blooms in the Netherlands Combined physical, chemical and biological factors shape Alexandriumostenfeldii Harmful Algae

(1)

Combined physical, chemical and biological factors shape Alexandrium ostenfeldii blooms in the Netherlands

Karen M. Brandenburg

^a,

*, Lisette N. de Senerpont Domis

^a,b

, Sylke Wohlrab

^c

, Bernd Krock

^c

, Uwe John

^c

, Yvonne van Scheppingen

^d

, Ellen van Donk

^a

, Dedmer B. Van de Waal

^a

aDepartmentofAquaticEcology,NetherlandsInstituteofEcology,Wageningen,theNetherlands

bAquaticEcologyandWaterQualityManagementGroup,WageningenUniversity,theNetherlands

cDepartmentofEcologicalChemistry,AlfredWegenerInstitute,HelmholtzCentreforPolarandMarineResearch,Bremerhaven,Germany

dWaterschapScheldestromen,Middelburg,theNetherlands

ARTICLE INFO

Articlehistory:

Received29November2016

Receivedinrevisedform9February2017 Accepted17February2017

Availableonlinexxx

Keywords:

Alexandrium Nutrients Grazing Windspeed Temperature Salinity

ABSTRACT

Harmfulalgal blooms(HABs)aregloballyexpanding,compromising waterqualityworldwide.HAB dynamicsaredeterminedbyacomplexinterplayofabioticandbioticfactors,andtheiremergencehas oftenbeenlinkedtoeutrophication,andmorerecentlytoclimatechange.ThedinoflagellateAlexandrium is one of themost widespreadHAB generaand itssuccess is based onkey functional traits like allelopathy,mixotrophy,cystformationandnutrientretrievalmigrations.Since2012,denseAlexandrium ostenfeldiiblooms(upto4500cellsmL¹)haverecurredannuallyinacreeklocatedinthesouthwestof theNetherlands,anareacharacterizedbyintenseagricultureandaquaculture.Weinvestigatedhow physical,chemicalandbiologicalfactorsinfluencedA.ostenfeldiibloomdynamicsoverthreeconsecutive years(2013–2015).Overall,wefoundadecreaseinthemagnitudeofthebloomovertheyearsthatcould largelybelinkedtochangingweatherconditionsduringsummer.Morespecifically,lowsalinitiesdueto excessiverainfallandincreasedwindspeedcorrespondedtoadelayedA.ostenfeldiibloomwithreduced populationdensitiesin2015.Withineachyear,highestpopulationdensitiesgenerallycorrespondedto hightemperatures,lowDIN:DIPratiosandlowgrazerdensities.Together,ourresultsdemonstratean importantroleofnutrientavailability,absenceofgrazing,andparticularlyofthephysicalenvironment onthemagnitudeanddurationofA.ostenfeldiiblooms.Ourresultssuggestthatpredictedchangesinthe physicalenvironmentmayenhancebloomdevelopmentinfuturecoastalwatersandembayments.

1.Introduction

Globalchangeisoccurringatanunprecedentedrate(Stocker et al., 2013), impacting ecosystems worldwide. In addition to climaticchanges, anthropogenic activities have accelerated the rateandextentofeutrophicationofmanyaquaticenvironmentsas well.Thesechangesgreatlyaffectphytoplankton,standingatthe base of aquatic food webs. Over the past few decades, some phytoplankton species have become an increasingnuisance by forming harmful algal blooms (HABs; Anderson et al., 2002;

Heisleretal.,2008).TheglobalexpansionofHABshasveryoften beenattributedtoeutrophicationofcoastalregions.Changesin nutrientloading,nutrientratiosandnutrientcompositionhavea

tremendous impact on phytoplankton communities living in rivers,estuariesandcoastalzones(Andersonetal.,2002;Smith and Schindler, 2009). For instance, enhanced use of urea as a fertilizerandincreasesinthenitrogenandphosphorustosilicate ratios may promote proliferation of toxic dinoﬂagellates over diatoms(Glibertetal.,2001;Riegman,1995).Inaddition,further changesinclimateinvolvingtemperatureshiftsandsubsequent weatherchanges,mayleadtoanexpansionoftheecologicalniche ofmanyHAB-formingspecies(Andersonetal.,2012b;Hallegraeff, 2010;Wellsetal.,2015).

HABs are known for their adverse effects on ecosystems throughtheircascadingimpactonhighertrophiclevels(Anderson et al.,2002;Hallegraeff,1993).For instance, HABscanproduce toxiccompoundsthatmayaccumulateinthefoodchain,leadingto thedeathofﬁsh,seabirdsandmarinemammals.Moreover,toxins accumulatedinseafoodmaycauseshellﬁshpoisoningsyndromes inhumans(Wang,2008).ProliferationsofHABscanthushavefar

*Correspondingauthor.

E-mailaddress:k.brandenburg@nioo.knaw.nl(K.M. Brandenburg).

ContentslistsavailableatScienceDirect

Harmful Algae

j o u r n a l h o m ep a g e: w w w . e l s e v i e r . c o m / l o c a te / h a l

(2)

reachingecologicalandeconomicconsequences.Thedinoﬂagel- lateAlexandrium ostenfeldii isa globally widespreadtoxicHAB- formingspecies(Balech,1995;FragaandSanchez,1985;Gribble etal.,2005;Johnetal.,2003;Levasseuretal.,1997;Mackenzie et al.,1996; Okolodkov and Dodge,1996; Wang et al., 2006).

Although it used to occur in low numbers in phytoplankton assemblages,inrecent yearsdensebloomsof this specieshave beenreported(Borkmanetal.,2012;Bursonetal.,2014;Hakanen etal.,2012;Krempetal.,2009;Tomasetal.,2012).Alexandrium ostenfeldii is also known to produce various toxins, including ParalyticShellﬁsh Poisoning (PSP)toxins and the cyclic imines gymnodimines and spirolides (Anderson et al.,1990; Cembella etal.,2000;Harjuetal.,2016;Krempetal.,2014).Inadditionto thesetoxins,A.ostenfeldiiproducesextracellularallelochemicals, which can lyse competing phytoplankton species and small protozoan grazers (Tatters et al., 2012; Tillmann et al., 2007;

TillmannandJohn,2002;VandeWaaletal.,2015).Productionof toxinsandlyticcompoundsarekeytraitssupportingAlexandrium proliferation (John et al., 2015; Wohlrab et al., 2016, 2010), particularlysincedinoﬂagellatesaretypicallypoorcompetitorsin terms of growth and nutrient uptake (Litchman et al., 2007;

Smayda, 2002). Other important traits include heterotrophic feeding,cystformation,andnutrientretrievalmigrations(Smayda, 1997).

Althoughthesetraitsareeffectiveforsupportinggrowth,the abioticenvironmentalsoplaysacrucialroleinHABinitiationand subsequentdevelopment.Dinoﬂagellatesresideasrestingcystsin thesediment and,besidesusingendogenousclocksobservedin some species (Anderson and Keafer, 1987), require various environmental stimuli in order to germinate (Anderson et al., 2005).Speciﬁcally,temperature,oxygenconcentrationandlight playaroleincystgermination(Andersonetal.,1987;Dale,1983).

Therefore, cyst resuspension induced by wind mixing may facilitate bloom initiation. Once emerged from the cystbank, however,dinoﬂagellatesareverysensitivetoturbulence(Berdalet et al., 2007; Berdalet and Estrada, 2005; Wyatt and Horwood, 1973),and theirbloomsareoftenassociatedwithcalmweather and water column stability (Berman and Shteinman, 1998;

Margalefet al.,1979WyattandHorwood,1973).Otherphysical controls, suchas temperatureand salinity, as wellas chemical controls, such as nutrient availability are also important in

determining thedevelopment of HABs. Thus, a combination of environmentalconditionswillsetthewindowofopportunityfor HABstodevelop(Andersonetal.,2012a).

Massive annual recurring A.ostenfeldii blooms onlyrecently emergedinaDutchbrackishwatercreek(Bursonetal.,2014),and werefirstobservedin2012.Theinflowofwaterisderivedfromthe agriculturalhinterlands,subsequentlysupplyingthecreekample nutrients.Theoutflowisregulatedbyapumpingstation,which discharges into the Eastern Scheldt, an estuary with the main shellfishfarmingareasoftheNetherlands(VanDerHeijden,2007).

Discharge of creek water with high A. ostenfeldii population densitiesthusformsapotentialthreattothepublichealth.Little, however,isknownaboutthedriversunderlyingtheproliferation of A. ostenfeldii in this creek. Therefore, we investigated how various physical, chemical and biological factors affected the timingand magnitudeofA.ostenfeldii bloomsinthesebrackish waters.Tothisend,wecloselyfollowedanA.ostenfeldiipopulation for three consecutive years, together with temperature, wind speed,rainfall,salinity,nutrients,andzooplankton.

2.Material&methods

ThebrackishwatercreekOuwerkerksekreek(5162⁰N,399⁰E) islocatedintheRhine-Muesse-ScheldtdeltaoftheNetherlands (Fig.1A;Bursonetal.,2014;VandeWaaletal.,2015).Ithasamean depthof 5mwitha maximum depthof 8m, covering roughly 0.12km².

Fielddatawascollectedforthreeconsecutiveyears,startingin April2013.Atthreelocationsinthecreek(Fig.1B),sampleswere taken for Alexandrium ostenfeldii population densities, toxin concentrations (2014–2015), and bacterial abundances (2015) onceeveryweekoreverytwoweeksfromspringuntilautumn.In thesameperiod,additionalmonthlysamplesweretakenattwo locations for phytoplankton and zooplankton determination.

Moreover, monthly samples were taken each year for salinity, inorganic nutrient concentrations and chlorophyll-a concentrations(Fig.1B).Hourlymeteorologicaldata,i.e.temperature,wind speedandprecipitation,wasderivedfromtheweatherstationin Vlissingen,approximately20kmfromourstudysite(with very similar meteorological conditions) and a moving average was calculated.

Fig.1.LocationofA)theOuwerkerksekreekintheNetherlands,andB)samplingpointsintheOuwerkerksekreek,wherethetriangle(northernpart)andsquare(southern part)representthesamplelocationsforallmeasurementsandthecircle(middlepart)theextrasamplelocationforA.ostenfeldiiabundancesandtoxins.

(3)

2.1.Alexandriumostenfeldiiabundances

Anintegratedwatersamplewastakenoftheupper1mofthe watercolumnatallsamplelocations(Fig.1B),fromwhicha50mL subsamplewasﬁxedwithLugol’siodinesolution(Lugol)toaﬁnal concentrationof1%andstoredinthedarkat4Cuntilanalysis.A.

ostenfeldii cells were counted in an Utermöhl chamber on an invertedmicroscope(DMI4000B;LeicaMicrosystemsCMSGmbH, Mannheim, Germany). A. ostenfeldii population densities were determinedbycountingatleast 200cellsor 100ﬁeldsof view (200magniﬁcation).Earliermorphologicalandmolecularanaly- sesshowedthat20isolatedclonesfromthispopulationbelonged toGroup1ofA.ostenfeldii(Krempetal.,2014;VandeWaaletal., 2015).ThestartandendofabloomineachyearwassetwhenA.

ostenfeldiipopulationdensitiesreached10%ofitsmaximum.

2.2.Chlorophyll-a

Chlorophyll-a extractions wereperformed with 80% ethanol accordingtoNusch(1980).Samplesweresubsequentlyanalyzed spectrophotometricallyat665and 750nm (UV-VISSpectropho- tometerUV1650-PC,ShimadzuEurope,Duisburg,Germany),with anadditionalmeasurementatbothwavelengthsafteradditionof 0.4M HCl to a final concentration of 4mM for phaeopigment correction.Calculationswereperformedafter Nuschand Palme (1975), using a chlorophyll-a extinction coefficient of 8.16Lg¹ mm¹, and a ratio between chlorophyll-a and phaeopigment extinctioncoefficientsof1.7.

2.3.Zooplankton

Forzooplanktondetermination,a5Lintegratedwatersample wasﬁlteredovera100

m

^m^mesh^andﬁxedwith96%ethanol.The sample was settled in a large sedimentation chamber with a cuvetteandanalyzedwithaninvertedmicroscope(DMI4000B;

LeicaMicrosystemsCMSGmbH,Mannheim,Germany).Zooplank- tonwas countedand identiﬁed uptogenuslevel. Zooplankton dimensions were assessed using image analysis, and organism volume was subsequently calculated from these dimensions assuminga most similar geometric shape (typicallya sphere).

Biovolume was subsequently calculated by multiplying the organismvolumewithitscounts.

2.4.Salinity

Salinitywascalculatedusingwatertemperatureandconduc- tivity from surface water (<1m) according to the Standard MethodsfortheExaminationofWaterandWastewater(Clesceri etal.,1999).Forfurtherdetailsonthesalinitygradientwereferto Martensetal.(2016).

2.5.Nutrients

Field samples of 15mL for inorganic nutrients were taken (Fig.1B)andstoredat20Cuntilanalysis.Dissolvedinorganic nitrogen(DIN) and dissolved inorganic phosphorus(DIP) were subsequentlymeasuredusingcontinuous ﬂowanalysiscoupled withspectrophotometricdetection(San++AutomatedWetChem- istry Analyzer, Skalar Analytical B.V., Breda, the Netherlands) according to the ISO 13395:1996 and the ISO 15681-2:2003 protocol,respectively.

2.6.Toxinmeasurements

Subsamplesfortoxinanalysesweretakenfromtheintegrated watersamplebyfiltrationof20–60mLoverglassfiberfilters(GF/F,

Whatman, Maidstone, UK), which were stored at 20C until furtheranalysis.

2.6.1.AnalysisofPSPtoxins

PSPtoxinsweredeterminedbyionpairliquidchromatography coupledtopost-columnderivatizationandﬂuorescencedetection, asdescribedinKrocketal.(2007)andVandeWaaletal.(2015).

2.6.2.Analysisofcycliciminetoxinsbytriplequadrupolemass spectrometry

Thecycliciminetoxin measurementswereperformedonan Agilent 1100 LC liquid chromatograph (Waldbronn, Germany) coupledtoa4000QTraptriple-quadrupolemassspectrometer (AB-Sciex,Darmstadt,Germany)withaTurboVionsource.Toxins werequantiﬁedbyexternalcalibrationcurvesofSPX-1andGYMA withstandard solutions rangingfrom10to1000pg

m

^L¹^,^each.

OtherSPXand GYMfor which nostandardsareavailable were calibrated against the SPX-1 and GYM A calibration curve, respectively,and expressedasSPX-1or GYMA equivalents.For furtherdetails,pleasealsoseeVandeWaaletal.(2015).

2.7.Statisticalanalysis

In order to assess the correlation between A. ostenfeldii abundances and different environmental variables, we used Spearman’srho(

r

^),^as^a^linearrelationshipcouldnotbeassumed.

Therelationshipsbetweenthesigniﬁcantenvironmentalvariables were also determined with Spearman’s rho (

r

⁾ ^to ^test ^for

collinearityamongvariables.

3.Results

3.1.Bloomdevelopment

An Alexandrium ostenfeldii bloom was observed in all three years. Themagnitude of thebloomvariedgreatly betweenthe years(Fig.2).In2013,thehighestdensitieswereobserved,with values upto4500cellsmL¹. Populationdensities in2014were somewhat lower, reaching 3200cellsmL¹. In 2015, population densities were substantially lower than previous years, only reaching 800cellsmL¹. The duration of the bloomalso varied greatlybetweentheyears,withthelongestbloomperiodin2014,

Fig.2. SeasonaldynamicsofA.ostenfeldiipopulationdensities,wheresymbols (trianglethenorthernpart,squarethesouthernpartandcirclethemiddlepartof thecreek)representthepopulationdensitiesatthethreesamplelocationsandthe blacklineindicatestheaverage.

(4)

whereitlastedatotalof139daysfromthe17thofJulyuntilthe4th ofOctober,followedby2013,with79daysfromthe4thofJune untilthe21stofOctober.In2015,thebloomlastedonly35days fromthe23rdofJuneuntilthe27thofAugust.

3.2.Physicalenvironment

A.ostenfeldiiabundancescorrespondedstronglytochangesin temperature(

r

⁼^0.74, ^P<0.001;Table 1).Duringbloomperiods temperatures were generally above 15C, corresponding to summerconditions(Fig.3A).Moreover,somecollinearrelation- shipsbetweentemperatureandotherenvironmentalfactorswere also found, namely wind speed (

r

⁼0.56, P<0.001), salinity (

r

⁼^0.40, ^P<0.05), and DIP concentrations (

r

⁼^0.85, ^P<0.001;

Table1).

We observed a negative correlation between A. ostenfeldii abundancesandwindspeed(

r

⁼^0.51,^P<0.001;Table1).Average windspeedwasgenerallylowerduringthebloomsof2013and 2014compared tothe restof the year(around 5ms¹ against 7ms¹),withincreasesinthemiddleandattheendofthe2014 bloomperiod(Fig.3B).In 2015,however,windspeed remained relativelyhighduringthebloomperiodwithonlyabriefdecline towardstheendofthebloom.

Precipitationﬂuctuatedstronglythroughouttheyear,including the bloom periods. Distinct peaks in precipitation could be

recognizedattheendofthebloomin2013,inthemiddleofthe bloomin2014andduringthebloomperiodof2015(Fig.3C).A.

ostenfeldiiabundancesdidnotshowacorrelationwithprecipita- tion (Table 1). Precipitation did, however, show a negative correlationwithsalinity(

r

⁼^0.41,^P<0.05;Table1).

Overthecourseofayear,largeﬂuctuationswereobservedin salinity(Fig.3D),withhighvaluesupto23insummerandlow valuesdownto4.9inwinter.Duringthebloomperiodsof2013and 2014 salinities were largely above 15, while in 2015 salinities remainedbelow15withextremelylowvaluesdownto3.7atthe end of thebloom. A.ostenfeldii abundancesgenerallyincreased withsalinity(

r

⁼^0.42,^P<0.05;Table1).

3.3.Chemicalenvironment

Dissolved inorganic nutrients showed a clear annual cycle (Fig.4A,B).Nitrogen concentrationswerehighinwinter(150

m

^mol^L¹^), ^decline ⁱⁿ ^spring^(down ^to6

m

^mol^L¹⁾^and^subse-

quently increased again in autumn (Fig. 4A). Thispattern was reversed forphosphorus,withconcentrations becominghighin summer (up to 20

m

^mol^L¹⁾ ^and ^low ⁱⁿ ^winter ^(down ^to

1

m

^mol^L¹^;^Fig.^4B).Consequently,DIN:DIPratiosduringtheA.

ostenfeldiibloomswerelow,withanaveragearound0.5(Fig.4C).

Towardstheendofeachbloom,DIN:DIPratiosincreasedagain.A.

ostenfeldii abundances showed a positive correlation with DIP (

r

⁼^0.63, ^P<0.05), while nocorrelation withDIN and DIN:DIP ratioswasobserved(Table1).

3.4.Biologicalenvironment

Chlorophyll-a concentrations generally showed two distinct peaks,withaphytoplanktonspringbloominallthreeyears,and anotherbloomduringthesummerorautumn(Fig.5).Thespring blooms were dominated by green algae and were followed by diatoms and subsequently dinoﬂagellates (data not shown).

Cyanobacteriawerenotprominentlypresentin2014,butreached high populationdensities in 2013and 2015during spring and autumn.

Table1

Spearman’srankcorrelationcoefﬁcients(r)describingrelationshipsbetweenA.

ostenfeldiiabundances,temperatureandsalinity(**P<0.001,*P<0.05).

A.ostenfeldii Temperature Salinity Temperature 0.74**

Salinity 0.42* 0.40*

Windspeed 0.51** 0.56** 0.38*

Precipitation 0.16 0.13 0.41*

DIN 0.11 0.57** 0.54*

DIP 0.64* 0.85** 0.47*

Copepods 0.38 0.11 0.17

Rotifers 0.17 0.39 0.35

Week 0.43** 0.41** 0.07

Fig.3.SeasonaldynamicsofA)temperature,B)windspeed,C)precipitation,andD)salinity,wheretheblacklinesrepresentthemovingaverage.Thelightgreyareasinthe backgroundofeachgraphindicatetheoccurrenceofanA.ostenfeldiibloomevent.

(5)

Each year, copepods were most abundant prior to the A.

ostenfeldii blooms, reaching total biovolumes of 6.88mm³L¹ (Fig.5).Copepoddensitiesstartedtodeclinerapidlyattheonsetof eachbloomandremainedlowthroughoutthebloomperiod.Other grazers, particularly rotifers, were also abundant in the creek.

Rotifersreachedbiovolumesof6.22mm³L¹priortoeachbloom, andshowedsubsequentsteepdeclinesduringthe2013and2014 blooms(Fig.5).In2015,rotiferdensitiesremainedgenerallylow, withaveragebiovolumes around1.9mm³L¹during thebloom period.AcorrelationbetweenA.ostenfeldiiabundancesandgrazer densitieswasnotfound.

3.5.Toxins

PSP toxins, gymnodimines and spirolides were measured duringthebloomsof2014and2015.Inbothyears,theamount of cell-bound toxins corresponded to A. ostenfeldii population densities,beinghighestatthepeakofthebloomsreachingaverage valuesof37.2and35.9

m

^g^L¹^for^PSP^toxins,^3.0^and^1.8

m

^g^L¹^for

gymnodiminesand3.2and5.1

m

^g^L¹^forspirolides,in2014and 2015,respectively(Fig.6).In2014,toxinswereonlymeasuredfrom June19thuntilAugust28th.

Nomajorchangesintoxincompositionwereobservedbetween and within the years (datanot shown). For PSP toxins, C1/C2 analoguescomprisedmostofthetoxinproﬁle(80%),andrelative contributionsofsaxitoxinandGTX2/3werecomparablewith10%

each.Thecycliciminetoxinproﬁleconsistedforgymnodiminesof gymnodimineA,andforspirolidesof13-desmethylspirolideC.

4.Discussion

Since thedetectionof highabundances of thedinoﬂagellate Alexandriumostenfeldiiin2012(Bursonetal.,2014),annualblooms recurred in the Ouwerkerkse Kreek. Yet, the dynamics of the bloomsvariedstronglybetweenthemonitoredyears(2013–2015).

Our resultsindicatethat thecombinationof physical,chemical, andbiologicalfactorsdeterminestheoccurrenceofA.ostenfeldii blooms,whilephysicalfactorsplayacrucialroleinthemagnitude anddurationofabloom.

Nutrientdynamicswerelargelycomparablebetweenthethree years,withlowDINconcentrationsaround10

m

^mol^L¹^and^high

Fig.4.SeasonaldynamicsinA)DIN,B)DIPandC)DIN:DIPratios.Thelightgrey areasinthebackgroundofeachgraphindicatetheoccurrenceofanA.ostenfeldii bloomevent.

Fig.5. Seasonaldynamicsintheabundancesofcopepods(blackline)androtifers (blackdottedline),andintheconcentrationsofchlorophyll-a(greyline).Thelight greyareasinthebackgroundofeach graphindicatetheoccurrenceofan A.

ostenfeldiibloomevent.

Fig.6.SeasonaldynamicsinPSPtoxins(blackline),spirolides (greyline)and gymnodimines(greydottedline)duringtheA.ostenfeldiibloomsof2014and2015.

ThelightgreyareasinthebackgroundofeachgraphindicatetheoccurrenceofanA.

ostenfeldiibloomevent.

(6)

DIP concentrations around 20

m

^mol^L¹ ^during ^each ^bloom

(Fig.4A,B).DINconcentrationswerelowasaresultofanearlier phytoplanktonspringbloomconsistingprimarilyofgreenalgae, while DIP concentrations werepresumably highdue to anoxic conditionsnearthebottomofthecreekleadingtothereleaseof phosphate from the sediment (Mortimer 1941; Fig. S1;

AppendixA).Suchanoxicconditions,inturn,maybetheresult of increasedsurface temperatures (

r

⁼^0.85, ^P<0.001;Table 1).

Alexandriumostenfeldiiabundancesshowedasigniﬁcantcorrela- tionwithDIPconcentrations(

r

⁼^0.66,^P<0.05;Table1),although thismightbeduethecollinearitybetweenDIPandtemperature, correspondingtosummerconditions. Alexandrium specieshave previouslybeenshowntoproliferateindependentlyofinorganic nutrientconcentrations,alsowhennitrogenconcentrationswere low(Collosetal.,2007;Hakanenetal.,2012;Laanaiaetal.,2013;

Vilaetal.,2005).Despitetheselownitrogenconcentrations,and thegenerallypoorcompetitiveabilityofdinoﬂagellatesintermsof growthandnutrientuptake(Litchmanetal.,2007;Smayda,1997;

Tang, 1996),denseA.ostenfeldiibloomsdidoccureachyear.Several Alexandrium species were shown to utilize organic substrates (Anderson etal.,2012a;Colloset al.,2007,2004;Jacobsonand Anderson, 1996). Such heterotrophic feeding strategies would allow A. ostenfeldii to reach high population densities under nutrientlimitedconditions,andfurthermoregainadvantageover strictlyautotrophicgrowingphytoplankton.Futurestudiesshould includetheassessmentoforganicsubstratestofurtherelucidate theroleofmixotrophyintheformationofHABs.

Top-down control on the A. ostenfeldii population was also largelycomparable betweenthethreeyears.Grazerbiovolumes increasedinspringandsharplydeclinedjustbeforeorattheonset of the A. ostenfeldii bloom (Fig. 5). Grazer biomass was likely supportedbythephytoplanktonspringbloom, andgrazingmay have caused chlorophyll-a concentrations to rapidly decline (Fig.5).Consequently,chlorophyll-aconcentrationswererelative- lylow at theonset of the A.ostenfeldii blooms, and mayhave resulted in food limitation for zooplankton. This, in turn will reducegrazingpressure,whichpossiblyfacilitatedthedevelop- mentoftheA.ostenfeldiiblooms.Alternatively,A.ostenfeldiimay havereducedgrazerabundancesduetofunctionaltraitssuchas toxinsandallelochemicalproduction,whichwereshowntoplaya roleingrazerdeterrence(ColinandDam,2003; Selanderetal., 2006;Sopanenetal.,2011;Teegarden,1999;Tillmannetal.,2007;

TillmannandJohn,2002;Wohlrabetal.,2010).In2014and2015, concentrations of PSP toxins, as well as gymnodimines and spirolides largely followed A. ostenfeldii dynamics and were comparable between both years. This is remarkable since cell densities were substantially differentin 2014and 2015, which implies that cells fromthe populationof2015 containedmore toxins.Theproductionof thesetoxinsas wellasofallelopathic compoundsbythisA.ostenfeldiipopulationmayhavebeneﬁtted bloom development (John et al., 2015; Tillmann et al., 2007;

TillmannandJohn,2002).

Temperatures were generallyabove 15C during the bloom period,rangingupto25C,andmayverywellbearequirementfor bloomdevelopment, as a strongcorrelationwas foundwith A.

ostenfeldiiabundances(

r

⁼^0.74,^P<0.001;Table1).Temperatures of15Cweresufﬁcienttoallowgrowthratesof0.22d¹forvarious isolatesfromtheDutchA.ostenfeldiipopulation(VandeWaaletal., 2015), and higher temperatures may further promote growth.

Alexandrium ostenfeldii isolates from other populations indeed showedhighergrowthratesattemperaturesof20–24C(Billetal., 2016;Kremp etal., 2012).However, temperatureswere largely comparable between years and can therefore not explain the observedvariationsinthemagnitudeanddurationsoftheblooms.

Salinities in the creek showed clear seasonal dynamics, presumablyas a result of variationsin precipitation (

r

⁼0.41,

P<0.05;Table1)andassociateddrainageviaapumpingstation.

HighA.ostenfeldiiabundancescorrespondedtosummerconditions with increased salinities, where highest cell numbers were reachedatsalinitiesabove11(Fig.3C).Recentexperimentswith aDutchA.ostenfeldiistrainshowedabroadsalinitytolerancewith largelycomparable growth ratesover a salinity rangeof 10–35 (Martensetal.,2016).TheirﬁndingssuggestedthatA.ostenfeldiiis a euryhalinespecies, growingbestathigher salinities,which is supported bythe correlationbetweensalinity and A.ostenfeldii populationdensitiesthatwasfound(

r

⁼^0.42,^P<0.05;Table1).

Indeed,populationdensitieswerehighestin2013and2014,when salinities remained above 15 during the bloom period. Lower salinities in 2015 could have contributed to reduced bloom densities.Thesharpdropinsalinityattheendofthe2015bloom period,causedbyexcessiverainfall(Fig.3D)anddrainageviathe pumping station, was likely responsible for the sudden bloom termination.Salinitiesdroppedto3.7,andtheselowvalueswere showntocausemortalityamongseveralA.ostenfeldiistrains(Gu, 2011;JensenandMoestrup,1997;LimandOgata,2005;Maclean etal.,2003;Suikkanenetal.,2013),includingaDutchstrainfrom thesamepopulation(Martensetal.,2016).Similarly,alargedrop insalinityinOctober2013,mayhaveterminatedtheA.ostenfeldii bloominthatparticularyear(Fig.3D).

Higher temperatures are also related toreductions in wind speed (

r

⁼0.56, P<0.001; Table 1) and together both factors contributetoamorestablewatercolumn,whichmaybeneﬁtA.

ostenfeldiibloomformation.Inaddition,A.ostenfeldiiissensitiveto turbulence, and minimal disturbances of the water column by wind mayaffect its growthand bloom development(Berdalet et al., 2007). A negative correlation between A. ostenfeldii abundances and wind speed was indeed observed (

r

⁼0.51, P<0.001; Table1).Speciﬁcally, in2013 and 2014bloomsof A.

ostenfeldiidevelopedwhenwindspeedwaslow.Inthemiddleof thebloomperiodin2014,windspeeddrasticallyincreased,which mayhaveattributedthesubsequentdropincelldensitiesatthe endofJuly.In2015,windspeedwasgenerallynotlowercompared totherestoftheyear.Consequently,thecombinationofrelatively high wind speed and excessive precipitation with resulting reductioninsalinitypresumablyledtothelowerbloomdensities aswellastheshorterbloomperiodin2015.However,atemporary disturbanceinthewatercolumnmaybecrucialforA.ostenfeldii bloominitiation.Oxygenandlightareconsideredtoplayakeyrole incystgermination(Andersonetal.,1987),butvaluesarelownear thesedimentofthecreek(Fig.S1;AppendixA).Thus,inorderto germinate,cystsneed tobere-suspendedin thewatercolumn.

Inﬂuences from tides, upwelling and stream ﬂows are absent withinthecreek,and resuspensionofthesediment istherefore mainlydependedonwindevents.Indeed,higherwindspeedwas observedjustbeforethestartofthebloomduringallthreeyears (Fig.3B).

Overall,changesin thephysicalenvironmentovertheyears, particularly wind speed and salinity, may largely explain the observeddifferencesinmagnitudeanddurationoftheA.ostenfeldii blooms.Variouslaboratoryexperimentsshowedthattemperature, salinity and turbulence can substantially affect Alexandrium growth(Billetal.,2016;Grzebyketal.,2003;JensenandMoestrup, 1997;Juhletal.,2001;Laabiretal.,2011;LimandOgata,2005;

White,1976),andtherebyformimportantconstraintsonbloom development.Ourresultsindeedindicatethatthemostbeneﬁcial combinationofphysicalfactorsforA.ostenfeldiibloomsincludes higher temperatures and salinities (above 15C and salinities above10),andcalmandstableweatherconditionswithreduced wind speed,and indirectlyprecipitation.Theseconditionswere observedin2013,whichalsocorrespondedtohighestpopulation densitiesreachingupto4500cellsmL¹.Severalﬁeldstudiesalso foundthatphysicalfactors,whichpromotewatercolumnstability,

(7)

playacrucialroleinthedevelopmentofAlexandriumblooms.For instance, a correlation between higher temperatures and A.

ostenfeldii bloom development was found in the Baltic Sea (Hakanen et al., 2012).Similarly,both salinityand wind speed werestronglyrelatedtoAlexandriumbloomdevelopment inthe EstuaryandGulfofSt.Lawrence,Canada(Weiseetal.,2002).In addition,astudyfromThauLagoonfoundthatparticularlywind speed and temperature play an important role in Alexandrium bloomdevelopment,wherewindeventsallowedbloominitiation byresuspensionofcysts,andlowwindspeedincombinationwith highand stable temperatures facilitated subsequent vegetative growth(Laanaiaetal.,2013).

Even though we only studied one system, our results do indicatehowthephysicalenvironmentmayshapeHABs.Thisnot onlysupports ourcurrent understanding, but also provides an exampleofhowclimate-drivenchangesinthephysicalenviron- mentmayfavorHABdevelopment.Earlierstudieshavesuggested thatglobalrisingtemperaturesmayleadtoarangeexpansionof HAB species (Thomas et al., 2012; Wells et al., 2015). Other consequences of global temperature increases are lower wind speeds (McVicar et al., 2012), which have already decreased globallyby5–15%overthelastdecades(Vautardetal.,2010).Such lowwindspeeds incombinationwithhighertemperatureswill enhancewatercolumnstability,whichisgenerallyfavorablefor HAB development (Hallegraeff, 2010; Hallegraeff et al., 1995).

Furthermore,saltwaterintrusionisbecominganincreasingglobal threatthroughsealevelrise(WernerandSimmons, 2009).This allowseuryhaline HABspecies,suchasA.ostenfeldiitodisperse intothesewatersimposinganincreasinghealthriskparticularlyto densely populated areas such as the Netherlands. Our results suggestthatpredictedchangesinthephysicalenvironmentmay enhance bloom development in future coastal waters and embayments.

Acknowledgements

The authors thank Dennis Waasdorp, Nico Helmsing, Erik Reichmanand SuzanneNaus-Wiezer fortheirassistanceduring theﬁeld workand technical support. We alsothank Annegret MüllerforanalysesofPSPtoxins.Furthermore,wearegratefulto WaterschapScheldestromenforcollectingasubstantialpartofthe dataoverthepastyearsandprovidingittous.Wewouldliketo thankErnstdeBokxfromBrachion-Ideeforzooplanktoncounts and biovolume assessments. The work of KBis funded by the Gieskes-StrijbisFoundation.FinancialsupportforBKandUJwas providedbythePACESresearchprogramoftheAlfred-Wegener- Institute,Helmholtz-ZentrumfürPolar-undMeeresforschung,and bytheGerman ResearchFoundation (DFG) PriorityProgramme DynaTrait(SPP1704;Jo702/7-1)forUJandSW.

AppendixA.Supplementarydata

Supplementarydataassociatedwiththisarticlecanbefound,in theonlineversion,athttp://dx.doi.org/10.1016/j.hal.2017.02.004.

References

Anderson,D.M.,Keafer,B.A.,1987.Anendogenousannualclockinthetoxicmarine dinoﬂagellateGonyaulaxtamarensis.Nature325,616–617.

Anderson,D.M.,Taylor,C.D.,Armbrust,E.V.,1987.Theeffectsofdarknessand anaerobiosisondinoﬂagellatecystgermination.Limnol.Oceanogr.32,340–351.

Anderson, D.M., Kulis, D.M., Sullivan,J.J., Hall, S.,Lee, C.,1990.Dynamics of physiologyofsaxitoxinproductionbythedinoﬂagellatesAlexandriumspp.Mar.

Biol.104,511–524.

Anderson,D.M.,Glibert,P.M.,Burkholder,J.M.,2002.Harmfulalgalbloomsand eutrophicationnutrientsources,composition,andconsequences.Estuaries25, 704–726.

Anderson,D.M.,Stock,C.A.,Keafer,B.A.,BronzinoNelson,A.,McGillicuddy,D.J., Keller,M.,Thompson,B.,Matrai,P.A.,Martin,J.,2005.Alexandriumfundyense cystdynamicsintheGulfofMaine.Deep.Res.PartIITop.Stud.Oceanogr.52, 2522–2542.

Anderson,D.M.,Alpermann,T.J.,Cembella,A.D.,Collos,Y.,Masseret,E.,Montresor, M.,2012a.ThegloballydistributedgenusAlexandrium:multifacetedrolesin marineecosystemsandimpactsonhumanhealth.HarmfulAlgae14,10–35.

Anderson,D.M.,Cembella,A.D.,Hallegraeff,G.M.,2012b.Progressinunderstanding harmfulalgalblooms:paradigm shiftsandnewtechnologies forresearch monitoring,andmanagement.Ann.Rev.Mar.Sci.4,143–176.

Balech,1995.TheGenusAlexandriumHalim(Dinoﬂagellata).SherkinIslandMarine StationPublication,SherkinIsland,CountyCork,Ireland.

Berdalet,E.,Estrada,M.,2005.Effectsofsmall-scaleturbulenceonthephysiological functioningofmarinealgae.In:Durvasula,S.R.(Ed.),AlgalCultures,Analogues andApplications.SciencePublishers,Ensﬁeld,NewHampshireUSA,pp.459–

500.

Berdalet,E.,Peters,F.,Koumandou,V.L.,Roldan,C.,Guadayol,O.,Estrada,M.,2007.

Species-specificphysiologicalresponseofdinoflagellatestoquantifiedsmall- scaleturbulence.J.Phycol.43,965–977.

Berman,T.,Shteinman,B.,1998.Phytoplanktondevelopmentandturbulentmixing inLakeKinneret(1992–1996).J.PlanktonRes.20,709–726.

Bill, B.D.,Moore,S.K., Hay,L.R.,Anderson, D.M.,Trainer,V.L.,2016. Effects of temperatureandsalinityonthegrowthofAlexandrium(Dinophyceae)isolates fromtheSalishSea.JounalPhycol.52,230–238.

Borkman,D.G.,Smayda,T.J.,Tomas,C.R.,York,R.,Strangman,W.,Wright,J.L.C.,2012.

ToxicAlexandriumperuvianum(Balechanddemendiola)balechandtangenin narragansettbay,rhodeisland(USA).HarmfulAlgae19,92–100.

Burson,A.,Matthijs,H.C.P.,deBruijne,W.,Talens,R.,Hoogenboom,R.,Gerssen,A., Visser,P.M., Stomp,M.,Steur, K.,Van Scheppingen,Y., Huisman,J.,2014.

TerminationofatoxicAlexandriumbloomwithhydrogenperoxide.Harmful Algae31,125–135.

Cembella, A.D., Lewis, N.I., Quilliam, M.A., 2000. The marine dinoﬂagellate Alexandriumostenfeldii(Dinophyceae)asthecausativeorganismofspirolide shellﬁshtoxins.Phycologia39,67–74.

StandardMethodsforExaminationofWaterandWastewater, In:Clesceri,L.S., Greenberg,A.E., Andrew,EatonD. (Eds.),20thed. Americanpublichealth association,WashingtonD.C.,USA.

Colin, S.P., Dam, H.G., 2003. Effects of the toxic dinoﬂagellate Alexandrium fundyenseonthecopepodAcartiahudsonica:atestofthemechanismsthat reduceingestionrates.Mar.Ecol.Prog.Ser.248,55–65.

Collos,Y.,Gagne,C.,Laabir,M.,Vaquer,A.,Cecchi,P.,Souchu,P.,2004.Nitrogenous nutritionofAlexandriumcatenella(Dinophyceae)inculturesandinthaulagoon, southernFrance.J.Phycol.40,96–103.

Collos,Y.,Vaquer,A.,Laabir,M.,Abadie,E.,Laugier,T.,Pastoureaud,A.,Souchu,P., 2007. Contributionof several nitrogen sourcesto growth ofAlexandrium catenelladuringbloomsinThaulagoon,southernFrance.HarmfulAlgae6,781–

789.

Dale,B.,1983.Dinoﬂagellaterestingcysts:benthicplankton.In:Fryxell,G.A.(Ed.), SurvivalStrategiesoftheAlgae.CambirdgeUniveristyPress,Cambridge,UK,pp.

69–136.

Fraga,S.,Sanchez,F.,1985.Toxicandpotentiallytoxicdinoﬂagellatesfoundin GalacianRias(NWSpain).ToxicDinoﬂagellates.ElsevierSciencePublishingCo., NewYork.USA,pp.51–54.

Glibert,P.M.,Magnien,R.,Lomas,M.W.,Alexander,J.,Fan,C.,Haramoto,E.,Trice,M., Kana,T.M.,2001.HarmfulalgalbloomsintheChesapeakeandcoastalbaysof Maryland,USA:Comparisonof1997,1998,and1999events.Estuaries24,875–

883.

Gribble,K.E.,Keafer,B.A.,Quilliam,M.A.,Cembella,A.D.,Kulis,D.M.,Manahan,A., Anderson, D.M., 2005. Distribution and toxicity of Alexandrium ostenfeldii (Dinophyceae)intheGulfofMaine,USA.Deep.Res.PartIITop.Stud.Oceanogr.

52,2745–2763.

Grzebyk,D.,Bechemin,C.,Ward,C.J.,Verite,C.,Codd,G.A.,Maestrini,S.Y.,2003.

Effectsofsalinityandtwocoastalwatersonthegrowthandtoxincontentofthe dinoﬂagellateAlexandriumminutum.J.PlanktonRes.25,1185–1199.

Gu,H.,2011.Morphology,phylogeneticposition,andecophysiologyofAlexandrium ostenfeldii(Dinophyceae)fromtheBohaiSea,China.J.Syst.Evol.49,606–616.

Hakanen,P.,Suikkanen,S.,Franzen,J.,Franzen,H.,Kankaanpaa,H.,Kremp,A.,2012.

BloomandtoxindynamicsofAlexandriumostenfeldiiinashallowembaymentat theSWcoastofFinland,northernBalticSea.HarmfulAlgae15,91–99.

Hallegraeff,G.M.,Mccausland,M.A., Brown,R.K.,1995.Early warningoftoxic dinoﬂagellatebloomsof Gymnodinium catenation in southernTasmanian waters.J.PlanktonRes.17,1163–1176.

Hallegraeff,G.M.,1993.Areviewofharmfulalgalbloomsandtheirapparentglobal increase.Phycologia32,79–99.

Hallegraeff, G.M., 2010. Ocean climate change, phytoplankton community responses,andharmfulalgalblooms: aformidablepredictivechallenge.J.

Phycol.46,220–235.

Harju,K.,Koskela,H.,Kremp,A.,Suikkanen,S.,DeLaIglesia,P.,Miles,C.O.,Krock,B., Vanninen, P., 2016. Identiﬁcation of gymnodimine D and presence of gymnodiminevariantsinthedinoﬂagellateAlexandriumostenfeldiifromthe BalticSea.Toxicon112,68–76.

Heisler,J.,Glibert,P.M.,Burkholder,J.M.,Anderson,D.M.,Cochlan,W.,Dennison,W.

C.,Dortch,Q.,Gobler,C.J.,Heila,C.,Humphries,E.,Lewitus,A.,Magnien,R., Marshall,H.G.,Sellner,K.,Stockwell,D.a.,Stoecker,D.K.,Suddleson,M.,2008.

Eutrophicationandharmfulalgalblooms:ascientiﬁcconsensus.HarmfulAlgae 8,3–13.