompany (EnBW)andbelongstothecompany.The Introduction meterologicaldatabelongstothirdparty,too (Landesanstaltfu¨rUmwelt,Messungenund Theuseofwaterfromnaturalsystemsasacoolantinthermalpowerstationsleadstoawide Naturschutz).Theauthorgotthepermissionto rangeofecologicalimpactsonaquaticcommunitiesinboththeintakeandouttakestretches. workwiththedata(statisticalanalyses),butnotto Thermalloadingofcoolingwaterinterferesdirectlywithphysiologicalprocessesofthebiota, publishthemindetail.Duetothoserestrictons,the suchasenzymeactivity,feeding,reproduction,respiration,growthandphotosynthesis[1].Of datasetcannotbemadepubliclyavailable. stillgreaterpotentialimpact,however,arethelossesofvariouslife-stagesofinvertebratesand Interestedreseachersshouldcontactthe fishescapturedonintakescreensorentrainedwithincoolingsystems.Itisnotuncommonfor Landesanstaltfu¨rUmwelt,Messungenund Naturschutz([email protected])torequest millionsoffishesandcrustaceanstobeimpingedonpowerplantintakescreenseachyear[2] PLOSONE|https://doi.org/10.1371/journal.pone.0183769 September15,2017 1/16 Invasionstrategyandactivitytriggersforgobiids meterologicaldata,Mr.Dr.Vogel(t.vogel@kk. andmanymoreeggs,planktonandlarvaeofinvertebratesandfishesarealsokilled.Absolute enbw.com)fromtheNPPPfordataregardingthe mortalityfiguresatsomepowerplantshavebeenshowntoexceed1010individualsannually nuclearpowerplant(amountintakewater)andMr. [2]. Dr.FrankHartmann([email protected]) Thehugedemandforcoolingwaterbylargethermalpowerstationsmeansmostarelocated forthegobiiddata. nearlargewaterbodies.Permanentmonitoringofthefishfaunaatsuchlocationsistime, Funding:Theauthorsreceivednospecificfunding labour,andcostintensiveandmethodsarelimitedbytheoftenconsiderabledepthofwater, forthiswork. strongcurrents,impairedvisibilityandotherconfoundingfactors.Standardsamplingtech- Competinginterests:Theauthorshavedeclared niques,suchaselectricfishing,divingortheuseofgillnets,allhavedifferentdisadvantages thatnocompetinginterestsexist. renderingthemmostlyineffectiveinthiscontext[3].However,thesheervolumeofintake watersinlargethermalpowerplants,likeinnuclearpowerplants,constitutesasignificant samplingoffishstocksinthesourceriverandstudiesoffishoccurrenceincoolingwater intakeshavefacilitatedthedevelopmentofimportanttimeseries[4],withwhichtostudygen- eralbiologicaldynamics,suchaschangesinfishcommunitycomposition.Nevertheless,most ofthesestudiesconductedatnuclearpowerplantsarelookingfortheindirecteffectofthermal loading[5–9],whilethecostandlogisticsofstudieslookingforthedirecteffectsoftheremoval offishfromtheintakewaterinnuclearpowerplantsmakessuchprojectscomparativelyrare [4].Howeversuchdatamightprovideaninsightintootherwiseunknowndynamicsandin particulartheinvasionstrategiesofvariousnon-nativefishesinlargerivers[10,11]. TheroundgobyNeogobiusmelanostomus(Pallas,1814)isoneoftheworld’smostwide- ranginginvasivefish,whoserapidexpansionanddeleteriousecosystemeffectsarewelldocu- mented[12].However,detailedstudiesofautecologicalaspectsofinvasionssuchasactivity triggersorco-occurrenceeffectsduringthecolonisationofanewhabitatarelargelymissing [13].Thepresentstudytakesadvantagesofauniquedatasetgeneratedbyhourlyscreeningof fishandlampreyspeciesfromthecoolingwaterintakeofanuclearpowerplantontheRiver Rhine.Inthepresentstudy,thishighlyresolveddataisusedtodocumentthechronologyand theintensityofinvasionsofthreenon-nativegobiidsandtocorrelatetheiroccurrencewith abioticfactors.Gobyinvasionsfirstreachedtheareaofthesamplingstationin2007:firstcame theWesterntubenosegobyProterorhinus semilunaris(Heckel,1837),thenthebigheadgoby Ponticolakessleri(Gu¨nther,1861)in2010andfinallytheroundgobyin2011[14].Allthree speciesmostprobablyarrivedviatheMain-DanubeCanal,builtin1992toconnecttheDan- ubewiththeRhinesystem[14].Thegoalofthisstudywastogainareliablepictureofthegoby invasionfront,inordertobetterunderstandgeneralinvadingmechanisms,andtoidentify abioticactivitytriggers. Materialandmethods Ethicsstatement Approvalofourpresentstudybyareviewboardinstitutionorethicscommitteewasnotneces- sarybecauseallfishwerecaughtunderthepermissionofthelocalfisheriesadministration(F. Hartmann)andallneededqualificationsfortheinvolvedpeople(fishinglicenses)were checkedregularlybythelocalmemberoftheanimalprotectioncommittee(F.Hartmann). Basedonthelocalfisherylaw(Landesfischereiverordnung,§2)allnon-nativefishhastobe removed,thereforeallgobiidswerestunnedbyablowontheheadandexpertlykilledimmedi- atelybyacardiacstabaccordingtotheGermanAnimalProtectionLaw(§4)andtheordi- nanceofslaughterandkillingofanimals(Tierschlachtverordnung§13).Nolivinggobiids wereused.Livenativefishcapturedinthecoolingwaterweremaintainedin50Lplastichold- ingtanks,thecontentsofwhichwerereleasedeverysixhoursinabucketloweredintothe mainchanneloftheriver.Electrofishingwasconductedunderalicensefromthefisheries administration(Regierungspra¨sidiumKarlsruhe). PLOSONE|https://doi.org/10.1371/journal.pone.0183769 September15,2017 2/16 Invasionstrategyandactivitytriggersforgobiids Fig1.LocationofthePhilippsburgnuclearpowerplant(NPPP)inGermanyandsituationontheRiverRhine(herenamedRhein, arrowindicatingdirectionofflow). https://doi.org/10.1371/journal.pone.0183769.g001 Studyarea ThePhilippsburgnuclearpowerplant(NPPP)islocatedinsouth-westernGermany,approxi- mately30kmnorthofthecityKarlsruhe,onasmallisland(Rheinschanzinsel),642.5km upstreamfromthemouthoftheRiverRhine(49˚15´N;08˚26´E)(Fig1).Waterfromthe Rhinehasbeenusedforcoolingthefirst(926Megawatt)unitofNPPPsince1979andthesec- ond(1468Megawatt)unitsince1984.AftertheFukushimanucleardisasterin2011,theGer- mangovernmentdecidedtoabandonnuclearpower.NPPPunit1ceasedoperationsthesame yearanditscoolingwaterintakewasreducedaccordingly.Unit2isduetobeshutdownin 2019accordingtoGermanatomiclaw(Atomgesetz,§7,1a). Duringthestudyperiod(01.01.2011–31.10.2014)thecoolingwaterintakeofNPPPwas between3and96ms-1(mean:40.8±standarddeviationSD20.9ms-1).Thiswaterisallowed tobewarmedby8–10˚Cabovethemeanrivertemperaturebeforebeingdischarged.The intakeofcoolingwaterhappensviaanartificiallakeconnectedwiththemainchannelofthe Rhine.Thelakehasameanwettedwithof250mandameanwettedlengthof720m(Fig1). Waterisreturneddirectlytothemainriver,500mdownstreamofLakeInlet(Fig1).The RiverRhineadjacenttothemouthofthelakeisapproximately220mwide,withanaverage depthof3.75mandayearlyaveragedischargeof1265ms-1.Despitethelocationofthecool- ingintakebeingonthelakesome300mfromtheriver,abstractionisconsideredtotakeplace fromthemainstreamingwaterbodyoftheRhine.Theintakestructurecandrawupto100m s-1ofwater,comparabletotheannualdischargeofamediumsizedriverinthispartofGer- many,andmorethanmostlargertributariesdelivertotheRiverRhine.Thisvolumeissuffi- cienttogenerateanobservableflowofwaterfromthemainriverchanneltotheintake structure.Theintakehappensthroughtenverticalconcretesluices,approximately5mhigh and3mwideeach.Waterisdrawnfromthefullwaterdepthoftheartificiallakeand PLOSONE|https://doi.org/10.1371/journal.pone.0183769 September15,2017 3/16 Invasionstrategyandactivitytriggersforgobiids transportedthroughdifferentbarracks,attheendofwhichfishes,crustaceans,plantmaterial anddebrisareremovedby12verticalscreenswithasquaremesh-sizeof1mm.Retained materialisflushedintoatank,thesamplingpoint. Sampling Duringthestudyperiod,thetotalbodylengths(TL)ofallfishlongerthan2cmflushedinto thesamplingtankweremeasuredtothenearestcm.Individualsofeachspeciesandeach1cm length-classwerecounted.Inordertoreducelabour-intensivespeciesidentificationworkat timeswhenfishloadswerehigh,fishoflessthan2cmTLwereassignedtoonegenericgroup called´fishlarvae’.Thesamplingtankwaschecked24timesdaily(attheendofeachhour),by ateamof20trainedpersonnel(biologistsandsupervisedstudents),exceptduringannual NPPPserviceandmaintenanceoperationswhencheckswereonlycarriedouteverysecond day(15.05.-11.06.2011;30.05–16.07.2012;05.05.-04.08.2013;22.06.-21.08.2014).Hourlysam- plingwasthusachievedon1308days. TheNPPPcoolingwaterintakeactsasapassivefishinggear[3]andthewaterincludesonly migratingordriftingindividuals.Thusinthecontextofthisstudy,theoccurrenceofanindi- vidualinthesamplingcontainerisregardedasaproxyforspeciesactivity. Single-passelectricfishing(EFKO8000,straightDC,300–600V,Leutkirch,Germany)was usedtodeterminethecatchperuniteffort(CPUE,individualsm-2)ofgobiidsattwostretches oftheriver.Bothlocationswerefishedfromaboat.Theseelectricfishingsurveysfocussed mainlyonshallow,rockyandsandyareas,thepreferredgobiidhabitats.Thesesiteswere10 kmdownstreamoftheNPPP(nearthevillageofKetsch,49˚21´N;08˚29´E)and35km upstreamoftheNPPP(nearthevillageofPlittersdorf,48˚53´N;08˚08´E).Atthedownstream sitea1000–2000mstretchwasfishedoneightoccasionsbetween2008and2015,whileatthe upstreamsite,fishingwascarriedoutona100to3330msectionsixtimesoverthesame period. Datatreatment Basedonaliteraturereview,allsampledgobiidslongerthan2cmTLwereclassifiedasjuve- niles((cid:20)4cmTL)orpotentiallymatureindividuals(>4cmTL)[12,15–17].Theparameters selectedforinterrogationwithmultivariatestatisticsinthecurrentstudyare:dailyarithmetic meansofdischargeoftheRhine(ms-1)nearNPPP(PegelMaxau,Wasser-undSchifffahrt- samtMannheim);watertemperature(˚C)oftheriver20kmupstreamofNPPPin(LUBW); dailydurationofsunshinehours;precipitation(mlperday)anddailychangeinairpressure (hPa)comparedtothepreviousday.Thelatterthreeparameterswereallmeasuredbya weatherstation3kmsouthofNPPP(DWD).Othervaluesincorporatedintotheanalyseswere coolingwaterintakeinms-1,timeofday(night,day,andnauticaltwilightaccordingtoGlar- ner2006[18])andstudyyear.Thegeneralregressionmodelspecifiedthenegativebinomial distributionfortheresponsevariablefollowingtheargumentationofO’HaraandKotze(2010) [19],becauseofthehighnumberofzeroobservations.ApowerfulLassotechnique[20],well suitedforlargedatasetswherecollinearityisoftenaproblem,wasusedtocondensetheinde- pendentvariables.Toassesstheeffectstrengthofsignificantindependentvariablesinthe modelformula,MonteCarlosampleswereobtainedbyresamplingtheobserveddata.This wasdoneundertheassumptionthatfactorsareuncorrelatedandthattheirlikelyvalueswould notberepresentedbyauniformdistribution. AllstatisticswererunonJMPPro13.1.0(64bit,SASInstitute). PLOSONE|https://doi.org/10.1371/journal.pone.0183769 September15,2017 4/16 Invasionstrategyandactivitytriggersforgobiids Table1. Absolutenumbersofthe15mostdetectedspeciesinthesamplingcontainerofthePhilippsburgnuclearpowerplantbystudyyear;the threegobiidspeciesinfocusarehighlightedinbold. Species 2011 2012 2013 2014 Total Roach(Rutilusrutilus) 118530 176123 525821 127774 948248 Bream(Abramisbrama) 51113 93584 538179 227809 910685 Perch(Percafluviatilis) 73693 90751 202005 428251 794700 Zander(Sanderlucioperca) 169635 33135 50791 195132 448693 Asp(Leuciscusaspius) 103932 7417 2366 5791 119506 Bleak(Alburnusalburnus) 17771 26151 24216 28515 96653 Roundgoby(Neogobiusmelanostomus) 1458 43033 26888 18899 90278 Three-spinedstickleback(Gasterosteusaculeatus) 10190 18505 37750 12207 78652 Commonnase(Chondrostomanasus) 28429 2854 3694 5237 40214 Ruffe(Gymnocephaluscernua) 6379 4161 11215 271 22026 Sealamprey(Petromyzonmarinus) 7290 5972 3596 1752 18610 Riverlamprey(Lampetrafluviatilis) 4405 6158 4065 1244 15872 Bigheadgoby(Ponticolakessleri) 4303 2102 8047 297 14749 Westerntubenosegoby(Proterorhinussemilunaris) 10466 1150 825 249 12690 Commondace(Leuciscusleuciscus) 5233 434 1.801 201 7669 https://doi.org/10.1371/journal.pone.0183769.t001 Results Duringthestudyperiodatotalof6654592fishwerecountedinthesamplingcontainer.45.1% ofthem(n=3001388)werelarvae(<2cmTL,mostlycyprinids)and46.6%(n=3102326) wereroach(Rutilusrutilus),bream(Abramisbrama),perch(Percafluviatilis)orzander (Sanderlucioperca)(cid:21)2cmTL.Theremainder(8.3%,all(cid:21)2cmTL,n=550878)belongedto 42differentteleosttaxaandthreelampreyspecies.Of46identifiedfishtaxa,33areconsidered autochthonous,constituting84.6%ofthenaturalfishfaunaoftheRhine(n=39),while13are allochthonous(Baeretal.2014).Inoverallabundanceduringthewholestudyperiod,round gobywererankedinseventhplace(n=90278),bigheadgobyatthirteenth(n=14749)and tubenosegobyatfourteenth(n=12690)(Table1). Gobiidswererecordedinthesamplingcontaineralmostyearround,on91.5%ofstudy days. Therelativeproportionsofjuvenilegobiidsinthesamplingtankdifferedbetweenspecies: 22.1%oftubenosegoby,28.7%ofbigheadgoby,and51.1%ofroundgobywerebelowthe5cm TLthresholdforpotentialadulthoodandwerethusregardedasjuveniles.Maximumrecorded sizeswere20cmTLforbigheadgoby,19cmTLforroundgobyand16cmTLfortubenose goby. Thethreenon-nativegobiidsoccurredatallhoursoftheday,butthemajoritywere recordedbetween9:00amand5:00pmthanatothertimes(Figs2and3).Theoccurrenceof juveniletubenosegobypeakedinthesecondhalfofthenight(1:00–4:00am)andaround dawn(7:00–8:00am)(Fig2);largerindividualswerefoundmainlybetween11:00pmand5:00 am,butweregenerallyabsentatdawn(Fig3).Nearly45%ofalljuvenilebigheadgobywere detectedinthehourjustaftermidnight,withanothersmallerpeakbetween4:00and5:00am (Fig2).Adultbigheadgobiesweredetectedaroundtheclock,withsmallpeaksinthelatter halfofthenightandarounddawn(Fig3).Theoccurrenceofjuvenileroundgobiesincreased aftersunsetandpeakedonehouraftermidnight,withasecondpeakarounddawn(Fig2). Adultroundgobieswererecordedatallhours,withpeaksduringthesecondhalfofthenight andarounddawnanddusk(Fig3). PLOSONE|https://doi.org/10.1371/journal.pone.0183769 September15,2017 5/16 Invasionstrategyandactivitytriggersforgobiids Fig2.Relativefrequenciesofjuvenilenon-nativegobiidsinanaverage24hourperiodbasedonindividuals detectedinthesamplingcontainerduringthestudyperiod;notethediscontinuityofthey-axis,n=samplesize. https://doi.org/10.1371/journal.pone.0183769.g002 Tubenosegobieswererecordedbetween1and1621individualsperday(mean±standard deviationSD:20.1±516.5)on631studydays(48.5%ofallstudydays).Theyappearedmainly duringApril,MayandearlyJuneandagaininSeptemberandOctober(Fig4).Bigheadgoby werefoundon598studydays(45.7%),mainlyduringsummer(Fig4)between1and1968 individualsperday(mean±SD:49.2±612.0).Roundgobieswererecordedon998days (76.3%)indensitiesbetween1and14193individualsperday(mean±SD:90.4±591.5), mainlyduringsummerorautumn(Fig4). Allthreespeciesofinvasivegobyexhibitedpulsedswarmingbehaviour,withmorethan 1000individualsdetectedwithinaonetotwohourperiodonafewdaysinsummer(Fig5) and,inthecaseofroundgobyalsoduringautumn.Swarmingbehaviourwasparticularlyobvi- ousfortubenosegobyonthreedaysin2011,forbigheadgobiesontwodaysin2013andfor roundgobiesonfivedaysin2012,fivedaysin2013andtwodaysin2014(Fig4).Theexcep- tionallyhighoccurrencesrecordedatthesetimesweremainlylimitedtothesecondhalfofthe night(Fig5),butonthe18thand24thofAugust2013morethan1000roundgobiesperhour werecaughtinthemorningbetween5:00–9:00amandononedayinOctober2012asimilar captureratewasobservedarounddusk(17:00–18:00pm).Themeanbodylengthsof PLOSONE|https://doi.org/10.1371/journal.pone.0183769 September15,2017 6/16 Invasionstrategyandactivitytriggersforgobiids Fig3.Relativefrequenciesofadultnon-nativegobiidsinanaverage24hourperiodbasedonindividualsdetected inthesamplingcontainerduringthestudyperiod;notethediscontinuityofthey-axis,n=samplesize. https://doi.org/10.1371/journal.pone.0183769.g003 individualsrecordedduringswarmingeventswere5.1cmTL±0.9SDfortubenosesand4.3 cmTL±1.1SDforbighead,withnoindividuals>7cmTL.Bycontrast,TLsofswarming roundgobiesrangedupto14cm(mean±SD:4.4cm±2.5). Mostgobiidsdetectedinthesamplingtankduring2011(Table1)weretubenosegoby (n=10466),withbigheadgobyadistantsecond(n=4303).Roundgobyaccountingforonly 8.9%ofallgobiidrecordsfortheyear2011(n=1458,Table1).Adrasticshiftwasrecordedin 2012whennearly93%ofallgobiidrecordsreferredtoroundgoby(n=43033,Table1).In 2013recordsofbigheadgobypeakedatupto8047individuals,representing25%ofallgobiids (Table1).However,roundgobyremainedthemostabundantgobiidin2013(n=26888)and by2014,bigheadandtubenosegobyhadalmostdisappeared,withroundgobymakingup 97.2%ofallgobiidrecords(n=18899,Table1). Intermsofmonthlycatchcomposition,tubenosegobydominatedthenon-nativegobiids untilJuly2011(Fig6),atwhichpointasharpincreaseofbigheadgobywasobservedresulting inashiftofrelativeshare(Fig6).However,fromautumn2011,occurrencesofroundgoby alsoincreasedandbyJuly2012thisspeciesbegantodominatethecatchofnon-nativegobiids andcontinuedtodosofortheremainderofthestudyperiod(Fig6).Theexceptionwas PLOSONE|https://doi.org/10.1371/journal.pone.0183769 September15,2017 7/16 Invasionstrategyandactivitytriggersforgobiids Fig4.Absolutedailyoccurrence(n)ofgobiidindividualsinthesamplingcontainerbetween1stofApriland31stof Octoberinthefourstudyyears;notethediscontinuityofthey-axis. https://doi.org/10.1371/journal.pone.0183769.g004 summer2012(June),whenduringaphaseofgenerallylowgobiidactivity,largernumbersof tubenosegobieswerefound(Figs4and6).Afinalpeakofbigheadgobyrecordsoccurredin July2013duringnightswithover1000individualsdetectedinthesamplingtankonsome nights(Fig4),butafterthatbothbigheadandtubenosegobyrecordsdwindledtoalmostnoth- ing(Fig6). Lookingintothemodelstatistics,studyyearwasrelevant(P<0.0001)fortheconsecutive appearanceofround,bighead,andtubenosegoby(Tables2and3).Watertemperaturecorre- latedpositively(P<0.01)withtheoccurrenceofjuvenilesofallthreespecies(Table2)and adultsofbigheadandroundgoby(Table3).Incontrast,theoccurrenceofadulttubenosegoby showsasignificant(P<0.0001)negativeinfluenceofwatertemperature(Table3).Fluvialdis- chargewaspositivelycorrelated(P<0.0001)withtheoccurrenceofjuvenilebigheadgoby (Table2)andnegatively(P<0.0001)withadulttubenosegobyrecords(Table3).Timeofday appearedtobeofminorimportancebutcorrelatedsignificantlywiththeoccurrenceofjuve- nile(Table2)andadulttubenoseandroundgoby(Table3),thoughratesdiffered.Thevolume ofcoolingwaterintakecorrelatedpositivelywiththeoccurrenceofjuvenilebighead(P<0.01) androundgoby(P<0.0001)(Table2)andcorrelatednegativelywithrecordsofadultbighead goby(P<0.01)(Table3).Themodelshowednosignificanteffectsofthemeteorological PLOSONE|https://doi.org/10.1371/journal.pone.0183769 September15,2017 8/16 Invasionstrategyandactivitytriggersforgobiids Fig5.ExamplesofpulsedswarmingbehaviourofP.semilunarisonthe13thofJune2011andofP.kessleriandN. melanostomusonthe26thofJuly2013;notethediscontinuityofthey-axis. https://doi.org/10.1371/journal.pone.0183769.g005 factorsprecipitation,sunshinedurationorchangeinairpressurefortubenosegoby,butrec- ordsofjuvenileroundgobywerenegativelyaffected(P<0.05)bychangingairpressure (Table2),occurrenceofadultroundgobiesincreasedwithprecipitation(P<0.01)(Table3) andthedetectionofadultbigheadgobywasnegativelyinfluencedbyincreasingsunshine hours(P<0.0001)(Table3).Theinteractionofthetwoparametersdischargeandwatertem- peraturehadaninfluenceonthedetectionrateofjuvenileroundgoby(P<0.0001)(Table2) andadulttubenosegoby(P<0.0001)(Table3).Overall,themodelstatisticsexplained15– 20%ofobservedvariationingobyrecords(Tables2and3). ElectricfishingsurveysintheRiverRhinein2008yieldednogobiidrecords.Thefirsttube- nosegobieswerecaughtupstreamoftheNPPP(Table4)in2009.Bigheadgobiesbegantobe recordedin2010,bothup-anddownstreamoftheNPPP,thoughtubenosegobystilldomi- natedthecatch(Table4).In2011thefirstroundgobieswerefoundandbigheadgobieshad increasedasaproportionofcatch(Table4).From2012onwardsroundgobiesincreasedenor- mouslyandbigheadandtubenosegobiesnearlydisappeared(Table4).Intherecentpast (2015)roundgobiescontinuedtobefoundinhighdensities(0.3–0.5individualsm-2)and tubenoseandbigheadgobiesarecomparablyrare(Table4).Anotherspeciesofgobyrecently PLOSONE|https://doi.org/10.1371/journal.pone.0183769 September15,2017 9/16 Invasionstrategyandactivitytriggersforgobiids Fig6.Changesinrelativeshareofthenon-nativegobiidsinthemonthlycatchduringthestudyperiod. https://doi.org/10.1371/journal.pone.0183769.g006 detectedapproximately250kmdownstream(monkeygobyNeogobiusfluviatilis,Borcherding etal.2011),wasnotrecordedinthisstudy. Discussion ResultsfromthemonitoringofthecoolingwaterintakeofNPPPagreewiththoseoftheelec- tricfishingsurveys,andbothtechniquesrecordthesametrendsingeneralabundanceofthree non-nativegobiids,includingtherapiddeclineoftubenoseandbigheadgobyaftertheappear- anceofroundgoby.However,thevolumeandexceptionallyhightimeresolutionofdatafrom thecoolingwaterintakeandtheinterpretationoffishrecordsthereinasaproxyforfishactiv- ityoffersauniqueopportunitytoinvestigatepotentialmechanismsunderlyingtheparallel invasionsoftubenose,roundandbigheadgoby. Borcherdingetal.(2013)[10]pointedouttheimportanceofcircadianfactorsinforming theecologicalnichesoftheRhine’sthreeinvasivegobiids.Informationpertainingtothe24 houractivityofthesespeciesisscarce,especiallyoutsidetheirnativerange.Dopazoetal. (2008)[13]reportedthattheproportionofroundgobyinnighttimecatcheswasdoublethat recordedduringtheday.Borcherdingetal.(2013)[10]recordedsignificantlygreaterdensities ofroundgobyinsurveysconductedtwohoursaftersunsetthaninsamplestakendirectlyafter sunsetorduringthedaytime.Jana´četal.(2013)[21]reportedthatthedriftoflarvaltubenose androundgobyoccurredalmostcompletelyduringhoursofdarknessandthatinbothspecies PLOSONE|https://doi.org/10.1371/journal.pone.0183769 September15,2017 10/16