Ocean acidification changes the structure of an Antarctic coastal protistan community

32 Pages·2017·1.23 MB·English

BiogeosciencesDiscuss.,https://doi.org/10.5194/bg-2017-224
ManuscriptunderreviewforjournalBiogeosciences
Discussionstarted:7June2017
c Author(s)2017.CCBY3.0License.
(cid:13)
Ocean acidification changes the structure of an Antarctic coastal
protistan community
AlyceM.Hancock1,2,3,AndrewT.Davidson3,4,JohnMcKinlay4,AndrewMcMinn1,2,3,KaiSchulz5,and
RickL.vandenEnden4
1InstituteofMarineandAntarcticStudies,UniversityofTasmania20CastrayEsplanade,BatteryPointTASAustralia7004
2AntarcticGatewayPartnership,20CastrayEsplanade,BatteryPointTASAustralia7004
3AntarcticClimate&EcosystemsCooperativeResearchCentre20CastrayEsplanade,BatteryPointTASAustralia7004
4AustralianAntarcticDivision203ChannelHwy,KingstonTASAustralia7050
5CentreforCoastalBiogeochemistryResearch,SouthernCrossUniversity
Correspondenceto:AlyceHancock([email protected])
Abstract.Antarcticnear-shorewatersareamongstofthemostvulnerableintheworldtooceanacidification.Microbesoccupy-
ingthesewatersarecriticaldriversofecosystemproductivity,elementalcyclingandoceanbiogeochemistry,yetlittleisknown
abouttheirsensitivitytooceanacidification.Anunreplicated,six-leveldose-responseexperimentwasconductedusing650L
incubationtanks(minicosms)adjustedtofugacityofcarbondioxide(ƒCO )from343to11641µatm.Theminicosmswere
2
5 filledwithnear-shorewaterfromPrydzBay,EastAntarcticaandtheprotistancompositionandabundancewasdeterminedby
microscopy analysis of samples collected during the 18 day incubation. No CO -related change in the protistan community
2
composition was observed during the initial 8 day acclimation period under low light. Thereafter, the response of protists to
ƒCO werespecies-specificforbothheterotrophicandautotrophicprotists.Theresponsebydiatomswasrelatedtocellsize,
2
large cells increasing in abundance with low to moderate ƒCO (634-953 µatm). Similarly, the abundance of Phaeocystis
2
10 antarctica increased with increasing ƒCO2 peaking at a ƒCO2 of 634 µatm. Above this threshold the abundances of large
diatomsandPhaeocystisantarcticafelldramatically,andsmalldiatomsdominated,thereforeculminatinginasignificantshift
inthecompositionoftheprotistancommunity.ThethresholdCO levelatwhichthecompositionchangedagreedwiththat
2
previouslymeasuredatthislocation,indicatingitremainsconsistentamongseasons.Thissuggeststhatnear-shoremicrobial
communitiesarelikelytochangesignificantlyneartheendofthiscenturyifanthropogenicCO releasecontinuesunabated,
2
15 withprofoundramificationsfornear-shoreAntarcticecosystems.
Copyrightstatement.
1 Introduction
Eukaryotic and prokaryotic microbes are the most abundant organisms in the oceans and comprise the base of all marine
foodwebs(Kirchman,2008;CooleyandDoney,2009;Doneyetal.,2012).Theircompositionandabundancedeterminesthe
1
BiogeosciencesDiscuss.,https://doi.org/10.5194/bg-2017-224
ManuscriptunderreviewforjournalBiogeosciences
Discussionstarted:7June2017
c Author(s)2017.CCBY3.0License.
(cid:13)
qualityandquantityoffoodavailabletohighertrophiclevels,affectingfisheryproductivityandtheconservationofbiological
diversity(CooleyandDoney,2009;Doneyetal.,2012).IntheSouthernOcean,microbesaredriversofproductivity,elemental
cyclesandoceanbiogeochemistry,meaningtheirresponsetoenvironmentalstressorsisakeydeterminantofSouthernOcean
feedbackstoglobalclimatechange(ArrigoandThomas,2004;Arrigoetal.,2008;Kirchman,2008).Despitetheirimportance,
5 relativelylittleisknownaboutthesensitivityofAntarcticamarinemicrobestooceanacidificationandthislimitsourabilityto
predicthowtheSouthernOceanwillbeimpactedinthefuture,andthefeedbackthismayhaveonglobalclimatechange.
TheSouthernOceanisparticularlyvulnerabletooceanacidificationduetoitscoldtemperatures,extensiveupwellingand
naturally large seasonal fluctuations in pH. Carbon dioxide (CO ) has a higher solubility at cold temperatures so that more
2
CO isbeingabsorbedinpolarwaterscomparedtothatinwarmerwaters.Addedtothis,thesurfacewatersoftheSouthern
2
10 Ocean are being exposed to increased CO2 from upwelling bringing deep CO2 rich waters to the surface (Orr et al., 2005).
Thisisenhancedinnear-shoreAntarcticawhereupwellingisgreater(McNeiletal.,2010;IPCC,2011).Thesewatersalready
experience large seasonal fluctuations in pH, upon which the anthropogenic decrease is superimposed. Prydz Bay, off Davis
Station East Antarctica, shows an annual cycle of CO levels ranging from as high as 450 µatm during autumn and winter
2
to below 100 µatm in summer (Gibson and Trull, 1999; Roden et al., 2013). During autumn and winter, sea-ice covers the
15 ocean,lightislimitingtophotosynthesis,thedraw-downofCO2byprimaryproductionandair-seagasexchangeisnegligible.
During spring and summer the sea-ice retreats and phytoplankton bloom, due to the increased light and nutrient availability
(Gibson and Trull, 1999; Roden et al., 2013). Thus, phytoplankton in these coastal Antarctic waters are exposed to highly
variablecarbonatechemistryovertheannualcycle.
TherehavebeenrelativelyfewstudiesontheresponseofmicrobialcommunitiestoincreasedCO intheSouthernOcean
2
20 andAntarcticwaters,andoutofthesefewerhavebeenatthecommunitylevel(Tortelletal.,2008;Fengetal.,2010;McMinn
etal.,2014;Coadetal.,2016;Davidsonetal.,2016;Tarlingetal.,2016;Thomsonetal.,2016;McMinnetal.,2017).These
studiesreportarangeofmicrobialresponsestoincreasedCO ,butonecommontrendwasashiftincommunitycomposition.
2
With increased CO Tortell et al. (2008) and Feng et al. (2010) found an increase in the dominance of Chaetoceros species
2
oversmaller,pennatediatomssuchasPseudonitzschiasubcurvata.Tortelletal.(2008)alsofoundanincreaseinPhaeocystis
25 antarcticaabundanceat800µatm,althoughFengetal.(2010)foundnoeffectonthesamespecies.Davidsonetal.(2016)and
Thomsonetal.(2016)ranaminicosmstudyatthesamelocationasthisstudy(DavisStation,EastAntarctica),andfounda
significantchangeinspeciesabundanceandbiomasswithincreasesinCO above643to1281µatm(Davidsonetal.,2016;
2
Thomsonetal.,2016).Belowthisthresholdthecommunityhadahighabundanceoflargediatoms,butabovewasdominated
by nano- and picoplankton and Fragilariopsis curta/cylindrus <10 µm long. This is consistent with microbial communities
30 studiedelsewhereintheworld,particularlyintheArctic,wherestudieshaveobservedashifttopico-andnanoplanktonathigh
CO (Hareetal.,2007;Brussaardetal.,2013).Schulzetal.(2017)recentlyreviewed31communitylevelstudies,findingan
2
increaseinpicoeukaryotesathighCO inmoststudies,particularlyprasinophytesandchlorophytes.Theeffectsofincreased
2
CO onlargermarinediatomswaslessclearwithevidenceforbothpromotionandinhibition.Schulzetal.(2017)concluded
2
that the effects on marine diatoms are likely to be at a species level rather than the community level, and therefore could be
35 moredifficulttodetect.
2
BiogeosciencesDiscuss.,https://doi.org/10.5194/bg-2017-224
ManuscriptunderreviewforjournalBiogeosciences
Discussionstarted:7June2017
c Author(s)2017.CCBY3.0License.
(cid:13)
Incubationsofnaturalcommunities,whichincludetheeffectsofspeciesinteractionsandcompetition,areessentialtopredict
theeffectsofoceanacidificationonmicrobialcommunities(Schulzetal.,2017).Thisstudywillassessthefollowingquestions
onanaturalmicrobialcommunityfromnear-shoreEastAntarcticwaters.
1. DoindividualspecieshavedifferenttolerancestoincreasedCO level?
2
5 2. DoestheprotistancommunitycompositionandabundancechangewithincreasedlevelsofCO2?AndwhatCO2 level
elicitsthischange?
3. DoesanacclimationperiodallowtheprotistancommunitytobettertolerateexposuretoelevatedCO ?
2
4. WhencomparedwithDavidsonetal.(2016)andThomsonetal.(2016),doesourexperimentindicatethattheresponse
by the protistan community at this site is consistent in nature and threshold, irrespective of seasonal and interannual
10 differencesinthecompositionofthecommunityandtheavailabilityofnutrients?
2 Methods
An unreplicated, six-level dose-response experiment was conducted on a natural microbial community over a range of CO
2
levels (343, 506, 634, 953, 1140 and 1641 µatm). Seawater was collected on the 19th November 2014 approximately 1 km
offshorefromDavisStation,Antarctica(68 C35’S,77 C58’E)fromanareaofice-freewateramongstbrokenfast-ice.The
◦ ◦
15 seawaterwascollectedusingathoroughlyrinsed720LBambibucketslungbeneathahelicopterandtransferredintoa7000L
polypropalenereservoirtank.Six650Lpolyethenetanks(minicosms),locatedinatemperature-controlledshippingcontainer,
wereimmediatelyfilledviateflonlinedhoseviagravitywithanin-line200µmArkalfiltertoexcludemetazooplankton.The
minicosmsweresimultaneouslyfilledtoensuretheycontainedthesamestartingcommunity.Theambientwatertemperature
attimeofcollectionwas-1.0 Candtheminicosmsweremaintainedatatemperatureof0 C 0.5 C.Atthecentreofeach
◦ ◦ ◦
±
20 minicosmtherewasanaugershieldedformuchofitslengthbyatubeofpolythene.Thisaugerwasrotatedat15rpmtogently
mixthecontentsofthetanks.Eachminicosmtankwascoveredwithanacrylicair-tightlidtopreventCO off-gassingoutside
2
oftheminicosmheadspace.Foramoredetaileddescriptionoftheminicosmset-upseeDavidsonetal.(2016).
Theminicosmexperimentwasconductedbetweenthe19thNovemberandthe7thDecember2014.Initially,thecontentsof
thetanksweregivenadaytoequilibratetotheminicosms.Thiswasfollowedbyafivedayacclimationperiodtoincreasing
25 CO2atlowlight(0.8 0.2µmolm−2s−1),allowingcellphysiologytoacclimatedtotheCO2increase(days1-5).Duringthis
±
periodtheCO wasprogressivelyadjustedoverfivedaystothetargetlevelforeachtank(343-1641µatm).ThereafterCO
2 2
wasadjusteddailytomaintaintheCO levelineachtreatment(seecarbonatechemistrysectionbelow).Followingacclimation
2
tothevariousCO treatments,lightwasprogressivelyadjustedto89 16µmolm 2s 1 ata19hlight:5hdarkcycle.The
2 − −
±
communitywasincubatedandallowedtogrowforafurther10days(days8-18)withCO adjustedbacktotargeteachday(see
2
30 carbonate chemistry section below). For a detailed description of lighting apparatus and climates see Davidson et al. (2016)
andDeppeleretal.(submitted).
3
BiogeosciencesDiscuss.,https://doi.org/10.5194/bg-2017-224
ManuscriptunderreviewforjournalBiogeosciences
Discussionstarted:7June2017
c Author(s)2017.CCBY3.0License.
(cid:13)
2.1 Carbonatechemistrymeasurementsandcalculations
Carbonatechemistrywascalculatedfromdissolvedinorganiccarbon(DIC),pH (totalprotonscale),salinityandtemperature.
T
MinicosmfugacityofCO (ƒCO )wasadjustedincrementallyduringtheacclimationphasetothetargetlevelforeachtank,
2 2
andthenfortheremainderoftheexperimentƒCO wasadjusteddailytomaintainthetargetlevelswithinthetanks.Dailyad-
2
5 justmentsweremadebasedonmorningandafternoonpHmeasurements,andwereachievedbytheaddingtherequiredvolume
of0.2µmfilteredCO saturatedseawater.Afteradjustment,fullcarbonatechemistrymeasurementswereperformed(DIC,pH,
2
salinityandtemperature)tochecktheadjustment,andifrequiredasecondaryadjustmentwasmade.DailypHmeasurements
wereconductedusingaportable,NBS-calibratedprobe(MettlerToledo)andsalinitymeasurementsviaaconductivitymetre
WTW197. DIC was measured by infra-red absorption on an Apollo SciTech AS-C3 analyzer equipped with a LICOR7000
10 andpHT withaGBCUV-Vis916spectrophotometerinatencentimetrethermostatedcuvettteandpHindicatordye(m-cresol
purple).Allsamplesforthecarbonatechemistrywerecollectedin500mLglassbottleswithstoppersfollowingtheDickson
etal.(2007)guidelines.FullcarbonatechemistrymeasurementsandcalculationsaredescribedinDeppeleretal.(submitted).
2.2 Microbialcommunitystructure
Samplesfromeachminicosmwerecollectedondays1,3,5,8,10,12,14,16and18formicroscopicanalysistodetermine
15 protistanidentityandabundance.Approximately960mLwerecollectedfromeachtank,oneachday.Sampleswerefixedwith
40mLofLugol’siodineandallowedtosedimentoutat4 Cfor 4days.Oncecellshadsettledthesupernatantwasgently
◦
≥
aspiratedtillapproximately200mLremained.Thiswastransferredtoa250mLmeasuringcylinder,againallowedtosettle
(asabove),andthesupernatantgentlyaspirated.Theremaining~20mLwastransferredintoa30mLamberglassbottle.
Afurther1Lwastakenondays0,6,13and18foranalysisbyFieldEmissionScanningElectronMicroscope(FESEM).
20 Thesesampleswereconcentratedto5mLbyfiltrationovera0.8µmpolycarbonatefilter.Cellswereresuspended,theconcen-
tratetransferredtoaglassvialandfixedtoafinalconcentrationof1%EM-gradegluteraldehyde(ProSciTechPtyLtd).
Allsampleswerestoredandtransportedat4 CtotheAustralianAntarcticDivision,Hobart,Australiaforanalysis.
◦
2.2.1 Electronmicroscopy
Gluteraldehyde-fixedsampleswerepreparedforFESEMimagingusingamodifiedpolylysinetechnique(MarchantandThomas,
25 1983).Inbrief,afewdropsofgluteraldehyde-fixedsamplewereplacedonpolylysinecoatedcoverslipsandpost-fixedwith
OsO (4%)vapourfor30min,allowingcellstosettleontothecoverslips.Thecoverslipswerethenrinsedindistilledwater
4
anddehydratedthroughagradedethanolseriesendingwithemersionin100%dryacetonebeforebeingcriticallypointdried
inaTousimisAutosamdri-815CriticalPointDrier.Thecoverslipsweremountedonto12.5mmdiameteraluminiumstubsand
sputter-coatedwith7nmofplatinum/palladiuminaCressington208HRDcoater.ImagingofstubswasconductedbyJEOL
30 JSM6701FFESEMandprotistsidentifiedusingScottandMarchant(2005).
4
BiogeosciencesDiscuss.,https://doi.org/10.5194/bg-2017-224
ManuscriptunderreviewforjournalBiogeosciences
Discussionstarted:7June2017
c Author(s)2017.CCBY3.0License.
(cid:13)
2.2.2 Lightmicroscopy
Lugols-fixedandsedimentedsampleswereanalysedbylightmicroscopybetweenJuly2015andFebruary2017.Between2to
10mL(dependingoncell-density)oflugols-concentratedsampleswasplacedintoa10mLUtermohlcylinder(Hydro-Bios,
Keil)andthecellsallowedtosettleovernight.Duetothelargevariationinsizeandtaxa,astratifiedcountingprocedurewas
5 employedtoensurebothaccurateidentificationofsmallcellsandrepresentativecountsoflargercells.Allcellsgreaterthan20
µmwereidentifiedandcountedat20xmagnification;thoselessthan20µmat40xmagnification.Forlargercells(>20µm),
20randomlychosenfieldsofview(FOV)at3.66x106µm2 werecountedtogainanaveragecellsperL.Forsmallercells(<20
µm), 20 randomly chosen FOVs at 2.51x105µm2 were counted. Counts were conducted on an Olympus IX 81 microscope
with Nomarski interference optics. Identifications were determined using Scott and Marchant (2005) and FESEM images.
10 Autotrophic protists were distinguished from heterotrophs via the presence of chloroplasts and based on their taxonomic
identity.
2.3 Statisticalanalysis
The minicosm experiment was a six dose, unreplicated experiment based on a repeated measures design. Due to the lack of
replication,noformalstatisticscouldbeundertakenontheinteractionsbetweentimeandƒCO treatment.Temporalchanges
2
15 inspeciesabundancesbetweentreatmentgroupswereinformallyassessedbyplottingthemeanmicrobialabundanceateach
timeforeachtreatment.MeansandstandarderrorswerecalculatedfromseparateFOVcounts;asthesearesub-samplesare
fromasingletreatment,theyshouldbeconsideredpseudo-replicatesandareindicativeoftreatment-levelsamplingvariability.
Abundant taxa with low variance were examined separately, but rare taxa with high variance were combined into functional
groups (see Table 1 for abbreviations). In plots, data points from the different ƒCO treatments were slightly offset at each
2
20 sampletimetoavoidover-plottingofthedata.ClusteranalysesandordinationswereperformedonBray-Curtisresemblance
matrixesformedfromsquare-roottransformedabundancedata.Thistransformationwasassessedasappropriateforreducing
the influence of abundant species, as judged from a one-to-one relationship between observed dissimilarities and ordination
distances(ie.Sheparddiagram,notshown).TheBray-Curtismetricwasusedasitisrecommendedforecologicaldataduetoits
treatmentofjointabsences(ie.thesedonotcontributetowardssimilarity),andgivingmoreweighttoabundanttaxaratherthan
25 raretaxa(BrayandCurtis,1957).Thedatadays1to8andthendays8to18wereanalysedseparatelytodistinguishcommunity
structureintheacclimationperiodandintheexponentialgrowthphaseduringtheincubationperiodoftheexperiment.
Hierarchical agglomerative cluster analyses, based on the Bray-Curtis resemblance matrix, were performed using group-
averagelinkage.SignificantlydifferentclustersofsamplesweredeterminedusingSIMPROF(similarityprofilepermutations
method) (Clarke et al., 2008) with an alpha value of 0.05 and based on 1000 permutations. An unconstrained ordination by
30 non-metricmultidimensionalscaling(nMDS)wasperformedontheresemblancematrixwithaprimary(‘weak’)treatmentof
ties(Kruskal,1964a,b).Thiswasrepeatedover50randomstartstoensureagloballyoptimalsolutionaccordingtoPeres-Neto
and Jackson (2001). Clusters are displayed in the nMDS using colour. Weighted average of sample scores are shown in the
5
BiogeosciencesDiscuss.,https://doi.org/10.5194/bg-2017-224
ManuscriptunderreviewforjournalBiogeosciences
Discussionstarted:7June2017
c Author(s)2017.CCBY3.0License.
(cid:13)
nMDStoshowtheapproximatecontributionofeachspeciestoeachsample.Theassumptionofalineartrendforpredictors
withintheordinationwascheckedforeachcovariate,andinallinstanceswasfoundtobejustified.
A constrained canonical analysis of principal coordinates (CAP) was conducted according to the Oksanen et al. (2017)
protocol using the Bray-Curtis resemblance matrix. This analysis was used to assess the significance of the environmental
5 covariates,orconstraints,indeterminingthemicrobialcommunitystructure.UnlikethenMDSordination,theCAPanalysis
usestheresemblancematrixtopartitionthetotalvarianceinthecommunitycompositionintounconstrainedandconstrained
components,withthelattercomprisingonlythevariationthatcanbeattributedtotheconstrainingvariables,ƒCO ,Si,Pand
2
NO .Randomreassignmentofsampleresemblancewasperformedover199permutationstocomputethepseudo-F statistic
x
asameasureofsignificanceofeachenvironmentalconstraintinthestructuralchangeofthemicrobialcommunity(Legendre
10 and Anderson, 1999). A forward selection strategy was used to choose a minimum subset of significant constraints that still
accountforthemajorityofthevariationwithinthemicrobialcommunity(Legendreetal.,2011).
AllanalyseswereperformedusingRv1.0.136(RCoreTeam,2016)andtheadd-onpackageveganv2.4-2(Oksanenetal.,
2017).
3 Results
15 3.1 Protistancommunityoverview
Thestartingmicrobialcommunitywascharacteristicofapostsea-icebreak-outcommunityinthenear-shoreseawaterofPrydz
Bay.Itwashighlydiversewithover100speciespresent,rangingfromsmallflagellates(<2µm)tolargediatoms(>100µm).
Thetotalprotistanabundanceatthebeginningoftheexperimentwasquitelow(approximately3x105cellsL 1),butincreased
−
tobetween6.4x106and1.9x107cellsL 1dependingonthetreatment.Abundancesremainedlowduringtheacclimationperiod
−
20 (days1to8)thenincreasedexponentiallyfromdays10to16inalltreatmentsexceptthehighestƒCO2 treatment(Figure1).
Atday18therewasadecreaseinabundanceinalltreatmentsexcept635and1641µatm.Thehighvarianceinthe635µatm
treatmentonday18islikelyaduetotheincreaseinafewrarelargediatomspecies,whichwerehighlyvariable.Thehighest
ƒCO treatmentshowedadifferenttrajectorytotheothertanks,withlowabundancemaintaineduntilday12,afterwhichcell
2
numbersincreasedexponentially;hereafterthisisreferredtoasalagingrowth.
25 3.2 SpeciesspecificƒCO2tolerances
3.2.1 Diatoms
DiatomsdominatedthemicrobialcommunityandshowedmarkedresponsestoincreasedƒCO levels.Theresponseofdiatoms
2
was size related, with small diatoms (<20 µm) showing little effect of exposure to higher ƒCO , while larger diatoms (>20
2
µm)increasedinabundanceatmoderatelevelsƒCO ( 634µatm)butdeclinedathigherƒCO .Thistrendwasparticularly
2 2
≤
30 evident in discoid centric diatoms (centric diatoms of the genera Thalassiosira, Landeria and Stellarima), which showed
decreasedƒCO tolerancewithincreasingsize.Thesmallestdiscoidtaxa,anunidentified1to2µmdiametercentricdiatom,
2
6
BiogeosciencesDiscuss.,https://doi.org/10.5194/bg-2017-224
ManuscriptunderreviewforjournalBiogeosciences
Discussionstarted:7June2017
c Author(s)2017.CCBY3.0License.
(cid:13)
showednosignificantresponsetoincreasingƒCO (Figure2a).Thalassiosiraantarctica,whichrangedinsizefrom10to20
2
µmdiameter,alsoshowednoeffectofincreasedƒCO untilthehighesttreatmentlevelof1641µatm(Figure2b).Landeria
2
annulata,amediumsizedcellof30to60µmdiameter,hadanincreasesinabundanceinlowtomoderateƒCO treatments
2
(343-634µatm)afterwhichabundancesdecreasedbutnotlowerthanthatofthecontroltreatment(Figure2c).Thetwolarger
5 dominantcentricspecies,Stellarimamicrotrias(40to80µmdiameter)andThalassiosiraritscheri(>50µmdiameter)showed
an increase in abundance with moderate increases in ƒCO ( 634 µatm), but they showed a decrease in abundance at the
2
≤
three higher ƒCO levels (Figure 2d and e respectively). Whilst these two larger species had a similar response, there were
2
subtledifferencesthatcorrelatewithdifferencesintheaveragesizeofthespecies;Stellarimawasslightlymoretolerantthan
T.ritscheri(averagesizesofStellarimaislessthanT.ritscheri).
10 Asimilarsize-relatedresponsewasobservedinFragilariopsiscells(mostlyF.cylindrusbutincludedoccasionalF.curtaand
F.kerguelensis).Fragilariopsiswasthedominantdiatom,comprisingbetween15to50%ofthetotalphytoplanktonabundance
and ranged in length from 2 µm to >50 µm. Toward the end of the incubation (days 16 and 18), the abundance of small
Fragilariopsiscells(<20µm)washigherintreatmentsexposedtomoderatelyenhancedƒCO (506-953µatm)comparedto
2
that of the control. Abundances in the ƒCO treatments >953 µatm were lower but less than those in the control treatment
2
15 (Figure3a).Incontrast,largercellsofFragilariopsis(>20µm),showedalowertolerancewithabundanceshighestinthethree
lowerƒCO treatmentsandconsiderablylowerabundancesinthetwohighestƒCO treatments.Theabundanceoflargecells
2 2
inthe953µmtreatmentfellbetweenthesetwoextremes(Figure3b).
Other diatoms showed similar responses with larger species having higher abundances in ƒCO treatments 634 µatm
2
≤
but lower abundances in the three highest ƒCO treatments ( 953 µatm). This included members of the centric diatom
2
≥
20 genus Odontella (mainly O.weissfloggi but also some O.litigiosa) and the pennant diatoms Pseudonitzschia subcurvata and
Pseudonitzschiaturgidulodies(Figure4a,bandcrespectively).
TheabundanceoftwodiatomtaxawasunrelatedtoƒCO treatment.ChaetocerosandProbosciatruncatabothshowedno
2
ƒCO relatedtrendintheirabundancesdespitebeingrelativelylargediatoms(P.truncatacanexceed100µminlength)(Figure
2
5). This may reflect a lack of precision in the microscopy counts due to high variance among replicate fields of views. The
25 response of Chaetoceros (mainly C.castracanei but C.tortissimus and C.bulbosus were also present) is difficult to interpret.
Atday14theirabundancewaslowerinthethreehigherƒCO treatments.Byday18noƒCO relatedtrendwasevidentin
2 2
Chaetocerosabundance(Figure5b).
3.2.2 Flagellates
Colonial life stage Phaeocystis antarctica occurred in much higher abundances in the three lower ƒCO treatments. It was
2
30 the most abundant flagellate in this study, reaching concentrations from 1.0x105 to 1.26x107cellsL−1 at ƒCO2 levels
∼
634 µatm (Figure 6). This starkly contrasted with abundance at ƒCO levels 953 µatm in which they did not exceed
2
≤ ≥
1.6x106cellsL 1.
−
Otherflagellatetaxainourstudyshowedavarietyofresponses.ThechoanoflagellateBicostaantennigerashowedaƒCO
2
relatedresponsesimilartothatofP.antarcticawithhigherabundancesinƒCO treatments 506µatmandlowerabundances
2
≤
7
BiogeosciencesDiscuss.,https://doi.org/10.5194/bg-2017-224
ManuscriptunderreviewforjournalBiogeosciences
Discussionstarted:7June2017
c Author(s)2017.CCBY3.0License.
(cid:13)
athigherƒCO levels(Figure7a).Otherchoanoflagellates(mainlyDiaphanoecamultiannulata)showednoconsistenttrend
2
inresponsetoƒCO level(Figure7b).
2
TheabundancesofothernanoflagellatesandotherheterotrophicprotistswerelowandseeminglyunrelatedtoƒCO treat-
2
ment. All other nanoflagellates were low in abundance and had high variance between field of view counts, therefore these
5 weregroupedtogether.Thisgroup,"OtherFlagellates",includingspeciesP.antarctica(gameteandflagellateforms),Telonema
antarctica,Leucocryptossp.,Polytomasp.,Pyramimonasgelidicola,Geminigerasp.,Mantoniellasp.,Bodosp.,Triparmalae-
vissubspramisping,T.laevissubsp.pinnatilobataandanunidentifiedhaptophyte.Similarlymicroheterotrophscomprisedonly
of 1%ofallcellsandwasdominatedbyanunidentifiedeuglenoid( 0.8%).Dinoflagellatesweregroupedintoautotrophic
∼ ∼
dinoflagellates(mainlyGymnodimiumandHeterocapsa)andheterotrophicdinoflagellates(predominantlyGyrodiniumsp.,G.
10 glaciale, G. lachryma and Protoperidinum cf. antarcticum). Ciliates were grouped together but were mostly comprised of
Strombidiumsp.WhilenoneofthesefunctionalgroupsshowedatrendinabundancecorrelatedwithƒCO treatment,thismay
2
beduetotheabundanceofthesetaxabeingpoorlyresolved.
3.3 Community-levelresponses
Analysisofthecommunity-levelresponseshasbeenseparatedintothe8dayacclimationand10daygrowthperiods.During
15 acclimationthegrowthofthecellswaslimitedbylowlight.SIMPROFanalysisofthemicrobialcommunityovertheacclima-
tionperiodidentifiedthreesignificantgroups(p<0.05)(Figure8).Group1iscomprisedofallthetreatmentsatday1,group
3containsofallthetreatmentsoverdays3,5and8exceptforthelowestƒCO treatmentonday3(D3T1,group2).
2
Cluster analysis and SIMPROF based on the composition of the protistan community identified ten significantly different
groups of samples (p-value <0.05) during the growth period (days 8 to 18) (Figure 9a). On days 8 and 10 the communities
20 did not differ among treatments, except for the highest ƒCO2 treatment on day 10 (D10T6), which was clustered with the
day 8 samples (clusters 8 and 9, Figure 9a). Day 12 treatments are scatter across the cluster groups but day 14 samples are
againgroupedtogether(alltreatmentsexcept634µatmtogetherincluster6).Onday16theƒCO treatmentswereclustered
2
togetherexceptatthehighestƒCO level.Byday18theƒCO treatmentshadseparatedintotwodistinctlydifferentgroups;
2 2
onewiththethreelowestƒCO treatmentsandthesecondwiththethreehighestƒCO treatments.Interestingly,thesethree
2 2
25 highest ƒCO2 treatments fall into the cluster with the day 16 samples (or nearby cluster 2). This shows that at day 18 the
higherƒCO treatments( 953µatm)containedaprotistancommunitymoresimilartothatofday16,andweresignificantly
2
≥
differenttothatoftheday18lowerƒCO treatments(Figure9a).
2
AnMDSintwodimensionsprovedareasonableapproximationtothefullmultivariatestructure(stress=0.05),andshows
thesimilarityamongclustersalongwiththerelativecontributionofeachspecifictaxon/functionalgroupstoeachcluster(the
30 morecloselyspeciesarelocatedtoasampleinthenMDS,themoreabundantitisinthatsample)(Figure9b).Thecommu-
nity at day 8 was dominated by discoid centrics diatoms below 40 µm, flagellates, ciliates, P.subcurvata and dinoflagellates
(heterotrophicdinoflagellatesarelocatedofftotheleftoftheplotinFigure9b).Byday18thecommunityhadshiftedtobe
dominatedbyFragilariopsis,T.antarctica,T.ritscheri,OdontellaandP.antarctica(Figure9b).Betweenthestartandtheend
oftheexperimentothertaxaemergedbetweendays10and14includinganunidentifiedeuglenoid,L.annulata,B.antennigera
8
BiogeosciencesDiscuss.,https://doi.org/10.5194/bg-2017-224
ManuscriptunderreviewforjournalBiogeosciences
Discussionstarted:7June2017
c Author(s)2017.CCBY3.0License.
(cid:13)
andothercentricdiatoms.Othertaxaincreasedinabundancebetweendays14and16,includingotherchoanoflagellates,large
centricdiatoms(ie.P.truncataandS.microtrias),andpennatediatoms(ie.P.turgidulodies)(Figure9b).Attheendoftheex-
periment(day18)largediatoms,T.ritscheri,OdontellaandFragilariopsis(>20µm)aswellasP.antarcticaarelocatedclose
tothelowerƒCO treatmentsforday18(cluster1),andresistantFragilariopsis(<20µm)isfoundnearthehighƒCO treat-
2 2
5 ments(cluster4).Interestingly,T.antarcticaisalsolocatedclosetothelowerƒCO2treatmentsdespitebeingquiteresistantto
increasedƒCO whenanalysedasasinglespecies(Figure9bandFigure2b).
2
From the nMDS the lag in community growth and succession in the highest ƒCO level (tank 6) can be seen. At day 10,
2
thecommunityisgroupedwiththatonday8,likewiseatday12,14,16and18itisconsistentlyclosertosamplesfromthe
previoustimepoint(Figure9b).Theseresultsareconsistentwiththechangesintotalcellabundance(Figure1).Thehighest
10 ƒCO2levelinhibitsgrowthandsuccessionintheprotistancommunitysuchthatitisconsistentlyatimepointbehindtheother
ƒCO treatments.
2
CAP analysis showed differences in the trajectory of the protistan community succession over time in the different ƒCO
2
treatments(Figure10).Thisanalysis,usingcovariatesƒCO ,NO ,SiandP,providedamodelwhichexplained71.61%of
2 x
the variation in similarity among samples (F = 19.544, p <0.001 based on 999 permutations). However, NO was not
4,31 x
15 significant (F1,32 = 1.3714, p >0.200 based on 999 permutations) and so was dropped from the model. In this reduced CAP
model,CAP1andCAP2werebothsignificant(p<0.05)(Table2),andtheremainingtermsƒCO ,SiandPtogetheraccounted
2
for70.35%ofthetotalvariance(F =25.308,p<0.001basedon999permutations).Consideringeachtermmarginaltoall
3,31
others(ie.thecontributionofatermafterfirstaccountingforallotherterms),ƒCO accountedfor2.92%,P22.68%andSi
2
5.21%ofthevariance(Table3a).Theremainingtermsinthereducedmodelwereallsignificantwhensequentiallyadded,but
20 themarginaleffectsshowedthatonlyPandSiweresignificant,notƒCO2(p>0.100)(Table3b).TheCAPanalysisshowsthere
is a separation of the protistan community between low ( 506 µatm), medium (634 to 1140 µatm) and high (1641 µatm)
≤
ƒCO treatments(Figure10).Atday18therearetwodistincttreatmentgroups,thosenamelyexposedtolowandmoderate
2
ƒCO (343,506and634µatm)andthoseexposedtohighƒCO (953,1140and1641µatm).
2 2
4 Discussion
25 4.1 AcclimationtohighƒCO2
Changesintheresponseoftheprotistancompositionandabundanceduringtheacclimationperiodoftheexperiment(days1
to8)werelikelyduetothetransferandestablishmentofthenaturalcommunitiesintheminicosmtanksratherthanexposure
toƒCO .Therewasasignificantshiftinspeciescompositionbetweendayoneandallothersamplestimesduringacclimation.
2
The collection of seawater (via Bambi bucket under a helicopter) and subsequent filling of the minicosms is likely to have
30 damaged delicate protists (Estrada and Berdalet, 1997), and the minicosm conditions may have been sub-optimal for some
species(Kimetal.,2008).Thereforeitislikelythischangeincommunitycompositionreflectedthedeclineindelicatespecies.
The total abundance of the protistan community showed an initial lag in growth at the highest ƒCO level. Our results
2
indicatethattheprotistsrequiredmorethan8daystoacclimatetothishighƒCO andthereforegrowthwasdelayedcompared
2
9
BiogeosciencesDiscuss.,https://doi.org/10.5194/bg-2017-224
ManuscriptunderreviewforjournalBiogeosciences
Discussionstarted:7June2017
c Author(s)2017.CCBY3.0License.
(cid:13)
to other treatments. Deppeler et al. (submitted) showed that during acclimation there is a decrease in photosynthetic health
of the community, but whilst all treatments had recovered by day 12, the highest ƒCO treatment had a greater decline in
2
photosynthetichealthandtooklongertorecover.Thislagwasalsoseeninaccumulationofchlorophylla,rateofproductivity
andnutrientconcentrationswheretheslowergrowthmeantthiswastheonlytreatmentinwhichnutrientswerenotlimitingby
5 theendoftheexperiment(Deppeleretal.,submitted).AfterthislagtherateofgrowthinthehighestƒCO2treatmentissimilar
toothertreatments.Toourknowledgethisisthefirststudytoreportsuchanacclimationresponseinashort-termexperiment.
4.2 Autotrophicprotisttaxaspecificresponses
Themicrobialcommunitywithinthisstudywashighlydiverse,andthedetailedtaxonomicclassificationemployedallowedthe
rangeofresponsesbytheindividualtaxatoberesolved.Indiatomstheresponsewasmainlysize-related.Smalldiatomsshowed
10 a strong resistance to high levels of ƒCO2 but large diatoms showed a non-linear response with an increase in abundance at
moderate ƒCO levels (634-953 µatm) but a decrease a higher ƒCO . This trend is even evident within Fragilariopsis cells
2 2
>20µmwhichshowedasimilarresponsetootherlargecentricdiatoms,butthose 20µmshowedahigherresistancesimilar
≤
toothersmallerdiatoms.Whilsttheresponseformostdiatomtaxawasrelatedtocellsize,acoupleofspeciesdidnotfollow
thistrend.ProbosciatruncataisalargecellbutdidnotshowaƒCO response,likewiseChaetocerosdidnotshowaresponse
2
15 toƒCO2 butinsteadreflectedthenutrientavailability(seeDeppeleretal.submittedfornutrientdata).AnumberofAntarctic
studieshaveshownanenhancementoflargerdiatomspecieswithincreasesinCO (ie.Engeletal.2008,Tortelletal.2008and
2
Fengetal.2010).Therearealsootherpreviousstudiestowhichourresultsdonotalign.Fengetal.(2009)foundanincreasein
abundanceofPseudonitzschiasubcurvataathighƒCO .InthisstudywerePseudonitzschiahadaverylowtolerancetoƒCO
2 2
levelsabovethecontrol,differenttothatofFengetal.(2009).
20 Unlikediatomspecies,thedominantautotrophicnanoplankton,Phaeocystisantarctica,dramaticallydeclinedinabundance
at the three highest ƒCO levels. There has been a common consensus in other ocean acidification studies that pico- and
2
nanoplanktonabundanceincreasesathighCO levels(ie.Hareetal.2007,Brussaardetal.2013,Davidsonetal.2016,Thom-
2
sonetal.2016andarecentreviewbySchulzetal.2017),butourstudyonlyfindsthisresponsesindiatoms.P.antarcticahad
anincreaseinabundancewithmoderateƒCO levelsbuthadastrongthresholdlevelbetween634and963µatm,abovewhich
2
25 abundances were very low. This contrasts with the findings of many ocean acidification studies that found that P.antarctica
abundancewaseithernoteffectedorincreasedbyelevatedCO attheexperimentalCO levelsusedintheirstudies(Tortell
2 2
etal.,2008;Fengetal.,2010;Trimbornetal.,2013).ItisnotedthatthesestudieswereconductedintheRossSeawhichisa
verydifferentecosystemtoothercoastalAntarcticareas(Smithetal.,2014;DeppelerandDavidson,2017).Thedifferencein
responseofP.antarcticacomparedtootherstudiesshowinganincreaseinpico-andnanoplankton(Schulzetal.,2017),may
30 bebecauseofthepeculiarphysiologyofthisalgainthecoloniallifestage(whichdominatedinourstudy).P.antarctica,when
incolonialform,protecttheircellsfromintimatecontactwiththesurroundingseawaterbyathintoughskinandbehaveina
similarfashiontoalargediatom(DavidsonandMarchant,1992;Hammetal.,1999;Hamm,2000).
Thisstudy’sƒCO treatmentsextendwellpasttherangeofotherstudies,wherecommonlythehighestlevelwasbetween
2
750and1000µatmcomparedto1641µatminthisstudy.WhentheresultsofthisstudyarecomparedinlightoftheseCO
2
10