Deep-SeaResearchII105(2014)1–16 ContentslistsavailableatScienceDirect Deep-Sea Research II journal homepage: www.elsevier.com/locate/dsr2 Phytoplankton blooms beneath the sea ice in the Chukchi sea Kevin R. Arrigoa,n,1, Donald K. Perovichb,c, Robert S. Pickartd, Zachary W. Browna, Gert L. van Dijkena, Kate E. Lowrya, Matthew M. Millsa, Molly A. Palmera, William M. Balche, Nicholas R. Batesf, Claudia R. Benitez-Nelsong, Emily Brownleeh, Karen E. Freyi, Samuel R. Laneyh, Jeremy Mathisj, Atsushi Matsuokak,l, B. Greg Mitchellm, G.W.K. Mooren, Rick A. Reynoldsm, Heidi M. Sosikh, James H. Swiftm aDepartmentofEnvironmentalEarthSystemScience,StanfordUniversity,Stanford,CA94305,USA bEngineerResearchandDevelopmentCenter,ColdRegionsResearchandEngineeringLaboratory,Hanover,NH03755,USA cThayerSchoolofEngineering,DartmouthCollege,Hanover,NH03755,USA dDepartmentofPhysicalOceanography,WoodsHoleOceanographicInstitution,WoodsHole,MA02543,USA eBigelowLaboratoryforOceanSciences,WestBoothbayHarbor,ME04575,USA fBermudaInstituteofOceanSciences,FerryReachGE01,Bermuda gMarineScienceProgram,andDepartmentofEarthandOceanSciences,UniversityofSouthCarolina,Columbia,SC29208,USA hBiologyDepartment,WoodsHoleOceanographicInstitution,WoodsHole,MA02543,USA iGraduateSchoolofGeography,ClarkUniversity,Worcester,MA01610,USA jNOAAPacificMarineEnvironmentalLaboratory,Seattle,WA98115,USA kUniversitéPierreetMarieCurie,Laboratoired’OcéanographiedeVillefranche,Villefranche-sur-Mer06238,France lTakuvikJointInternationalLaboratory,DépartementdeBiologieandQuébec-Océan,UniversitéLaval,Canada mScrippsInstitutionofOceanography,UniversityofCaliforniaSanDiego,LaJolla,CA92093,USA nDepartmentofPhysics,UniversityofToronto,Toronto,Ontario,CanadaM5S1A7 a r t i c l e i n f o a b s t r a c t Availableonline3April2014 IntheArcticOcean,phytoplanktonbloomsoncontinentalshelvesareoftenlimitedbylightavailability,andare therefore thought to be restricted to waters free of sea ice. During July 2011 in the Chukchi Sea, a large Keywords: Arctic phytoplanktonbloomwasobservedbeneathfullyconsolidatedpackiceandextendedfromtheiceedgeto Seaice 4100km intothe pack.Thebloomwas composed primarilyof diatoms,with biomass reaching1291mg Phytoplankton chlorophyllam(cid:1)2andratesofcarbonfixationashighas3.7gCm(cid:1)2d(cid:1)1.Althoughtheseaicewherethe bloomwasobservedwasnear100%concentrationand0.8–1.2mthick,30–40%ofitssurfacewascoveredby meltpondsthattransmitted4-foldmorelightthanadjacentareasofbareice,providingsufficientlightfor phytoplanktontobloom.Phytoplanktongrowthratesassociatedwiththeunder-icebloomaveraged0.9d(cid:1)1 andwereashighas1.6d(cid:1)1.Wearguethatathinningseaicecoverwithmorenumerousmeltpondsoverthe pastdecadehasenhancedlightpenetrationthroughtheseaiceintotheupperwatercolumn,favoringthe developmentoftheseblooms.Theseobservations,coupledwithadditionalbiogeochemicalevidence,suggest that phytoplankton blooms are currently widespread on nutrient-rich Arctic continental shelves and thatsatellite-basedestimatesofannualprimaryproductioninwaterswhereunder-icebloomsdevelopare (cid:3)10-foldtoolow.Thesemassivephytoplanktonbloomsrepresentamarkedshiftinourunderstandingof Arcticmarineecosystems. &2014ElsevierLtd.Allrightsreserved. 1. Introduction (Maslaniketal.,2011;Nghiemetal.,2007).Thishasproducedan icecoverthatissubstantiallythinnerandmorepronetomelting Overthelastseveraldecades,theArcticOceanhasundergone and transport, leading to a markedly extended period of open unprecedented changes in sea ice, with summer minimum ice water, particularly over the last decade (Arrigo and van Dijken, extentdeclining 440%since1979(Comisoetal.,2008)andfirst- 2011). year ice largely replacing the once prevalent multi-year pack ice Associatedwiththelossinseaiceontheseshelveshasbeenan increaseintheamountoflightpenetratingthesurfaceoceananda dramatic rise in the productivity of phytoplankton (Arrigo and nCorrespondingauthor. vanDijken,2011),theorganismsresponsibleforthebulkofArctic E-mailaddress:[email protected](K.R.Arrigo). Oceanprimary production and constitute the base of the marine 1DepartmentofEnvironmentalEarthSystemScience,473ViaOrtega,Stanford University,Stanford,CA94305-4216,USA. food web. This is particularly true for the Pacific sector of the http://dx.doi.org/10.1016/j.dsr2.2014.03.018 0967-0645/&2014ElsevierLtd.Allrightsreserved. 2 K.R.Arrigoetal./Deep-SeaResearchII105(2014)1–16 Arctic Ocean where annual production increased by 130% in the masspropertiessetwithintheBeringSea(Coachmanetal.,1975; East Siberian Sea and 48% in the Chukchi Sea between 1997 and Overland and Roach, 1987; Weingartner et al., 2005), they also 2009 (Arrigo and van Dijken, 2011). Because sea ice and snow differsignificantlywithrespecttonutrientconcentrations(Walsh strongly reflect and attenuate incident solar radiation (Perovich, et al., 1989; Cooper et al., 1997; Codispoti et al., 2005, 2013). 1998;PerovichandPolashenski,2012),thegrowthofphytoplank- The easternmost Alaska Coastal Water (ACW) is relatively warm ton at high latitudes is generally thought to begin in the open (1–61C),fresh(So31.8),andnutrient-poor(NO(cid:1)o10mM)duetothe 3 waters of the marginal ice zone (MIZ) once sea ice retreats in inputofriverrunoffandthebiologicaldrawdownofnutrientsinthe spring, as solar elevation increases and surface waters become eastern Bering Sea. The Central Channel pathway consists of Bering stratified by the addition of sea ice melt water (Alexander and Shelf Water with moderate nutrients (NO(cid:1)(cid:3)15mM) and salinity 3 Niebauer,1981;Loengetal.,2005;Sakshaug,1997;HillandCota, (31.8–32.5). The westernmost Herald Canyon pathway, containing 2005;Perretteetal.,2011).Infact,allcurrentlarge-scaleestimates mostlyAnadyrWater,hasthehighestsalinity(32.5–33)andnutrient of primary production in the Arctic Ocean assume that phyto- concentrations(pre-bloomNO(cid:1)425mM)duetotheunder-utilization 3 plankton production in the water column under sea ice is of nutrients in the western Bering Sea (Hansell et al., 1993). negligible (Subba Rao and Platt, 1984; Sakshaug, 2003; Pabi Upon reaching the shelf break, some of this water turns eastward et al., 2008; Arrigo and van Dijken, 2011). However, an intense andflowstowardtheBeaufortSeainarelativelyswiftshelfbreakjet phytoplankton bloom was recently observed in the Chukchi Sea (Weingartneretal.,2005). growing beneath fully consolidated sea ice ranging in thickness WatermasspropertiesintheChukchiSeaarefurtherinfluenced from 0.8 to 1.3m (Arrigo et al., 2012). The development of this bytheseasonalcycleofseaice.Inthewinter,brinerejectionduring bloomhasbeenattributedtoincreasedlighttransmissionthrough seaiceformationmixestheentirewatercolumntoextremelycold seaice(Freyetal.,2011)aswellashighnutrientconcentrationson temperatures((cid:1)1.81C)(Woodgateetal.,2006).Seaiceformationin theChukchishelf. polynyas and leads continue to convectively form cold and dense The Chukchi Sea is an inflow shelf (Carmack and Wassmann, winterwater(WW)ontheChukchishelfthroughoutthewinter.A 2006) that ventilates the upper halocline of the Arctic Ocean large fraction of this becomes nutrient-rich Pacific Winter Water (Woodgate et al., 2005). Water flows northward through the thatdrainsthroughtheChukchiSeainthespringandeventuallyfills BeringStraitduetotheseasurfaceheightdifferentialthatresults theArcticOceanhalocline.Asseaiceretreatsinspringandsummer, fromthesalinitydifferencebetweenthePacificandArcticOceans the water column becomes re-stratified as surface waters freshen (Coachman et al.,1975), with the volume flux increasing by 50% and warm (Woodgate and Aagaard, 2005). WW remaining on between 2001 and 2011 (Woodgate et al., 2012). Upon reaching theChukchishelfinthesummerisgraduallyreplacedbyrelatively the Chukchi Sea, the Pacific-origin water separates into three warmPacificSummerWater(Weingartneretal.,2005). branches that flow around or between Herald and Hanna Shoals Becauseofitshighnutrientcontent,theChukchiSeaisaregion (Fig. 1). Although these branches are identified based on water of intense summer biological activity with a rich benthic 150 m 100 m 3000 m 2000 m 80 m 1000 m 60 m 40 m Hanna 40 m Central Shoal Channel 40 m Herald Shoal 20 m Herald Canyon 40 m 20 m 40 m 40 m 20 m 80 m 60 m 40 m Fig.1. MapoftheChukchiSeashowingthemajorbathymetricfeaturesandthepredominantflowpathsfromtheBeringSea.CurrentsmodifiedfromWeingartneretal. (2005). K.R.Arrigoetal./Deep-SeaResearchII105(2014)1–16 3 community that supports abundant populations of marine mam- exceeded the surface value by 0.05. Water was collected at mals and seabirds (Loeng et al., 2005; Dunton et al., 2005; standard depths (5, 10, 25, 50, 75, 100, 150, and 200m) and at Grebmeier et al., 2006). Depth-integrated primary production the depth of the fluorescence maximum into 30L Niskin bottles. andchlorophylla(Chla)intheChukchiSeaareamongthehighest Currentspeedanddirectionweremeasuredwithahullmounted in the world (Springer and McRoy, 1993; Arrigo et al., 2012). OceanSurveyor150KHzADCP(TeledyneRDInstruments,OS150). Phytoplankton contribute 490% of the total primary production Sea ice thickness. Sea ice thickness was measured directly by in the Chukchi Sea (Hill and Cota, 2005), although sea ice algal drilling holes through the ice. Surveys of ice thickness were also productioncanbeimportantearlyintheseason(Gradinger,2009). conducted using a Geonics EM-31 electromagnetic induction Intensephytoplanktonbloomshavebeenobservedthroughoutthe sensor that determines the ice thickness with an accuracy of Chukchi shelf, with spatial variation attributed to differences in 70.05mbyexploitingthelargeconductivitydifferencebetween water masses, nutrient availability, and environmental forcing sea ice and the underlying seawater (Eicken et al., 2001). The (Cota et al., 1996; Wang et al., 2005; Lee et al., 2007). The instrument transmits a primary electromagnetic field and mea- southeastern Chukchi Sea and Barrow Canyon have specifically suresthestrengthofasecondaryfieldinducedintheocean.The been identified as hot spots for primary production (Sukhanova induced field strength is related to the distance from the instru- et al., 2009), while waters near the coast of Alaska in the ACW menttotheseawaterandthereforetheicethickness. (Coupel et al., 2012) and off the continental slope generally have lowerbiomass(Leeetal.,2012). HerewepresenttheresultsfromtherecentICESCAPE(Impacts 2.2. Phytoplankton of Climate on EcoSystems and Chemistry of the Arctic Pacific Environment)programdocumentingingreaterdetailthephysical, Chlorophylla.Underwayseawaterwaspumpedfromadepthof chemical,andbiologicalcharacteristicsofanintensephytoplank- 8m as part of the Uncontaminated Science Seawater System tonbloomobservedbelowthefirstyearseaiceoftheChukchiSea (USSS). Some of this seawater was diverted to a WETLabs WET- (Arrigoetal.,2012).Thepresentpaperprovidesanoverviewofthe STAR flow-through fluorometer. Fluorescence was converted to observed under-ice phytoplankton bloom. Companion papers in Chl a concentration by calibrating against Chl a concentrations this issue address more specific aspects of the bloom, including measured in discrete seawater samples (see below) that were (1)thehydrographicconditionsleadingtothelargeaccumulations collectedatleastfourtimesdailyfromtheUSSS. of algal biomass beneath the ice (Spall et al., 2014), (2) the ice Samples for fluorometric analysis of Chl a were filtered onto conditions required for under-ice bloom development (Palmer 25mmWhatmanGF/Ffilters(nominalporesize0.7mm),placedin etal.,2014),(3)thetypesofphytoplanktonobservedinassociation 5mL of 90% acetone, and extracted in the dark at 31C for 24h. with the bloom (Laney and Sosik, 2014), (4) the impact of the Chl a was measured fluorometrically (Holm-Hansen et al.,1965) under-ice bloom on the optical properties of the water column usingaTurner10-AU(TurnerDesigns,Inc.).Thefluorometerwas (Neukermans et al., 2014), the response of the bacterial commu- calibratedusingapureChlastandard(Sigma). nitytotheunder-icephytoplanktonbloom(Ortega-Retuertaetal., Particulate organic carbon. Particulate organic carbon (POC) 2014),andasatellite-basedanalysisofthefrequencyofunder-ice samples were collected by filtering water samples onto pre- phytoplankton blooms on the Chukchi shelf in the recent past combusted (4501C for 4h) 25mm Whatman GF/F filters. Filter (Lowry et al., 2014). Observations made within the sea ice zone blanks were produced by passing (cid:3)50mL of 0.2mm filtered (SIZ)ontheChukchishelfcontradictthecommonperceptionthat seawater through a GF/F. All filters were then immediately dried watersbeneaththeconsolidatedicepackharborlittlephytoplank- at 601C and stored dry until analysis. Prior to analysis, samples ton biomass. Instead, our results suggest that phytoplankton andblankswerefumedwithconcentratedHCl,driedat601Cand growing beneath a thinner Arctic ice pack composed mostly of packed into tin capsules (Costech Analytical Technologies, Inc.) firstyearicecancontributegreatlytoannualprimaryproduction, for elemental analysis on a Elementar Vario EL Cube (Elementar and evenexceedrates measured in openwatersafterthesea ice Analysensysteme GmbH, Hanau, Germany) interfaced to a PDZ hasretreated. Europa 20–20 isotope ratio mass spectrometer (Sercon Ltd., Cheshire,UK).Standardsincludedpeachleavesandglutamicacid. Fast repetition rate fluorometry. The maximum efficiency of 2. Methods photosystem II (Fv/Fm) of discrete samples was determined by fast repetition rate fluorometry (FRRf) (Kolber et al., 1998) on a 2.1. Studyregion custommadeinstrument(Z.Kolber).Samplesweredarkadapted for 30min at in situ seawater temperatures beforemeasurement During the ICESCAPE program, we sampled the Chukchi Sea with the FRRf. Blanks for individual samples analyzed by FRRf continental shelf between the Bering Strait and the western werepreparedbygentlefiltrationthrougha0.2mmpolycarbonate BeaufortSeafrom18Junethrough16July2010andfrom28June syringe filter before measurement using identical protocols. All through24July2011onboardtheUSCGCHealy.Wesampled140 reported values of Fv/Fm have been corrected for blank effects stations (including 10 sea ice stations) in 2010 and 172 stations (CullenandDavis,2003). (including 9 sea ice stations) in 2011. Here we focus on stations Photosynthesis vs. Irradiance. P–E parameters (Pn – maximum m 46–57(Transect1)and57–71(Transect2)sampledfrom3–8July photosynthetic rate, αn – photosynthetic efficiency, and E – k 2011 (Fig. 2A) but also refer to stations 38–55 (29–30 June) and photoacclimation parameter) were determined using a modified 71–84 (6–7 July) sampled in 2010 (Fig. 2C). These stations were 14C-bicarbonate incorporation technique (Lewis and Smith,1983; located along four different transects that extended from open Arrigo et al., 2010). Carbon uptake, normalized by Chl a concen- waterintothepackiceandallexhibitedevidenceofanunder-ice tration, was calculated from radioisotope incorporation, and the phytoplanktonbloom(Fig.2). datawerefitbyleastsquaresnonlinearregressiontotheequation Ateachstation,watercolumnprofilesoftemperature,salinity, of Platt et al. (1980). P–E parameters were used with under-ice and oxygen were measured using a conductivity, temperature, light profiles to estimate rates of depth-integrated daily gross anddepthsensor(Sea-BirdElectronics)andoxygensensor(SBE43, primary production. Specific growth rate (m, d(cid:1)1) for a given Sea-Bird Electronics) attached to a rosette. Mixed layer depth sample was calculated by multiplying the photosynthetic rate (MLD) was calculated as the depth where potential density (Pn)bytheChla:POCratiofromthatsample. 4 K.R.Arrigoetal./Deep-SeaResearchII105(2014)1–16 170°W 165°W 160°W 74°N July 29 July 22 Beaufort July 22 Sea 57 59 Chukchi 73°N 61 Sea 73°N 56 63 Transect 2 ALASKA 65 55 67 Transect 1 52 50 69 71 July 1 72°N 48 46 72°N June 17 71°N June 3 71°N 74°N 150 m 3000 m 150 m 1000 m 73°N 100 m 73°N 80 m 80 m 60 m 60 m 50 m 72°N Hanna Shoal 72°N Central Channel 50 m 40 m 71°N 40 m 71°N 40 m 170°W 165°W 160°W 0.01 0.1 0.2 0.5 1 25 10>20 Chlorophyll a (µg L-1) 170°W 160°W 150°W 74°N Aug 10 Aug 3 July 27 73°N 73°N July 20 July 20 July 27 71 7375 July 27 72°N Chukchi 7779 Aug 3 72°N 81 North transect 84 71°N July 20 71°N Chukchi South 40 44 38 42 46 50 transect 70°N 53 70°N 55 ALASKA 69°N June 15 69°N 170°W 160°W 150°W Fig.2. MapsofthenorthernChukchiSeashowing(A)thedistributionofseaiceon8July2011(MODIS-Aqua)andthelocationofstationssampledduringtheICESCAPE2011 cruise.Blackindicatesopenwater.Linesshowthepositionoftheiceedgeontheindicateddates(AMSR-E).Inthetext,stations46–57arereferredtoasTransect1andstations 57–71areTransect2.(B)Bathymetryofthestudyareaoverlaidbysurfacechlorophyllaconcentrationcalculatedfromfluorescencemeasuredbythecontinuousseawatersystemof theUSCGCHealyinJuly2011.(C)MapoftheChukchiSeashowingthedistributionofseaiceon8July2010andthelocationofselectstationssampledduringtheICESCAPE2010 cruise(numberedwhitesquares).Anunder-icephytoplanktonbloomofsimilarintensitywasobservedalongtheChukchiNorthtransectduringICESCAPE2010atthesametimeof year(6–7July2010)andinthesamelocationasTransect1inICESCAPE2011(forreference,otherstationsfromICESCAPE2011areshownasun-numberedgrayboxes). Unfortunately,theChukchiNorthtransectstartedintheseaiceandprogressedsoutheastwardtowardopenwaterduringICESCAPE2010,withoutenoughtransitintotheicepack tosamplethefullextentoftheunder-icebloom. K.R.Arrigoetal./Deep-SeaResearchII105(2014)1–16 5 Net primary production. Simulated in situ daily net primary using an automated imaging flow cytometric approach. Small production (NPP) was determined by measuring 14C-bicarbonate volumes of seawater (2.5 or 5mL) were subsampled from bottle incorporation in water samples collected from different light casts and then analyzed with an Imaging FlowCytobot (Olson and depths and incubated at corresponding light intensities in an Sosik, 2007). This generated large numbers of images of individual on-deckincubatorfor24h.Weadded0.74MBq14C-bicarbonateto cells of length scale (cid:3)10 to 4100mm, which were then sorted 150mL of sample in a 250mL Falcon flask and covered the flask manually into appropriate taxonomic classes, typically to the genus with 0 to 9 layers of neutral density screens to simulate light level.Thebiomasscontributedbycellsineachclassateverystation intensities of 85, 65, 25,10, 5 and 1% of surface irradiance. After and depth was estimated from the cross-sectional area of cells, incubation, 30mL of sample was filtered in triplicate under very quantified by automated image processing (Sosik and Olson, 2007) lowvacuum(o50mmHg).Filterswereacidifiedwith0.1mL6N and calibrated with a fluorescent microsphere standard of known HCl to drive off inorganic C. After 24h of acidification, 5mL of diameter.Taxonomiccompositionofnano-andmicro-phytoplankton scintillation cocktail (Ecolume) was added and samples were were also verified onboard ship. Samples of 50–100mL were first countedafter 43honaPerkinElmerTri-Carbliquidscintillation filteredonto0.4mmpolycarbonatefilters(HewesandHolm-Hansen, counter.Totalactivitywasdeterminedoneachsamplebycombin- 1983)andthenviewedwithacompoundlightmicroscopeequipped ing50μLofsamplewith50μLofethanolamine,0.5mLoffiltered withstandardbrightfield,epifluorescence,orpolarizedillumination seawater, and 5mL of scintillation cocktail. Time zero controls for imaging all cells, chlorophyll/phycoerythrin-containing phyto- werefiltered(30mLintriplicate)andacidifiedatthestartofthe plankton,orcalcifyingplankton,respectively. incubationperiod. Specificgrowthrate(m,d(cid:1)1)insurfacewaters wascalculatedbynormalizingtheNPPratebyPOCconcentration. Phytoplanktoncommunitycomposition.Thetaxonomiccomposition 2.3. Nutrients,oxygen,anddissolvedinorganiccarbon(DIC) of the nano- and microphytoplankton assemblage was determined Nitrate (NO ), Oxygen, and DIC concentration. Discrete water 3 columnsampleswereanalyzedforNO andnitrite(NO )concen- 3 2 trations with a Seal Analytical continuous-flow AutoAnalyzer 3 74°N (AA3) using a modification of the procedure by Armstrong et al. (1967). Seawater samples for DIC were drawn from the Niskin 73°N samplersintopre-cleaned (cid:3)300mLborosilicatebottles,poisoned with HgCl to halt biological activity, sealed, and returned to 2 the Bermuda Institute of Ocean Sciences (BIOS) for analysis. 72°N DIC samples were analyzed using a highly precise ((cid:3)0.025%; o0.5mmoleskg(cid:1)1) gas extraction/coulometric detection system 71°N (Bates et al., 2005). Analyses of Certified Reference Materials (provided by A. G. Dickson, Scripps Institution of Oceanography) 70°N ensured that the accuracy of the DIC measurements was 0.05% ((cid:3)0.5mmoleskg(cid:1)1). The oxygen sensor on the CTD rosette 69°N (SBE43)wascalibratedusingstandardWinklertitrations. NitrateandDICdeficitcalculations.WatercolumnNO andDIC 3 deficitswerecalculatedbyfirstassumingthatthewintertimeNO 170° 165° 160° 155°W 3 and DIC concentrations were equal to the maximum WW con- centrationsmeasuredalongthestudytransects.WWwasdefined as water deeper than 10m with a temperature below (cid:1)1.651C. 0 170°W Thus, WW NO and DIC concentrations over the Chukchi shelf 3 were24.1and2300mmolL(cid:1)1,respectively.Thevalueweusedfor WW NO was virtually the same as that calculated previously 3 (Hanselletal.,1993)fortheChukchiShelf(23.674.86mmolL(cid:1)1). 50 Assumingthatthewatercolumnwasthoroughlymixedvertically 165°W duringthewinter(Fig.3),NO andDICdeficitswerecalculatedas 3 m) thedifferencebetweenthedepth-integratedwatercolumnNO3or pth ( DorICDdICurminegastuhreewdidnuterirnagntdhethIeCEdSeCpAthP-Ein2t0e1g1ractreudiswe.aAtellrNcoOl3umanndNDOIC3 e 100 concentrationswerenormalizedtoaconstantsalinityof33.1prior D tomakingdeficitcalculationstocorrectformeltwaterinput. 160°W 2.4. Optics 150 Inherentoptical properties.Thebeamattenuationcoefficientat 488nmwasmeasuredinsituwithaC-Startransmissometer(WET 155°W Labs,Inc.)alonga25cmpathlengthwithinseawater.Thespectral 0 5 10 15 20 absorptioncoefficientofcoloreddissolvedorganicmatter(CDOM) Nitrate concentration (µmol L-1) and particles was determined at 1nm resolution from freshly- collected Niskin bottle samples. CDOM absorption in the sample Fig.3. VerticalnitrateprofilesfromtheChukchiSeameasuredduringtheShelf– filtrate (o0.2mm) was measured at 200–735nm using an Ultra- BasinInteractions(SBI)program.(A)Mapshowinglocationsofstationssampledin Path liquid waveguide system (World Precision Inc.) equipped Maypriortosignificantphytoplanktongrowth.(B)Nitrateconcentrationsarehigh with a 2m pathlength. Reference water salinity was adjusted to insurfacewatersinthewesternportionofthestudyarea,reflectingconvective thatofthesamplewithpre-combustedNaClandMilliQwater.The processesontheshallowshelfthatmixnutrientstothesurfaceinwinter.Colors denotelongitudewheresamplesweretaken. particle absorption coefficient (ap) from 300 to 800nm was 6 K.R.Arrigoetal./Deep-SeaResearchII105(2014)1–16 determinedbycollectingparticlesontoa25mmfilter(Whatman GF/F) and measuring its optical density relative to a blank reference filter in a Cary 1E spectrophotometer. Spectral absorp- tion coefficients were calculated as described in Mitchell (1990). Following measurement of a , sample filters were extracted in p 100%methanolandre-measuredtoyielddetritalabsorption(a ). d Phytoplankton absorption (a ) was determined by difference as ph a ¼a (cid:1)a . ph p d Thedrymassconcentrationofsuspendedparticleswasdeter- minedgravimetricallyonsamplesfilteredontopre-weighedGF/F filters, rinsed with deionized water to remove sea salt, dried at 601C, and measured with a Mettler-Toledo MT5 microbalance witharesolutionof0.001mg.Particlenumberconcentrationsand sizeestimatesweredeterminedwithaCoulterCounterMultisizer III(Beckman-Coulter)equippedwith30mmand200mmaperture Bare ice Melt pond tubes (Reynolds et al., 2010). The estimates of particle number reported here represent the total concentration of particles over Surface 70% 100% 20% 100% the size range of 1–100mm. Particle median diameters were scattering layer calculatedfromthe volumedistribution overthis size range, and 0.05 m representthediameterofavolume-equivalentsphere. 0.1 m Under-ice optical measurements. Light transmittance through theiceandtheupperwatercolumnwasmeasuredatstations55, 56, and 57 (Fig. 2A). At each station, vertical light profiles were obtained beneath both representative bare ice and melt ponds. Congelation Scattering and Individual sampling sites werechosen tomaximize the ability to 1.0 m ice layer absorption 0.9 m uniformly represent each ice surface type (i.e., in the center of a relativelylargemeltpondorinthecenterofarelativelylargebare icesurface).AnAnalyticalSpectralDevicesDualDetectorspectro- radiometer was used to measure transmitted spectral irradiance overthewavelengthrangeof380–850nm.A modified Compact- Optical Profiling System (C-OPS, Biospherical Instruments Inc.) Seawater radiometer with a cosine collector was lowered to a depth of (cid:3)50m through an auger-drilled (cid:3)25cm hole in the ice. At the 15% 55% bare ice surface site, the hole was re-filled with ice tailings to Fig.4. MeltpondsandbareiceintheChukchiSea.(A)Photographtaken13July mimic the previously undisturbed bare ice surface. At the melt 2011byChristieWood.(B)Schematicshowinglighttransmissionthroughbareice pond site, the C-OPS was offset horizontally from the hole such andmeltponds.Bareiceconsistsofathingranularsurfacescatteringlayerthat that the cosine collector on the instrument looked directly up at givesiceitswhiteappearanceandathickcongelationicelayerthatconsistsof the underside of the ice. The underwater C-OPS and surface columnaricecontainingnumerousbrineinclusions.Seaicebeneathmeltponds referencesensormeasureddownwellingirradianceat19channels hasnosurfacescatteringlayerandthecongelationiceisgenerallythinner.The albedoofbareice(70%)andmeltponds(20%)includesspecularreflectionatthe (320, 340, 380, 395, 412, 443, 465, 490, 510, 532, 555, 560, 625, air–ice interface and scattering of light back out of the ice interior. Light not 665, 670, 683, 710, 780nm, and photosynthetically active radia- backscatteredorabsorbedwithinthecongelationicelayeristransmittedtothe tion(PAR,400–700nm)).Asurfacereferencewasmountedontop upper water column. Because light transmitted through the ice spreads in all of a tripod ((cid:3)2.5m above the ice surface) that stood on the ice directions,lightlevelsbelowbareiceandmeltpondsconvergewithin (cid:3)10mof within (cid:3)1.5mofwheretheC-OPSmeasurementswerecollected. theiceinterface. stationsalong Transect2.By22July,ourentirestudyregionwas 3. Results ice-free. Thesnowcoverhadmeltedcompletelybythetimeofourstudy 3.1. Seaice and sea ice thickness increased significantly with distance from the ice edge. Near the edge at station 55 (Fig. 2A), mean ice Duringthe2011campaign,theseaiceedgeintheChukchiSea thickness was 0.7770.09m, including melt ponds that were generallyretreatedfromthesoutheasttothenorthwest(Fig.2A). 15–20% thinner than bare ice (Table 1). As we penetrated In the vicinity of our study area, sea ice was fully consolidated (cid:3)60km deeper into the pack, ice thickness increased to 1.037 ((cid:3)100% concentration) and largely undeformed, with a surface 0.15m (at station 56). By station 57, approximately 120km from composed of avariable mixof bareice and melt ponds (Fig. 4A). theiceedge,icethicknesshadincreasedto1.2170.10m.Surface Occasionally, a northerly wind shift would push broken pack ice meltpondfractionatthethreeseaicestationsaveraged(cid:3)0.3.The southward a short distance, thereby temporarily reducing ice meltpondshadwell-definedboundariesandsurfacesthatwereat concentrations. However, these events were short-lived and the sealevel,implyingthattheywerebeyondtheinitialfloodingstage iceedgenearourstudyareawasusuallywell-defined(Fig.2A). (Polashenski,2011). WhilemuchofthesouthernChukchiSeawasice-freeinmid- June2011,ourstudyareawasstillcoveredbyseaice(Fig.2A).By 3.2. Phytoplankton thetimewestartedsamplingTransect1(stations46–71)on3July 2011, ice had retreated northwestward and the southern half of From4–8July2011,wesampledthefull40–150mdeepwater Transect1wasinopenwater.SamplingalongTransect2(Stations column along Transects 1 and 2, both of which extended from 57–71)endedon8Julyandbythistime,afewmorestationsalong openwaterto 4100kmintotheicepack(Fig.2A).Atthetimeof Transect 1 were in open water, as were the six southernmost sampling,thesoutheasternendofTransect1(station46)hadbeen K.R.Arrigoetal./Deep-SeaResearchII105(2014)1–16 7 Table1 Depth-integrated phytoplankton biomass along Transect 2 was Icethicknessesandmeltponddepthsaatseaicestationsassociatedwithunder-ice even higherthan onTransect1, rangingfrom921 to 1292mg Chl phytoplanktonbloom. a m(cid:1)2 beneath the ice at stations 59–61 (Fig. 6B). The under-ice bloom along Transect 2 extended for (cid:3)133km between the ice Station Allice(m) Bareice(m) Pondedice(m) Meltpond depth(m) edgeanditsterminationattheshelfbreak. Mean N Mean N Mean N Mean N 3.3. Lighttransmissionthroughseaice 55 0.7770.09 113 0.8170.07 73 0.6970.06 40 0.0570.03 20 56 1.0370.15 198 1.1070.12 138 0.8870.07 60 0.1170.04 59 Vertical light profiles taken through the sea ice at stations 57 1.2170.10 333 1.2670.06 239 1.0870.05 94 0.1270.04 97 55–57showedthatbareicetransmittedapproximately12.7–17.5% ofincidentsurfacelighttothewatercolumnbelow(Table2).Due aMean7standarddeviation. tothelackofsignificantsnowcover,lightattenuationinbareice wasdominatedbythepresenceofahighlyscatteringsurfacelayer consistingofdrainedicecrystals.Incontrast,theicebeneathmelt ice-free for two weeks (Fig. 5A) and was characterized by chlor- pondslackedthesurfacescatteringlayerandwasthinnerthanthe ophylla(Chla)concentrationsthatwerelow(0.38mgL(cid:1)1)within bare ice (Fig. 4B). Hence, ponded ice transmitted approximately the 25m upper mixed layer and increased to 2.5mgL(cid:1)1 toward 46.7–58.6% of incident surface light to the underlying water the sea floor (Fig. 5B). Moving northwest along Transect 1 from column,4-foldmorethanbareice(Table2). open water toward the ice edge, a prominent subsurface Chl a Interestingly, the euphotic depth (0.1% of visible incident sur- maximum (SCM) that tracked the shoaling MLD (Fig. 5C) was face light) associated with both bare and ponded ice at a given increasinglyapparent,withpeakconcentrationsintheSCMrising stationwasidentical(Table2).Thisisbecauselighttransmittedto from5.4mgL(cid:1)1atstation48to12.9mgL(cid:1)1atstation51.TheSCM the upper ocean through melt ponds scattersintoadjacent areas and the MLD became progressively shallower with proximity to overlain by bare ice (Fig. 4B). Consequently, with increasing theice edge.TotalwatercolumnChla alsoincreasedtoward the distancebelowtheundersideoftheice,theunderwaterlightfield iceedge,from43.6mgm(cid:1)2atstation46to250mgm(cid:1)2atstation became a more uniform mix of light transmitted through both 51(Fig.6A). bareiceandmeltponds(Freyetal.,2011). Nearstation52,weenteredtheSIZ,whichhadawell-defined Euphoticdepthdidvarybymorethanafactorofthreebetween ice edge coinciding with the westward limit of the warm water stations (Table 2) as a result of changes in light attenuation by (Fig. 5D) being advected northward in the upper portion of the particulate and dissolved materials within thewatercolumn. For shelfbreakjet(Fig.5E)andaMLDofapproximately10m(Fig.5C). example, beam attenuation at 488nm increased 6-fold from Althoughweexpectedthatreducedlightavailabilitybeneaththe 0.26m(cid:1)1 in surface waters outside the ice pack (station 46) to 0.7770.09m thick first-year ice at the edge of the SIZ would 1.54m(cid:1)1withintheunder-icephytoplanktonbloom(station56), result in lower phytoplankton abundance, water column Chl a due to large increases in both particle scattering (5-fold) and continued to increase as we traveled further into the ice pack particle absorption (3.5-fold). Light absorption within the bloom (Fig. 2B). Between stations 52 and 56, depth-integrated phyto- wasalmostentirelyattributable to particles (91%), mainlyphyto- planktonbiomassbeneaththeseaicerosefrom212mgChlam(cid:1)2 plankton (78%), and to a lesser extent by CDOM. The patterns in (8.8gC,m(cid:1)2)toaremarkable1016mgChlam(cid:1)2(28.7gCm(cid:1)2) particle absorption and scattering are consistent with observed at station 56 (Fig. 6A), despite sea ice cover of 100% and an differences in particle number (0.9–1.8(cid:4)1011m(cid:1)3), particle dry increaseinseaicethicknessto1.0370.15m(Table1).Mostofthis mass(0.4–2.7mgL(cid:1)1),andparticlemediandiameter(2.2–3.6mm) phytoplankton biomass was within the upper 30m of the water betweenstations(moredetailscanbefoundinNeukermansetal., column, with the highest Chl a concentrations near the ice/ 2014). Due to the strong light attenuation within the water seawater interface (Fig. 5B). The depth over which this bloom column, the euphotic depth shoaled considerably within the extended(30–70m)andthehighdepthintegratedphytoplankton under-ice phytoplankton bloom, decreasing from 46m at station biomassdistinguishesthisunder-icebloomfromthemuchsmaller 46(inopenwater)toonly (cid:3)10matstation56. bloomsthathavebeenobservedpreviouslyinshallowmeltwater lensesbeneaththeicethatonlyextend (cid:3)1mbelowtheicewater interface (Gradinger, 1996; Spilling, 2007). The ship’s underway 3.4. Dissolvednutrientsandgases fluorometershowedthatthismassivediatombloomcontinuedfor another32kmnorthofstation56beforeterminatingjustnorthof Along Transect 1, the nitracline (Fig. 5C) largely tracked the the shelf break near station 57 (Fig. 2B). In total, the under-ice depthofthemixedlayer(Fig.5C)andwasdeepestinopenwater phytoplankton bloom along Transect 1 extended for (cid:3)105km southeast of the sea ice edge, extending to (cid:3)30m at station 46. from the ice edge to the shelf break. Diatoms were the most Surface water NO concentrations at this stationwere belowour 3 abundantphytoplanktongroupbeneaththeice,includingspecies detection limit (0.01mmol L(cid:1)1), increasing to 1.65mmolL(cid:1)1 at a inthegeneraChaetoceros,Thalassiosira,andFragilariopsis,andata depthof27mandreachingamaximumof10.5mmolL(cid:1)1nearthe few stations, Odontella (more details can be found in Laney and seafloor(44m).Thenitraclineshoaledasweapproachedtheice, Sosik.,2014). reaching (cid:3)10mattheiceedge(station52),withNO concentra- 3 Transect 2, which also terminated north of the shelf break at tions of 0.3mmolL(cid:1)1 at the surface and 4.5mmolL(cid:1)1 at 11m. station57,wassampledatahigherspatialresolutionwithintheSIZ Near-bottom NO was also higher at this station (18mmolL(cid:1)1) 3 (Fig.2A)andexhibitedaspatialpatterninphytoplanktonbiomass than they were at station 46. The nitracline was shallowest that was almost identical to Transect 1. Starting in open water beneath the sea ice, reaching its minimal depth ((cid:3)7m) near southeast of the SIZ (station 71, Fig. 5H), the SCM was located station 55 (Fig. 5C). The nitracline along Transect 2 exhibited a progressively closer to the surface with proximity to the sea ice similarspatialpatterntoTransect1,beingdeepestinopenwater edge (station 66, Fig. 5I) and depth-integrated Chl a increased southeast of the sea ice edge. Unlike Transect 1, however, the dramaticallyfrom54mgm(cid:1)2to576mgm(cid:1)2(Fig.6B).Beneaththe nitricline depth (25–30m) exceeded the MLD ((cid:3)10m) in open ice,phytoplanktonbiomassincreasedwithdistanceinfromtheice water,indicatingthattheMLDhadshoaledafterthenitriclinehad edge despite increasing ice thicknesses (to 1.2170.10m, Table 1). formed (Fig. 5J). Maximum nitracline depths ((cid:3)40m) were also 8 K.R.Arrigoetal./Deep-SeaResearchII105(2014)1–16 100 100 3 July 3 July 50 Sea ice Sea ice 50 4 July 4 July 0 concentration (%) concentration (%) 0 57 56 55 52 50 48 46 57 56 55 52 50 48 46 0 30 20 25 15 50 20 m) h ( 15 10 pt e D 100 10 5 Transect 1 5 Transect 1 Chlorophyll a (µg L-1) Nitrate (µmol L-1) 150 0 0 0 4 20 3 50 10 2 m) h ( 1 0 ept 100 0 D -10 Transect 1 -1 Transect 1 Potential temperature (°C) Cross-track velocity (cm s-1) 150 -2 -20 0 700 2200 2150 600 m) 50 2100 h ( 500 2050 pt e D 100 2000 400 Transect 1 Transect 1 1950 Oxygen (µmol L-1) DIC (µmol L-1) 150 300 1900 250 200 150 100 50 0 250 200 150 100 50 0 Distance along transect (km) Distance along transect (km) 100 100 50 Sea ice 7 July Sea ice 7 July 50 8 July 8 July 0 concentration (%) concentration (%) 0 57 59 61 63 65 67 69 71 57 59 61 63 65 67 69 71 0 30 20 25 m) 50 20 15 h ( 15 pt 10 e D 100 10 5 Transect 2 5 Transect 2 Chlorophyll a (µg L-1) Nitrate (µmol L-1) 150 0 0 200 150 100 50 0 200 150 100 50 0 Distance along transect (km) Distance along transect (km) Fig.5. HydrographicsectionsfromICESCAPE2011showingunder-icephytoplanktonbloomcharacteristics.(A–G)arefromTransect1andshow(A)seaiceconcentration, (B)chlorophylla,(C)nitrate,(D)potentialtemperature(θ),(E)cross-trackvelocity(positivevaluesdenotenortheastwardflow),(F)dissolvedoxygen,and(G)dissolved inorganiccarbon(DIC).(H–J)arefromTransect2andshowconcentrationsof(H)seaice,(I),chlorophylla,and(J)nitrate.Stationnumbersareaboveeachpanelandblack dotsrepresentsamplingdepths.Solidcontourlinesin(B),(D),and(E)arepotentialdensity(sθ).Theblacklinein(C)and(J)denotesthedepthofthemixedlayer. somewhatgreateralong Transect 2than onTransect1,with NO withproximitytotheiceedgeandwasshallowestundertheice, 3 being completely consumed (i.e. below the detection limit) to a althoughnitraclinedepthsweremorespatiallyvariablethanalong depthof26matstations69and70.Thenitraclineagainshoaled Transect1. K.R.Arrigoetal./Deep-SeaResearchII105(2014)1–16 9 Table2 1200 TUrnadnesr-eiccet 1 30000 -2m) Oblpotoicma.lcharacteristicsaatseaicestationsassociatedwithunder-icephytoplankton bloom C -2)mg m 1800000 2205000000 on (mg Station Icethickness(m) LBiagrhetitcreansmPitotnadneced(i%ce) EBuarpehoicteicdepPthon(mde)dice b a ( ar 55 0.7770.09 17.570.7 53.871.3 17.570.06 17.970.40 yll 600 15000 c c 56 1.0370.15 12.770.5 46.771.3 10.070.03 11.070.03 h ni 57 1.2170.10 13.670.7 58.671.2 32.470.43 33.271.01 p a o 400 10000 g or or Euphoticdepthincludesthicknessofoverlyingseaice. Chl 200 5000 ate Threelightcastsweredoneateachstation. ul aMean7standarddeviation c 0 0 arti 58 56 54 52 50 48 46 P Fv/Fm 0.0 0.1 0.2 0.3 0.4 0.5 0.6 Transect 2 ) 0 1500 35000 -2m Station 55 Under-ice C 10 a-2hyll (mg m) 1692000000 bloom 12235050000000000000 nic carbon (mg Depth (m) 234000 SSttaattiioonn 5662 p a 50 o g or 10000 or hl 300 e 60 C 5000 at ul 70 c 0 0 arti Fig. 7. Depth profiles of the efficiency of electron flow through photosystem II 57 59 61 63 65 67 69 71 P (Fv/Fm)withintheunder-icephytoplanktonbloom.Allvaluesexceed0.4,suggesting Station number thatphytoplanktonwerephysiologicallycompetentthroughoutthewatercolumn. the nitracline, the depth of maximum O concentration shoaled 1 2 with proximity tothe ice edge.Beneaththe ice pack, O concen- mean = 46.4 2 trations were highest in the upper (cid:3)25m of the water column, mode = 20-40 reachingvaluesashighas583.1mmolL(cid:1)1atthesurfaceofstation n = 12,048 on 0.1 61. This station also exhibited enhanced O2 concentrations down cti toadepthofapproximately75m. un Dissolved inorganic carbon (DIC) concentrations in the deep y f samples along Transect 1 varied from approximately 2250 to sit 2303mmolL(cid:1)1. The DIC depletion pattern (Fig. 5G) strongly n 0.01 mean of gray max of gray y de area in B area in B sreosuetmhebalsetdothfatthoefNseOa3,icweitehdlgoew(D2I0C6v4a–l2u0e9s8inmmopoelnLs(cid:1)u1rfaatcestwataitoenrss bilit 46–48)andsubstantialdepletiondowntoatleast30m.Thedepth a b ofDICdepletiondiminishedtowardtheiceedge,eventuallyshoaling o 0.001 Pr to (cid:3)10m at ice edge station 52. There was less DIC depletion in waters beneath the sea ice (except station 57), with surface DIC valuesreducedto2105–2190mmolL(cid:1)1atstations52–55. 0.0001 3.5. Phytoplanktonphotophysiologyandgrowthrate 0 200 400 600 800 1000 1200 1400 Depth-integrated chlorophyll a (mg m-2) Microscopic analysis indicated that under-ice phytoplankton Fig.6. Depth-integratedchlorophylla(diamonds)andparticulateorganiccarbon at all stations were healthy, with few observations of empty (squares) along (A) Transect 1 and (B) Transect 2. Gray areas denote samples or senescing diatom frustules. This conclusion is supported by collectedwithintheunder-icephytoplanktonbloom.Probabilitydensityfunction observationsofgenerallyhighefficiencyofelectronflowthrough (C) for depth integrated water column Chl a from the global SeaBASS pigment photosystem II (Fv/Fm), which exceeded 0.4 at all depths shal- database(http://seabass.gsfc.nasa.gov/).Notelogscaleofy-axis.Themeandepth- integratedbiomassintheunder-icebloomalongTransect2ranksinthetop0.1%of lower than 70m and exceeded 0.5 in 12 out of the 15 samples globalvalues.Depth-integratedChlaatstation60ishigherthananyvalueinthe whereFv/Fmintheunder-icebloomwasmeasured(Fig.7). database. Photosyntheticparametersdeterminedforphytoplanktonsampled atthesurfaceandat (cid:3)25m(theapproximate depthof theSCMin Dissolved oxygen (O ) concentrations were very high along openwater)atagivenstationwerenotsignificantlydifferent(t-test, 2 bothTransect1(Fig.5F)andTransect2(notshown).Atstation47, p40.05)acrossourstudyregion(Table3).However,photosynthetic locatedsoutheastoftheiceedge,O concentrationswerehighest parameters of under-ice phytoplankton differed significantly from 2 at a depth of 25m (466.1mmolL(cid:1)1), although there was still those of phytoplankton in nearby open water. The maximum Chl significant supersaturation in shallower waters. Like Chl a and a-specificphotosyntheticrates(Pn)withintheunder-icephytoplankton m 10 K.R.Arrigoetal./Deep-SeaResearchII105(2014)1–16 bloom(1.4070.33mgCmg(cid:1)1Chlahr(cid:1)1)weredoublethoseoftheir Table3 openwatercounterparts(0.7170.14mgCmg(cid:1)1Chlahr(cid:1)1),despite PhotosyntheticParametersofPhytoplanktonatthesurfaceandtheSCMfromthe growing beneath heavysea ice cover (t-test, po0.05). Similarly, the under-icebloom(bloom)andadjacentopenwater(non-bloom). plahrgoetotshyantthoefttihceeirffiocpieenncwy(aαten)rocfouunntdeerrp-aicretsp(ht-ytteospt,lapnokt0o.n05w),aasvtewraicgeinags Pnm αn Ek m 0.02470.007 and 0.01270.004mg C mg(cid:1)1 Chl ahr(cid:1)1 (mmol Surface photonsm(cid:1)2s(cid:1)1)(cid:1)1,respectively(Table3).Themeanphotoacclima- Bloom 1.35(0.30) 0.021(0.006) 67.6(22.4) 0.85(0.47) tionparameter(E)fortheunder-icebloom(61.2718.0mmolphotons Non-bloom 0.68(0.11) 0.011(0.003) 67.9(25.1) 0.05(0.02) m(cid:1)2s(cid:1)1) was nokt significantly different (t-test, p40.05) from that SCM Bloom 1.45(0.38) 0.027(0.007) 54.9(10.0) 0.92(0.30) of phytoplankton growing in open waters (66.0726.6mmol Non-bloom 0.74(0.17) 0.013(0.005) 64.0(31.9) 0.24(0.23) photonsm(cid:1)2s(cid:1)1). The high Pn, high αn, and low E indicate that All m k the under-ice phytoplankton were well adapted to grow at the Bloom 1.40(0.33) 0.024(0.007) 61.2(18.0) 0.89(0.39) Non-bloom 0.71(0.14) 0.012(0.004) 66.0(26.6) 0.15(0.18) reducedlightlevelsencounteredbelowtheArcticicepack. The largest physiological difference between the under-ice aMean7standarddeviation. phytoplanktonandtheiropenwatercounterpartsistheirmaximum Pnm–mgCmg(cid:1)1Chlahr(cid:1)1,αn–mgCmg(cid:1)1Chlahr(cid:1)1(mmolphotonsm(cid:1)2s(cid:1)1)(cid:1)1, biomass-specific growth rate, calculated from both short-term Ek–mmolphotonsm(cid:1)2s(cid:1)1,m–d(cid:1)1. estimates of P–E parameters and from 24-h NPP measurements. Bloomstations:52,54,55,56,58,60,62,and64.Non-bloomstations:46,48,57,and70. Light-saturated growth rates (near the ice/water interface) within theunder-icebloomalongTransect1wereextraordinarilyhigh,rang- is further supported by the coincidence of extraordinarily high ingfrom0.86d(cid:1)1neartheiceedgeto1.59d(cid:1)1atthehigh-biomass algalbiomassandlargenutrientdeficitsintheupper25–30mof station56.Light-saturatedgrowthratesalongTransect2weresome- the water column beneath the ice, which can only be the result whatlowerbutstillveryhigh,varyingfrom0.21–1.20d(cid:1)1.Fiveof of nutrient uptake by phytoplankton growing within the water the eight stations associated with the under-ice bloom exhibited column. Sea ice algae simply cannot deplete nutrients from that phytoplankton light-saturated growth rates 41.0d(cid:1)1, which are farbelowtheice/waterinterface.Itispossiblethatalgaereleased muchhigherthanexpectedforpolarwatersnearthefreezingpoint from the sea ice may have helped to seed the under-ice phyto- of (cid:1)1.81C(0.53d(cid:1)1,Eppley,1972). planktonbloom,althoughwemeasuredverylowicealgalbiomass The mean light-saturated growth rate within the under-ice duringourstudy. bloom was 0.89d(cid:1)1 (Table 3), approximately 6-fold higher than The most intense phytoplankton blooms previously reported the mean growth rate of phytoplankton collected in open water inpolarwatersweregenerallyassociatedwithopenwatersalong (0.15d(cid:1)1).ThislargedifferenceisdueinparttotheelevatedPOC theMIZ(AlexanderandNiebauer,1981;Sakshaug,2004;Perrette concentration associated with a higher detrital content in the et al., 2011), where meltwater-induced surface stratification cre- older open water phytoplankton blooms (m¼NPP/POC or Pn(cid:5)Chl ateslightconditionsthatarefavorableforphytoplanktongrowth. a/POC).However,itisalsoareflectionofthefactthatphytoplankton Occasionally, these MIZ blooms can become ice covered if either growinginopenwatershadalreadyconsumedmuchoftheNO in seaiceadvancesbackintowaterswhereaMIZbloomhasalready 3 near surface waters by the time of our cruise and were concen- developed(Mundyetal.,2009)orlocalcirculationadvectstheMIZ tratedatthesub-surfacenitraclinewherebothnutrientsandlight bloom beneath the ice pack. As a result, it is important to levels are relatively low. In contrast, phytoplankton growing distinguish ice-covered MIZ blooms from phytoplankton blooms beneaththeicehadnotyetconsumedalloftheavailablenutrients thatdevelopentirelybeneaththeseaice.Onthebasisofageand fromwithin the euphotic zone and were able to maintain much location of the bloom relative to both the ice edge and bottom highergrowthrates. topography,andnutrientandbiogenicgasconcentrationsbeneath the ice and in the adjacent open water, we determined that the under-ice phytoplankton bloom we observed during ICESCAPE 4. Discussion 2011wasneitheraresidualMIZbloomnoranopenwaterbloom thathadbeentransportedbeneaththeice. 4.1. Phytoplanktonundertheice First,satelliteobservationsshowthattheseaiceintheChukchi Sea consistently retreated in a northwestward direction prior to It is noteworthy that the mean biomass associated with the oursamplingthebloom,rulingoutthepossibilitythaticehadever under-ice phytoplankton bloom we observed during ICESCAPE reversed direction and drifted over a previously developed MIZ 2011 was comparable to depth-integrated values from the most bloom.Second,phytoplanktonbiomassassociatedwiththeunder- productive pelagic ecosystems in the global ocean. In fact, the icebloomalongbothTransect1and2reacheditsmaximumvalue maximum depth-integrated biomass we measured at station 60 approximately 120km from the ice edge at a location where was higher than any of the 12,048 open ocean values currently nutrientswereenhancedduetoshelfbreakupwellingfromdepth contained in the SeaBASS pigment database managed by NASA (see below). This bathymetric constraint strongly supports the (Fig.6C).Althoughthesehighbiomasslevelswereobserveddeep conclusion that the bloom developed beneath the ice, with its within the ice pack under relatively thick (0.8–1.2m) ice, it is northwesternboundary(theboundarylocateddeepestwithinthe importanttodeterminewhetherthisbloom(1)originatedasasea icepack)controlledbythepositionoftheshelfbreak. ice microalgal bloom that was subsequently released into the Finally, if the bloom had advected beneath the ice from open water column, (2) advected beneath the ice from the site of an water, it would have been in a much later stage of development earlierMIZbloom,or(3)developedexclusivelybeneaththeice. thanweobserved,giventhehighgrowthrateofthephytoplank- The majority of the stations within the under-ice the bloom tonbeneaththeiceandthelocalcirculationpatterns.Theshortest were overwhelmingly (480% bycell cross-sectional area) domi- route from ice-free waters to our study region beneath the ice natedbypelagic diatomsof the genera Chaetoceros,Thalassiosira, would require advection of the bloom to the northwest in the and Fragilariopsis, indicating that this was not a remnant sea ice direction of sea ice motion. This is unlikely since no known algalbloomthathadsloughedoffintothewatercolumn,although currentsinthevicinityofourstudyregionflowfromthesoutheast the diatom Odontella aurita was abundant at station 56 and is to the northwest. Instead, one current flows northward through occasionallyfoundinseaice(McMinnetal.,2008).Thisconclusion Central Channel then northeastward around Hanna Shoal while
Description: