Domain Editor-in-Chief: Jody W. Deming; Department of Biological Oceanography, University of Washington, Seattle, Washington, United States
Associate Editor: Jean-Éric Tremblay; Department of Biology, Université Laval, Québec, Québec, Canada


Introduction

Antarctic shelf waters are strong sinks for atmospheric CO2 due to high biological productivity, intense winds, high air-sea gas exchange, formation of bottom water and extensive winter ice cover. These factors make these regions important for the biogeochemical cycling of elements, particularly of carbon (C) (Sarmiento et al., 2004; Arrigo et al., 2008). Specifically, coastal polynyas (areas of open water surrounded by ice) are hot spots for energy and C transfer between the atmosphere and ocean (Smith and Barber, 2007, Mu et al., 2014). The reduced ice cover increases air-sea gas exchange and results in enhanced light availability in the water column in early spring, thereby increasing primary productivity through phytoplankton photosynthesis. In addition to its importance for the global C cycle, phytoplankton productivity supports the biota occupying higher trophic levels including krill, penguins, and whales (Arrigo et al., 2003; Ainley et al., 2006).

Phytoplankton productivity in the Southern Ocean is often limited by the availability of iron (Fe) (Boyd et al., 2007, and references therein), although light limitation may also occur due to deep vertical mixing below the critical depth (Mitchell et al., 1991; De Baar et al., 2005). The Fe supply for phytoplankton growth in polynyas is enhanced compared to the open ocean due to input of dissolved Fe (DFe) from melting sea ice (Sedwick and DiTullio, 1997; Lannuzel et al., 2010), shelf sediments (Hatta et al., 2013; De Jong et al., 2013), icebergs (Raiswell et al., 2008; Raiswell, 2011; Shaw et al., 2011), upwelling Circumpolar Deep Water (CDW) (Klunder et al., 2011), remineralized Fe in winter water (WW), and melting glaciers (Raiswell et al., 2006; Gerringa et al., 2012, Sherrell et al., 2015). Despite these sources, phytoplankton growth is often still seasonally Fe-limited in later stages of blooms in polynyas such as in the Ross Sea Polynya (Sedwick and DiTullio, 1997; Sedwick et al., 2000; Tagliabue and Arrigo, 2005) and the Weddell Sea (Buma et al., 1991). Bioassay experiments in the Ross Sea Polynya revealed that phytoplankton growth was Fe-limited later in the season, but not early (Sedwick and DiTullio, 1997; Sedwick et al., 2000). It is believed that phytoplankton blooms gradually draw down a “winter stock” of DFe in the WW that eventually limits the bloom, especially away from DFe sources such as sea ice (Hopkinson et al., 2013). However, recent measurements show low and potentially limiting concentrations of DFe early in the season (Sedwick et al., 2011; Marsay et al 2014), implying that significant sources of new DFe are required to sustain a phytoplankton bloom throughout the season.

Satellite estimates of primary productivity in Antarctic polynyas reveal the highest productivity per surface area in the Amundsen Sea (Arrigo and Van Dijken, 2003). The Amundsen Sea contains two polynyas, the Pine Island Polynya (PIP) in the east and the Amundsen Sea Polynya (ASP) in the west. The Amundsen Sea is located in the western Antarctic, where rates of ice sheet thinning are the highest in all of Antarctica (Pritchard et al., 2009; Rignot et al., 2013). Several fast-flowing glaciers that are rapidly thinning drain into the Amundsen Sea where they form floating ice shelves or glacier tongues that are also thinning (Randall-Goodwin et al., 2015). These are the Pine Island Glacier in the PIP and the Getz and Dotson glaciers in the ASP. The Thwaites and Crosson glaciers form ice tongues between the PIP and ASP and may affect both polynyas (Pritchard et al., 2009; Rignot et al., 2013).

The thinning of the ice shelves is mainly attributed to regional bathymetry and oceanography. As the Antarctic Circumpolar Current (ACC) flows close to the continental shelf break, Circumpolar Deep Water (CDW) intrudes southward through deep troughs onto the Antarctic continental shelf (Jacobs et al., 1996, 2011; Arneborg et al., 2012). On the shelf, the CDW mixes with WW and becomes modified CDW (mCDW) that is warm (> 0.6°C) and salty (> 34.5) relative to WW. Near the coast, mCDW has access to ice shelf cavities and drives basal melting of floating ice shelves (Jenkins et al., 2010; Jacobs et al., 2011, 2013; Dutrieux et al., 2014). The resulting mCDW mixed with glacial melt water (meltwater-laden mCDW) becomes fresher (< 34.0), colder (−1.1 to −0.5°C), and more buoyant. At the surface of the ASP, Antarctic Surface Water (AASW) shows a range in salinity (33.6 to 34.1) and temperature (−1.8 to > 0°C), depending on length of time since sea ice melt, degree of solar warming, and wind- or buoyancy-induced mixing with the underlying waters (Yager et al., 2012; Ha et al., 2014). The meltwater-laden mCDW outflowing at subsurface depths from under the Dotson ice shelf (DIS) appears to be a major source of DFe and Particulate Fe (PFe) to the phytoplankton bloom in the ASP (Gerringa et al., 2012; Sherrell et al., 2015). This subsurface source is likely made available to AASW at the surface through horizontal diffusivity (Gerringa et al., 2012), advective eddy transport (e.g., Årthun et al., 2013), mixing along the Dotson trough (e.g., St-Laurent et al., 2013), and by wind- and iceberg-induced mixing (Randall-Goodwin et al., 2015). Similarly, the meltwater-laden mCDW from the Pine Island Glacier was found to be a major source of DFe for the phytoplankton bloom in the PIP (Gerringa et al., 2012). These Fe sources support phytoplankton blooms with high biomass and productivity in both the PIP and ASP (Alderkamp et al., 2012a; Yager et al., 2012). Moreover, Fe addition bioassay experiments at the peak and during the decline of the phytoplankton bloom revealed that Fe was not limiting phytoplankton growth in either polynya at these times (Mills et al., 2012), suggesting relatively high Fe availability to the phytoplankton.

Light availability for phytoplankton in Antarctic polynyas is temporally and spatially variable. Early in the season, the shrinking sea ice cover, in combination with increasing day length and solar elevation, results in greater light availability in surface waters. After the polynya opens up, light availability is determined by cloud cover, mixed layer depth (MLD), and attenuation in the water column that is controlled primarily by phytoplankton biomass. In general, melting sea ice introduces fresh water at the surface, which stabilizes the water column and creates a well-lit shallow mixed layer. Conversely, basal ice shelf melt introduces meltwater-laden mCDW into the water column at the base of the ice shelf, at 150-400 m depth, depending both on the draft of the ice shelf (Jacobs et al., 2012; Mankoff et al., 2012) and the degree of buoyancy driven upwelling. Introducing buoyant water at depth destabilizes the water column and may increase MLD and decrease light availability to the phytoplankton, as was observed at the face of the Pine Island ice shelf (Alderkamp et al., 2012a).

Phytoplankton photoacclimate to low light by increasing their cellular pigment concentration to maximize the capture of photons (Falkowski and LaRoche, 1991; MacIntyre et al., 2002). However, Fe availability affects photoacclimation because biosynthesis of pigments requires Fe and the photosynthetic apparatus has a high Fe content (Raven, 1990; Greene et al., 1992). Thus, phytoplankton in low Fe regions may be impaired in their ability to photoacclimate (Greene et al., 1992; Vassiliev et al., 1995) or may have adapted mechanisms to increase the capture of photons without increasing Fe requirements (Strzepek and Harrison, 2004; Strzepek et al., 2012).

In this study, we investigated the effects of Fe additions on phytoplankton growth and photosynthesis rates during the build-up phase of a phytoplankton bloom in the ASP and assessed how Fe availability affected phytoplankton photoacclimation. We measured phytoplankton biomass, photosynthetic parameters, and productivity rates in surface waters of the ASP (Figure 1) and performed Fe addition bioassay experiments throughout the ASP to test effects of Fe additions on these parameters. In addition, effects of Fe addition on bacterial productivity were evaluated in the experiments. Moreover, interactions between Fe and light availability were tested in the experiments by including incubations at different light levels. Experimental results were then used to draw conclusions about Fe limitation of phytoplankton photosynthesis and productivity in the ASP.

doi: 10.12952/journal.elementa.000043.f001.
Figure 1.  

ASPIRE stations in the Amundsen Sea Polynya.

Hydrographic stations sampled during the ASPIRE cruise are projected on the Amundsen Sea Polynya (ASP) bathymetry. Fe addition bioassay experiments were performed at red stations; the water column was sampled at all stations. The dashed line shows the sea ice edge on 1 January 2011; the Getz Ice Shelf (GIS), Dotson Ice Shelf (DIS), and Twaites Ice Tongue are shown in white. Inset shows the location of the ASP in Antarctica.

Methods

Sampling

Seawater samples were collected during the NBP 10-05 cruise on the RVIB Nathaniel B. Palmer in the ASP during the austral spring and summer, 13 December 2010 to 12 January 2011. Water for analysis of parameters listed in the section on Analytical methods was sampled from discrete depths in the upper 300 m of the water column at 22 stations (Figure 1) during the middle of the day within three hours of solar noon. Water was sampled with trace metal clean (TMC) techniques using externally-closing 12 L Niskin bottles (Model 110BES, Ocean Test Equipment, Ft. Lauderdale, FL, USA) mounted on a GEOTRACES-style non-contaminating CTD-Rosette deployed on a coated aramid cable (see Sherrell et al., 2015, for details). Continuous vertical profiles of temperature, salinity, irradiance, fluorescence, and suspended particle abundance were obtained from the water column using a SeaBird 911+ CTD, a Chelsea fluorometer, photosynthetically active radiation (PAR) sensor (Biospherical), and a 25-cm WETLabs transmissometer, respectively, mounted on a TMC rosette.

Fe addition bioassay experiments

At five stations (Figure 1), large volumes of seawater (∼ 100 L) were collected at select depths (see below) within the upper 50 m of the water column for bioassay experiments to study the effects of Fe addition under different light conditions, using TMC techniques throughout the experiments. Acid-washed polycarbonate bottles (2 L) were rinsed three times with MilliQ water and once with seawater from the same station before being filled to the brim with unfiltered seawater. Triplicate bottles for each treatment were incubated at in-situ water temperature in transparent deck incubators under incident irradiance shaded with different levels of neutral transmission screening. FeCl3 was added to the Fe treatments from a 1000x stock in weakly acidified, 0.2 µm filtered seawater, in a final concentration of 4 nmol L−1 (Mills et al. 2012). Nothing was added to the control treatments. Bottles were capped and caps were wrapped with parafilm to prevent contamination by water from the incubator. All treatments were sampled within three hours of solar noon at the beginning of the experiment (T = 0 days), at 4 days, and at 7 or 8 days for all parameters listed in the section on Analytical methods. At three of the five stations (Stations 5, 13, and 35), parallel experiments were conducted with water from two different depths, the surface (S: 8–12 m) and subsurface (deep, D: 35–50 m). Water from each depth was incubated at 10% of incident irradiance. At the two remaining stations (Stations 57 and 66), only surface water was incubated in three parallel experiments under different light conditions: high (50%), medium (10%), and low (1%) incident irradiance.

Analytical methods

Nutrients and iron concentrations

Concentrations of nitrate (NO3), ammonium (NH4), phosphate (PO4) and silicate (Si(OH)4) from seawater samples at 22 stations and bioassay experiments were determined by flow injection analysis using a Lachat Instruments Quickchem 8000 Autoanalyzer according to standard protocols (see Vernet et al., 2011). Samples were collected directly from the experimental bottles, filtered through 0.2 µm Acrodisc® filters and stored at 4°C until analysis on the same day.

Concentrations of DFe from seawater samples at 16 stations and T = 0 samples of the bioassay experiments were determined by preconcentration and isotope dilution ICP-MS and are described in full in Sherrell et al. (2015). Concentrations of particulate Fe (PFe) from seawater samples at 11 stations and T = 0 samples of the bioassay experiments were determined by ICP-MS by methods following Planquette and Sherrell (2012) and have been reported elsewhere (Harazin et al., 2014).

Particulate organic carbon (POC) and particulate organic nitrogen (PON)

Duplicate samples from the bioassay experiments (100–1000 ml) were filtered onto precombusted (450°C for 4 h) 25 mm Whatman GF/F filters and dried at 60°C for analysis of POC and PON on a Carlo-Erba NA-1500 elemental analyzer using acetanilide as a calibration standard.

The POC concentrations at T = 0 and at 4 days were used to calculate phytoplankton growth rates (µ: d−1) in the bioassay experiments using the equation:

POCT4=POCT0eμT.
(1)

Pigment analysis

Seawater samples at 19 stations and bioassay experiment samples (50–500 mL) for chlorophyll a (Chl a) were filtered onto 25 mm Whatman GF/F filters, extracted overnight at 4°C in 5 mL of 90% acetone, and analyzed on a Turner Model 10AU fluorometer before and after acidification (Holm-Hansen et al., 1965).

The full pigment composition was analyzed for seawater samples at 12 stations and T = 0 samples of the bioassay experiments by High Performance Liquid Chromatography (HPLC). Samples (100–2000 mL) were filtered onto 25 mm Whatman GF/F filters, flash-frozen in liquid N2, and stored at −80°C until analysis. Filters were extracted for two hours in 98% methanol: 2% ammonium acetate [vol:vol] in the dark at −20°C after disruption by sonication. Pigments were separated on a SPD-M10AVP HPLC system (Shimadzu, Inc.) using a Agilent 4.6 x 250 mm C18 column kept at 30°C according to Wright et al. (1991), using standards for Chl a, chlorophyll b (Chl b), chlorophyll c3 (Chl c3), peridinin (Per), 19′-butanoyloxylfucoxanthin (19′-But), fucoxanthin (Fuc), 19′-hexanoyloxyfucoxanthin (19′-Hex), neoxanthin (Neo), prasinoxanthin (Pras), violaxanthin (Viol), alloxanthin (Allo), lutein (Lut), antheraxanthin (Anth), diadinoxanthin (Diad), diatoxanthin (Diat), and -carotene (-Car). The Chl a breakdown product chlorophillide a was detected in few samples at < 2% of Chl a concentrations.

The ratios of the first 12 of these pigments were used to determine the phytoplankton class abundance using the CHEMTAX analysis package, version 1.95 (Mackey et al., 1996; Wright et al., 1996). The initial input ratios (Table 1A) consisted of specific pigment ratios for eight phytoplankton classes that generally dominate Antarctic waters (Wright et al., 2010), including prasinophytes, chlorophytes, cryptophytes, diatoms (with a separate class for Chl c3 containing diatoms such as Pseudonitzschia), dinoflagellates, and two classes of P. antarctica. The two classes of P. antarctica account for variations in pigment ratios between strains (Zapata et al., 2004) and changes in response to Fe limitation (Van Leeuwe and Stefels, 2007, DiTullio et al., 2007, Alderkamp et al., 2012b). The pigment ratios in the output matrix (Table 1B) were within those reported in the literature (Zapata et al., 2004, Van Leeuwe and Stefels, 2007, DiTullio et al., 2007, Wright et al., 2010, Alderkamp et al., 2012b). Diatoms and Pseudonitzschia are presented together as diatoms and the two P. antarctica classes are presented together as P. antarctica.

10.12952/journal.elementa.000043.t001a

Table 1A.

Initial pigment: Chlorophyll a (Chl a) ratios used in the CHEMTAX analysis of pigment dataa

Phytoplankton class Pigment used in ratio with Chl a
Chl c3 Lut Per 19′-But Fuc 19′-Hex Neo Pras Allo Viol Chl b
Prasinophytes 0 0.006 0 0 0 0 0.030 0.315 0 0.056 0.620
Chlorophytes 0 0.220 0 0 0 0 0.062 0 0 0.031 0.180
Cryptophytes 0 0 0 0 0 0 0 0 0.220 0 0
Diatoms 0 0 0 0 0.520 0 0 0 0 0 0
Pseudonitzschia 0.033 0 0 0 0.610 0 0 0 0 0 0
P. antarctica H 0.130 0 0 0.010 0.080 0.400 0 0 0 0 0
P. antarctica L 0.270 0 0 0.001 0.010 1.100 0 0 0 0 0
Dinoflagellates 0 0 1.060 0 0 0 0 0 0 0 0
10.12952/journal.elementa.000043.t001b

Table 1B.

Optimized pigment: Chl a ratios after CHEMTAX analysisa

Phytoplankton class Pigment used in ratio with Chl a
Chl c3 Lut Per 19′-But Fuc 19′-Hex Neo Pras Allo Viol Chl b
Prasinophytes 0 0.006 0 0 0 0 0.030 0.094 0 0.056 0.620
Chlorophytes 0 0.220 0 0 0 0 0.062 0 0 0.031 0.180
Cryptophytes 0 0 0 0 0 0 0 0 0.220 0 0
Diatoms 0 0 0 0 0.520 0 0 0 0 0 0
Pseudonitzschia 0.033 0 0 0 0.624 0 0 0 0 0 0
P. antarctica H 0.201 0 0 0.010 0.080 0.284 0 0 0 0 0
P. antarctica L 0.098 0 0 0.001 0.010 1.493 0 0 0 0 0
Dinoflagellates 0 0 1.060 0 0 0 0 0 0 0 0

Phytoplankton photosynthesis rates

Photosynthesis vs irradiance (P-E) relationships were determined in surface water samples (2–10 m depth) at 12 stations and in one sample from pooled replicates for each incubation treatment. P-E relationships were determined using the 14C-bicarbonate incorporation technique by incubating 2 mL aliquots of seawater in a photosynthetron for two hours over a range of 20 different light intensities ranging from 3 to 542 µmol photons m−2 s−1 at 0°C (Lewis and Smith, 1983; the full method is outlined in Arrigo et al., 2010). CO2 incorporation normalized by Chl a concentration was calculated from radioisotope incorporation and the data were fit by least squares nonlinear regression to the equation of Webb et al. (1974):

P=Pmax(1exp(αEPmax))
(2)

where Pmax is the maximum rate of photosynthesis (CO2 incorporation in g C g−1 Chl a hr−1) and α is the initial slope of the P-E curve (g C g−1 Chl a hr−1 [µmol photons m−2 s−1]−1) where photosynthesis rates are light-limited. The photoacclimation parameter, Ek, was calculated as Pmax/. P-E data were also fitted to the model of Platt et al. (1980), which contains the photoinhibition parameter ß (g C g−1 Chl a hr−1 [µmol photons m−2 s−1]−1). However, ß was not significantly different from zero in any of the P-E curves and, therefore, this model was disregarded.

Phytoplankton optical absorption ()

The spectrally averaged optical absorption cross section (, m2 mg−1 Chl a) was determined in surface water samples (2–10 m depth) at 12 stations and in one sample from pooled replicates for each bioassay incubation treatment. Aliquots of the seawater sample (100–1000 mL) were filtered onto 25 mm Whatman GF/F filters for measurement of particulate absorption spectra (ap, 300–800 nm) and detrital absorption (adet, 300–800 nm) on a Perkin-Elmer Lambda 35 spectrophotometer equipped with an integrating sphere (Labsphere) using the filter pad method and optical corrections in Mitchell and Kiefer (1988) and the coefficients of Bricaud and Stramski (1990). Detrital absorption (adet, 300–800 nm) was assayed after methanol extraction according to the method of Kishino et al. (1985). Chl a-specific optical absorption cross sections (aph) at each wavelength () were calculated as:

aph(λ)=ap(λ)adet(λ)[Chla]
(3)

where [Chla] is the Chl a concentration of the sample.

Spectrally averaged Chl a-specific optical absorption cross sections (, m2 mg−1 Chl a) were calculated using the equation:

a¯*=Σλ=400λ=700aph*E(λ)Σλ=400λ=700E(λ)
(4)

where E() (µmol photons m−2 s−1) is the spectral irradiance of the photosynthetron light source.

Quantum yield of photosynthesis

The quantum yield of photosynthesis (m in C mol−1 photons) was calculated as:

Φm=α43.2a¯
(5)

after first confirming that m was maximal at the lowest light level used in each of the assays (Johnson and Barber, 2003).

Variable fluorescence

A Satlantic Fluorescence Induction and Relaxation (FIRe) system was used to determine the maximum photochemical efficiency (Fv/Fm) and the functional absorption cross section (PSII) (Å2 photon−1) of photosystem II (Gorbunov et al., 1999) for water samples at 11 stations and bioassay experiment samples. Prior to analysis, the FIRe was blanked with GF/F-filtered seawater from the same station. After sampling from the sample bottles, samples were acclimated in the dark at 2°C for 30 min to fully oxidize the photosynthetic reaction centers.

Bacterial productivity

Bacterial productivity was estimated in the bioassay experiments from incorporation of 3H-leucine into protein (Williams et al., 2015). Each of the incubation bottles served as a replicate in the measurements.

Water column analysis

Diffuse attenuation coefficient

The diffuse attenuation coefficient of downwelling PAR (Kd) in the water column was determined by fitting the equation:

Ez=E0 e Kdz
(6)

to each PAR profile, where Ez is the irradiance at depth z and E0 is the irradiance just below the sea surface.

Mixed layer depth (MLD)

The MLD was determined from each CTD profile as the shallowest depth at which the density (T) was 0.02 kg m−3 greater than at the surface (Cisewski et al., 2008; Alderkamp et al., 2012a).

Mean light level in the mixed layer (EUML)

To calculate the mean daily PAR in the upper mixed layer (EUML, mol photons m−2 day−1), we used the equation of Riley (1957):

EUML=E¯surfT(1eKdzUML)KdzUML
(7)

where surf is the total daily surface PAR averaged over five days and T is the mean transmittance through the sea surface (0.85 for open water, 0.20 for grey ice and nilas, and 0.05 for snow covered and multiyear ice).

Water column productivity

Phytoplankton productivity throughout the water column was estimated at each station from the Chl a concentrations, light availability in the water column, and P-E parameters, as described in Alderkamp et al. (2012a). Briefly, at depth intervals of 1 m, Chl a concentrations were estimated from continuous vertical fluorescence profiles that were calibrated to the measured Chl a concentrations at similar depths (Chl a = 0.71 fluorescence; R2 = 0.69). The daily light cycle was binned in 10-min intervals and the mean over the previous five days was used to estimate the sinusoidal light cycle at 1-m depth intervals at each station based on the measured Kd of that station. These light levels were then used to calculate the phytoplankton productivity at each depth using P-E parameters of the phytoplankton collected at 10 m depth. These P-E parameters were assumed to be representative of phytoplankton in the upper mixed layer (UML) where > 99% of the phytoplankton productivity occurred (virtually no light penetrated below the MLD because the high phytoplankton biomass levels resulted in high Kd in all stations).

Statistical analysis

Effects of Fe addition were tested by comparing Fe addition treatments to unamended controls in each bioassay experiment using one-way ANOVA analysis. All parameters for phytoplankton photosynthesis and cellular composition, as well as bacterial productivity, were tested at day four to eliminate potential effects of NO3 limitation (see Phytoplankton biomass response to Fe addition). Interactions between Fe addition and original sample depth, as well as interactions between Fe addition and light availability, were tested using two-way ANOVA analysis. Differences were considered significant at p < 0.05. Simple linear regression was used to test relationships between phytoplankton parameters in seawater samples.

Results

Phytoplankton bloom characteristics

The physical properties of the upper water column, which affected the Fe and light availability to the phytoplankton, varied markedly within the ASP (Randall-Goodwin et al., 2015). In the southern ASP, near the ice sheets, AASW was relatively salty, due to little overall sea ice or near surface glacial melt, and/or enhanced mixing with WW, as observed in the southwest ASP close to the Getz Ice Shelf (GIS) (e.g., Station 5, Figure 2), or due to enhanced mixing with meltwater-laden mCDW waters as observed in front of the Dotson Ice Shelf (DIS) and at Station 57 (Figure 2). In the central ASP, AASW was moderately fresh, due to less recent sea ice melt and/or wind mixing (e.g., Stations 13 and 35, Figure 2). In the northern ASP, AASW was very fresh due to recent sea ice melt, as observed along the sea ice edge and in the MIZ (e.g., Station 66, Figure 2). These surface water characteristics affected the phytoplankton throughout our sampling during the buildup of a dense spring-summer phytoplankton bloom that typically peaks in mid-January (Arrigo et al., 2012). The phytoplankton biomass increased over the course of the cruise (Yager et al., 2012), conforming to this typical bloom development.

doi: 10.12952/journal.elementa.000043.f002.
Figure 2.  

Water properties of Fe addition bioassay experiment stations.

Depth profiles of salinity are shown in (A), and temperature and salinity properties in (B), where AASW is Antarctic Surface Water, WW is Winter Water, and mCDW is modified Circumpolar Deep Water. Open symbols indicate surface waters (10 m depth); solid symbols indicate deeper subsurface waters (D, 35–50 m depth).

Near the Getz Ice Shelf

The salty AASW with the largest contribution of WW was observed in the southwest ASP close to the GIS (Station 5; Table 2A) and had a MLD of ∼ 35 m (Figure 3C), resulting in moderate light levels in the UML (EUML 158 µmol photons m−2 s−2). DFe in these surface waters were > 0.18 nmol L−1 (Figure 3E), whereas PFe was two orders of magnitude higher than DFe at > 18 nmol L−1 (Figure 3F). Phytoplankton biomass was 2.3 mg Chl a m−3 in surface waters (Figure 3G) and 177 mg m−2 integrated over depth (Figure 3H). The phytoplankton community was dominated by Phaeocystis antarctica, but diatoms and prasinophytes were also present and constituted up to 12% of the phytoplankton community at different depths (Table 2B, Figures 3I, 3J). Of the five stations sampled for the Fe addition bioassay experiments (see Phytoplankton biomass response to Fe addition), Station 5 is the most representative of salty AASW with WW influence (Figure 2).

10.12952/journal.elementa.000043.t002a

Table 2A.

Oceanographic data on water samples used to initiate bioassay experiments

Type of data Station-specific data
Station 5 13 35 57 66
Latitude (°S) 73°57.72 73°34.22 73°16.78 73°43.69 72°44.45
Longitude (°W) 118°01.18 112°40.03 112°06.28 113°15.20 116°01.21
Date 14 Dec 2010 19 Dec 2010 26 Dec 2010 1 Jan 2011 5 Jan 2011
Mixed layer depth (m) 28 42 22 80 25
Mean irradiance during experiment (µmol photons m−2 s−1) 1243 932 1098 1082 776
Sample type (surface, S; deeper subsurface, D) S D S D S D S S
Sample depth (m) 8.8 50.3 9.9 40.0 12.0 35.0 9.9 10.2
Salinity 34.01 34.05 33.94 33.94 33.83 33.93 33.93 33.69
Temperature (°C) −1.46 −1.71 −1.25 −1.25 −0.26 −1.23 −0.55 −1.33
NO3 (µmol L−1) 26.3 26.7 21.1 24.8 12.5 22.0 24.9 15.8
PO4 (µmol L−1) 1.76 1.78 1.33 1.61 0.88 1.49 1.67 1.18
Si(OH)4 (µmol L−1) 88.0 93.0 77.9 83.8 70.8 72.5 96.5 74.1
DFe (nmol L−1) 0.18 0.30 0.09 0.22 0.32 0.20 0.10 0.116
PFe (nmol L−1) 18.88 12.54 ND ND 7.28 6.78 21.07 15.43
10.12952/journal.elementa.000043.t002b

Table 2B.

Biological data on water samples used to initiate bioassay experimentsa

Type of data Station 5 Station 13 Station 35 Station 57 Station 66
S D S D S D S S
Chl a 3.8 3.7 8.4 6.5 14.0 7.4 5.6 10.8
POC 244 179 386 287 903 471 237 544
POC/Chl a 65.4 48.5 46.4 44.4 66.8 63.7 42.7 51.1
POC/PON 7.05 4.50 5.66 6.20 7.79 6.02 5.32 5.66
Fv/Fm 0.49 0.46 0.35 0.39 0.27 0.39 0.40 0.36
PSII 525 561 737 677 608 648 544 638
Pmax 3.52 3.00 2.97 2.61 2.27 2.92 3.18 2.48
0.080 0.065 0.051 0.058 0.034 0.075 0.032 0.067
Ek 44 46 58 45 67 39 99 37
0.015 0.008 0.013 0.011 0.014 0.014 0.013 0.010
Fm 0.127 0.191 0.091 0.123 0.057 0.128 0.060 0.119
Dominant phytoplankton (fraction of total) Phaeocystis antarctica (0.80) P. antarctica (0.85) P. antarctica (0.87) P. antarctica (0.88) P. antarctica (0.84) P. antarctica (0.84) P. antarctica (0.87) P. antarctica (0.84)
2nd most abundant phytoplankton Prasinophytes (0.11) Diatoms (0.11) Prasinophytes (0.09) Prasinophytes (0.08) Prasinophytes (0.14) Prasinophytes (0.15) Diatoms (0.05) Diatoms (0.11)
BP 6.4 5.6 63.6 55.6 74.1 61.7 49.5 89.6
doi: 10.12952/journal.elementa.000043.f003.
Figure 3.  

Surface water properties in the Amundsen Sea Polynya.

Properties shown for surface waters (2–10 m depth) are: salinity (A); temperature (B); mixed layer depth (MLD), the depth where density t increased 0.02 from surface waters (C); EUML, mean light in the upper mixed layer (D); dissolved Fe (DFe) concentrations in surface water (data from Sherrell et al., 2015) (E); particulate Fe (PFe) concentrations in surface waters (data from Harazin et al., 2014) (F); chlorophyll (Chl) a concentrations (G); depth-integrated Chl a (H); fraction of phytoplankton community as Phaeocystis antarctica (I); and fraction of phytoplankton community as diatoms (J). In each panel, the dashed line shows the sea ice edge on 1 January 2011 and the Getz Ice Shelf (GIS) and Dotson Ice Shelf (DIS) are shown in white.

Near the Dotson Ice Shelf

Salty AASW with the largest contribution of mCDW was observed in the southern ASP close to the DIS (Stations 8, 9, 10, 11, and 57). The buoyancy-driven upwelling of mCDW destabilized the upper water column resulting in deep MLDs (40 to 70 m, Figure 3C). The deep MLD, in combination with variable light attenuation due to variable phytoplankton biomass, resulted in variable EUML (50 to 220 µmol photons m−2 s−2, Figure 3D). The meltwater-laden mCDW flowing from under the DIS appears to be the main source of DFe and PFe to the ASP (Sherrell et al., 2015) and the salty AASW contained 0.11–1.31 nmol L−1 DFe (Figure 3E) and 15–50 nmol L−1 PFe (Figure 3F). Phytoplankton biomass was < 2.0 mg Chl a m−3 in surface waters (Figure 3G) and < 200 mg Chl a m−2 integrated over depth near the face of the DIS where the mCDW signature was strongest. At Station 57, approximately 50 km away from the DIS, surface Chl a increased to 8.6 mg m−3 (Figure 3G) and depth-integrated Chl a to 618 mg m−2 (Figure 3H).

Central Amundsen Sea Polynya

In the center of the polynya, at increasing distance from the Dotson and Getz Ice Shelves (Stations 6, 18–32, 35–50), solar warming increased the temperature of AASW (−0.7 °C, Figure 3B) and recent sea ice melt lowered the salinity (< 33.9, Figure 3A). Both processes increase water column stratification, which decreased the MLD to 10–30 m (Figure 3C). The highest phytoplankton biomass was observed in the central ASP with mean surface Chl a of 12.6 ± 5.7 mg m−3 (Figure 3G) and mean depth-integrated Chl a of 604 ± 150 mg m−2 (Figure 3H). The highest Chl a concentrations were found in the upper 20 m of the water column, dropping to 5–10 mg m−3 at a depth of 20–50 m (Figure 5A). The shallow MLD, in combination with rapid light attenuation with depth due to the high phytoplankton biomass, resulted in a relatively low EUML (< 50 µmol photons m−2 s−1, Figure 3D). Moreover, the rapid light attenuation with depth resulted in a shallow euphotic zone with a 1% light depth shallower than the MLD. DFe and PFe in surface waters of the central polynya were on average 0.18 ± 0.07 nmol L−1 (Figure 3E) and 8.7 ± 3.6 nmol L−1 (Figure 3F), respectively. The phytoplankton community was dominated by P. antarctica at almost all stations throughout the central polynya. Diatoms contributed up to 60% to the phytoplankton biomass at two stations (Figures 3I, 3J), but contributed less or were almost absent at most stations. Prasinophytes were present in low numbers at most stations (< 10%) and contributed up to 14% to the phytoplankton community at some. The bioassay experiments at Stations 13 and 35 were conducted in the central ASP, where sea ice melt influence was stronger at Station 35 than at Station 13 (Figure 2).

Northern Amundsen Sea Polynya and Marginal Ice Zone (MIZ)

Very fresh AASW that was strongly affected by sea ice melt was found in the MIZ (Station 66) and in the northern waters of the ASP near the sea ice (Station 34). Sea ice melt waters were relatively fresh (< 33.85) and cold (< −1.5°C) (Figures 3A, 3B), but warmed over time. Freshening of surface waters by sea ice melt water increased stratification, which resulted in a shallow MLD (< 20 m, Figure 3C) and a relatively high EUML > 150 µmol photons m−2 s−2 (Figure 3D). DFe in very fresh AASW was relatively low (0.11 ± 0.03 nmol L−1; Figure 3E), as was PFe (6.9 ± 1.6 nmol L−1; Figure 3F) when compared to other regions in the ASP. Phytoplankton biomass in very fresh AASW ranged from 6 to 10 mg Chl a m−3 (Figure 3G) and depth-integrated Chl a exceeded 400 mg m−2 (Figure 3H). All very fresh AASW stations were dominated by P. antarctica, although diatoms and prasinophytes were also present and constituted up to 4% and 13% of the phytoplankton assemblages, respectively. The bioassay experiment at Station 66 was situated in the MIZ where phytoplankton biomass was high, even though waters were still partially ice covered (Table 2B).

Phytoplankton photosynthesis in the ASP

Variable fluorescence parameters

In general, the Fv/Fm of phytoplankton was low in surface waters of the central ASP (< 0.35), and higher (> 0.4) in the salty AASW of the southern ASP near the DIS and GIS, at stations with a relatively strong WW and mCDW influence, respectively (Figure 4A). Phytoplankton in the very fresh AASW of the northern exhibited Fv/Fm values of 0.35. In general, Fv/Fm increased with depth in the upper 80 m of the water column (Figure 5B). The lowest Fv/Fm was observed in surface waters in the central ASP with the highest phytoplankton biomass. The relationship between Fv/Fm and Chl a in surface waters was negative (Table 3).

doi: 10.12952/journal.elementa.000043.f004.
Figure 4.  

Photosynthetic parameters of surface phytoplankton.

Photosynthetic parameters shown for phytoplankton in surface waters (2–10 m depth) are: maximum photochemical efficiency of Photosystem (PS) II (Fv/Fm) (A), functional cross section of PS II (PSII) (B), maximum photosynthesis rates normalized to chlorophyll a (Chl a) (Pmax) (C), initial light-limited slope of the photosynthesis–irradiance normalized to Chl a () (D), and photoacclimation parameter (Ek) (E). Water column productivity is shown in (F). In each panel, the sea ice edge on 1 January 2011, is shown by dashed line and the ice shelves are white: Getz Ice Shelf (GIS) and Dotson Ice Shelf (DIS).

doi: 10.12952/journal.elementa.000043.f005.
Figure 5.  

Depth profiles of phytoplankton variable fluorescence.

Mean and standard deviation of chlorophyll a (Chl a, n = 18) (A), maximum photochemical efficiency of Photosystem (PS) II (Fv/Fm, n = 14) (B), and functional cross section of PS II (PSII, n = 14) (C) are shown with depth in the water column.

The PSII of phytoplankton was variable (500–900 Å photon−1) throughout surface waters in the ASP (Figure 4B). The highest PSII (800–900 Å photon−1) was observed in salty AASW near the DIS, whereas the lowest (500 Å photons−1) was observed close to the GIS (Figure 4B). The PSII in the central polynya ranged from 473 to 938 Å photon−1 and was similar to that in waters affected by sea ice melt. There was no trend in PSII with depth (Figure 5C); moreover, there was no significant relationship between PSII and Chl a (Table 3).

P-E parameters

The spatial distribution of Pmax resembled that of Fv/Fm. Pmax was highest in the salty AASW in the southern ASP near the DIS and GIS (Figure 4C), exceeding 3.0 µg C µg−1 Chl a h−1. The Pmax was lowest (< 2.0 µg C µg−1 Chl a h−1) in the central ASP in the areas with the highest phytoplankton biomass. In fresh AASW waters affected by sea ice melt water, Pmax was intermediate. There was no trend in Pmax with depth (Table 1B). The relationship between Pmax and Fv/Fm was positive (Table 3). There was no relationship between Pmax and DFe, but a positive relationship between Pmax and PFe was detected (Table 3).

The spatial distribution of resembled that of Pmax, with high (> 0.06 µg C µg−1 Chl a h−1 [umol photons m−2 s−1]) in the southern ASP near the DIS and GIS (Figure 4D) and low (< 0.03 µg C µg−1 Chl a h−1 [umol photons m−2 s−1]) in the central ASP. The parameter was intermediate in fresh AASW affected by sea ice melt water in the northern ASP (Figure 4D). In general, α increased with depth (Table 2B). There were strong positive relationships between and Fv/Fm, and between and Pmax (Table 3). There was no relationship between and DFe, but the relationship between and PFe was positive (Table 3).

10.12952/journal.elementa.000043.t003

Table 3.

Simple linear regressions for phytoplankton photosynthesis parameters in surface waters of the Amundsen Sea Polynyaa

Variable 1 Variable 2 Equation n R2 p
Fv/Fm Chl a Fv/Fm=−0.008 Chl a + 0.433 14 0.594 < 0.01
PSII Chl a 14 0.005 0.820
Pmax Fv/Fm Pmax = 8.84 Fv/Fm − 0.13 13 0.411 < 0.05
Pmax DFe 14 0.154 0.208
Pmax PFe Pmax = 0.09 PFe + 1.67 8 0.745 < 0.01
Fv/Fm = 0.26 Fv/Fm − 0.035 13 0.678 < 0.001
Pmax = 0.02 Pmax − 0.009 14 0.820 < 0.001
DFe 14 0.081 0.370
PFe = 0.002 PFe + 0.024 8 0.506 < 0.05
WCP Chl a 11 0.001 0.928
WCP Depth-integr Chl a 11 0.041 0.552
WCP MLD 11 0.002 0.906
WCP EUML 11 0.059 0.473
WCP Pmax 11 0.043 0.538
WCP 11 0.037 0.571
WCP DFe 11 0.005 0.829
WCP PFe 8 0.063 0.548

The Ek ranged from 40 to 100 µmol photons m−2 s−1 (Figure 4E) and was generally lower than EUML (Figure 4E). The strong positive relationship between Pmax and resulted in no spatial pattern in Ek, although Ek did decrease with depth (Table 2B).

Water column productivity in the ASP ranged from 2.0 to 6.5 g C m−2 d−1 (Figure 4F). In general, productivity was highest (> 5.0 g C m−2 d−1) in the central polynya in waters with relatively shallow MLD (< 30 m) that resulted in high values for EUML (> 130 µmol photons m−2 s−2) and Chl a concentrations (> 7 mg m−3 and > 400 mg m−2). Water column productivity depends on phytoplankton biomass, light availability, and P-E parameters of phytoplankton. However, productivity showed no relationship with any of these parameters, such as surface Chl a, depth-integrated Chl a, MLD, EUML, Pmax, or (Table 3). Similarly, there was no relationship between productivity and DFe in surface waters or between productivity and PFe (Table 3).

Phytoplankton biomass response to Fe addition

In general, phytoplankton biomass increased significantly over the course of the incubation in all experiments at 10% and 50% irradiance (Figure 6, left panels). The increase in biomass was accompanied by a rapid drawdown of NO3 (Figure 6, right panels), which was depleted by the end (7 or 8 days) in all incubations at 10% and 50% irradiance, both in the control and the Fe treatments (Figure 6, right panels). Moreover, all NO3 was depleted by day 4 in both control and Fe treatments of surface water from Station 35, such that these incubations were likely NO3-limited (Figure 6F).

Fe addition resulted in higher biomass compared to controls in four of the bioassay incubations in the ASP, whereas there was no significant effect of Fe in eight incubations (Figure 6). In the incubations of surface water from Station 5, in the salty AASW near the face of the GIS, Fe addition enhanced phytoplankton biomass 1.37-fold by day 4, which increased growth rate to 0.29 d−1 compared to 0.26 d−1 in the control (one-way ANOVA, p < 0.05, Table 4). By day 7, when NO3 was depleted, biomass in the Fe treatment was still 1.27-fold higher than in the control treatment (one-way ANOVA, p < 0.05, Figure 6A). In the incubations of subsurface water from Station 35, in the moderately fresh AASW in the central ASP, Fe addition enhanced phytoplankton biomass 1.32-fold by day 4 (one-way ANOVA p < 0.05, Figure 6E), which increased the phytoplankton growth rates to 0.26 d−1 compared to 0.12 d−1 in the control (one-way ANOVA, p < 0.05, Table 4). By day 8, when NO3 was depleted, biomass in the Fe treatment was still 1.13-fold higher than in the control treatment (one-way ANOVA, p < 0.05, Figure 6E). In the incubations of subsurface water from Station 57, in the salty AASW affected by mCDW, Fe addition enhanced biomass 1.14-fold in the 10% irradiance incubation by day 4 (one-way ANOVA, p < 0.05), which increased phytoplankton growth rates to 0.31 d−1 compared to 0.29 d−1 in the control (one-way ANOVA, p < 0.05, Table 4). By day 7, when NO3 was depleted, Fe addition enhanced the biomass 1.42-fold in the 50% irradiance treatment (one-way ANOVA, p < 0.05, Figure 6G). Fe addition also enhanced NO3 drawdown when compared to controls in two subsurface water incubations (Stations 13 and 35), whereas there was no significant Fe effect in the ten other incubations.

doi: 10.12952/journal.elementa.000043.f006.
Figure 6.  

Phytoplankton responses to iron addition.

Mean and standard deviation of the measured parameters are shown for triplicate incubations of Fe additions (solid symbols) and unamended controls (open symbols). Left panels indicate change in biomass (particulate organic carbon: POC) over time of incubation; right panels indicate change in nitrate (NO3). The upper three rows of panels show responses during incubation of surface (black) and deeper subsurface (blue) waters from Station 5 (A, B) near the Dotson glacier, Station 13 (C, D) in the central Amundsen Sea Polynya (ASP), and Station 35 (E, F) in the central ASP. The lower two rows of panels show responses at different levels of incident irradiance (1%, green; 10%, black; 50%, red) for surface waters from Station 57 (G, H) close to the Dotson Ice Shelf and Station 66 (I, J) in the marginal ice zone (MIZ). An asterisk indicates significant difference between Fe addition and control of the same color at the time of sampling (one-way ANOVA, p < 0.05). In (F), data for SC are identical to data for SFe.

10.12952/journal.elementa.000043.t004

Table 4.

Physiological characteristicsa of phytoplankton in the bioassay experimentsb

Characteristic Bioassay experiment Station-specific data by sample type Station-specific data by light level
Station 5 Station 13 Station 35 Station 57 Station 66
S D S D S D 1% 10% 50% 1% 10% 50%
Phytoplankton growth rate (d−1) C 0.26 (0.01) 0.31 (0.01) 0.25 (0.01) 0.24 (0.02) 0.10 (0.01) 0.12 (0.01) 0.06 (0.01) 0.29 (0.00) 0.28 (0.01) 0.07 (0.01) 0.18 (0.01) 0.19 (0.01)
Fe 0.29 (0.00) 0.33 (0.01) 0.24 (0.01) 0.30 (0.00) 0.19 (0.01) 0.26 (0.01) 0.08 (0.01) 0.31 (0.00) 0.33 (0.00) 0.06 (0.02) 0.23 (0.01) 0.19 (0.01)
POC/Chl a (wt/wt) C 58 (5) 102 (17) 73 (20) 89 (17) 75 (2) 56 (12) 36 (3) 60 (10) 89 (9) 65 (7) 79 (5) 108 (7)
Fe 72 (14) 84 (15) 229 (212) 67 (19) 75 (19) 79 (15) 44 (3) 64 (7) 101 (26) 73 (10) 116 (15) 131 (29)
POC/PON (mol/mol) C 5.1 (1.5) 5.4 (0.5) 7.3 (0.6) 6.8 (0.2) 6.3 (0.2) 6.2 (0.3) 6.3 (3.5) 6.6 (0.2) 4.9 (1.3) 5.6 (0.4) 5.8 (0.6) 4.6 (0.3)
Fe 4.0 (0.7) 5.1 (1.0) 6.8 (1.1) 6.5 (0.2) 8.5 (1.9) 6.5 (0.5) 4.6 (0.3) 5.4 (0.8) 5.1 (1.1) 6.2 (1.1) 6.6 (0.6) 5.4 (1.1)
- (photons m−2) C 0.014 0.016 0.014 0.017 0.015 0.013 0.010 0.012 0.014 0.011 0.012 0.012
Fe 0.015 0.015 0.016 0.013 0.011 0.009 0.014 0.012 0.011 ND 0.013 0.013
m (mol C mol−1 photons) C 0.076 0.102 0.051 0.051 0.066 0.061 0.134 0.100 0.056 0.120 0.080 0.044
Fe 0.139 0.119 0.073 0.081 0.150 0.173 0.144 0.099 0.137 ND 0.093 0.096

Interactions between effects of Fe and sample depth on phytoplankton growth rates and NO3 drawdown were studied by analyzing data from day 4 of Stations 5, 13 and 35 bioassay experiments together (Table 5). Both growth rates and NO3 drawdown were higher in the subsurface than surface water incubations over these three experiments (one-way ANOVA, p < 0.05). These results are likely due to NO3-limitation in the incubations of surface water from Station 35, resulting in lower growth rates and NO3 drawdown compared to subsurface water incubations that were not NO3-depleted. There was no interaction between Fe and sample depth on phytoplankton growth rates, nor on NO3 drawdown (Table 5).

10.12952/journal.elementa.000043.t005

Table 5.

Significance of effectsa of Fe addition, sample depth, and light levels on phytoplankton variablesb and bacterial productivity

Variable p valuesa for bioassay test (number of experiments)
Fe (5) Depth (3) Fe x Depth (3) Light (2) Fe x Light (2)
Phytoplankton growth rate 0.018 0.009 0.668 0.000 0.723
NO3 drawdown 0.037 0.034 0.202 0.000 0.280
Fv/Fm 0.008 0.164 0.866 0.012 0.093
PSII 0.471 0.087 0.665 0.617 0.017
Pmax 0.039 0.500 0.527 0.217 0.868
0.006 0.851 0.473 0.024 0.322
Ek 0.419 0.359 0.701 0.000 0.002
0.234 0.379 0.219 0.381 0.822
Fm 0.005 0.796 0.935 0.005 0.051
POC/PON 0.977 0.665 0.731 0.119 0.586
POC/Chl a 0.124 0.440 0.170 0.000 0.785
Bacterial productivity 0.311 0.928 0.533 0.618 0.951

Data from day 4 of the Stations 57 and Station 66 incubations were analyzed together to study interactions between Fe and light on phytoplankton growth rates and NO3 drawdown (Table 5). Whereas both higher Fe and light availability individually increased the phytoplankton growth rates in the incubations (one-way ANOVA p < 0.05), there was no interaction between these factors. Similarly, there were no interactions between Fe and light effects on NO3 drawdown (Table 5).

Phytoplankton photosynthesis response to Fe additions

Variable fluorescence responses

All photosynthesis responses to Fe additions were studied on day 4 of the incubations, when NO3 was not depleted, except for the surface water incubation from Station 35. The Fv/Fm ranged from 0.25 to 0.48 (Figures 7A, 7B) and was in the same range as the initial Fv/Fm of surface phytoplankton (Figure 4A). Fe addition increased the Fv/Fm in almost all of the 10% and 50% irradiance incubations when compared to control treatments (one-way ANOVA, p < 0.05, Figures 7A, 7B) by an average of 1.17-fold, whereas there was no effect in the 1% irradiance incubations. Despite this trend, the interaction between Fe and light was not significant (Table 5) when experiments at Stations 57 and 66 were analyzed together, likely due to the small but significant decrease in Fv/Fm after Fe addition in the 10% irradiance incubation at Station 57 (Figure 7B). In addition, the original sampling depth did not affect the Fv/Fm in the incubations, nor was there an interaction between Fe and sample depth (Table 5).

doi: 10.12952/journal.elementa.000043.f007.
Figure 7.  

Variable fluorescence in incubation experiments.

Mean and standard deviation of variable fluorescence by day 4 of the incubation experiments are shown for triplicate incubations of Fe additions (dark bars) and unamended controls (white bars). An asterisk indicates significant difference between Fe addition and control (one-way ANOVA, p < 0.05). Maximum photochemical efficiency of Photosystem (PS) II (Fv/Fm) is shown for surface (S) and deeper subsurface (D) incubations from Stations 5, 13 and 35 (A) and for different levels of incident irradiance (1%, black; 10%, grey; 50%, light grey) for Stations 57 and 66 (B). Functional cross section of PS II (PSII) is shown for the surface and subsurface depths (C) and levels of incident irradiance (D).

The PSII by day 4 of the incubations ranged from 442 to 765 Å photon−1, which was within the range of the initial PSII of the incubations (Table 2B) and that of surface phytoplankton in the ASP (Figure 4B). Fe effects on PSII were inconsistent (Figures 7C, 7D), reducing PSII in two incubations (Stations 5, S incubation; Station 57, 10% irradiance incubation), increasing PSII in another incubation (Station 13, D incubation), and not affecting PSII in the remaining incubations (Figures 7C, 7D). Analyzing all incubations together revealed no overall Fe effect on PSII (Table 5). The PSII of the surface water incubations was slightly higher than the subsurface water incubations, but this effect was not significant (Table 5). Moreover, there was no interaction between effects of Fe and sample depth (Table 5). Light did not affect PSII, but there was a significant interaction between Fe and light (two-way ANOVA, p < 0.05), where Fe addition slightly decreased PSII at low light, but did not affect PSII at high light (Figure 7D).

Photosynthesis vs irradiance (P-E) parameters

High Pmax exceeding 2.2 µg C µg−1 Chl a h−1 were observed in all incubations by day 4 (Figures 8A, 8B). Fe addition increased Pmax compared to controls in almost all incubations, on average 1.31-fold. This increase was greatest in the experiments where Fe also affected phytoplankton growth (S for Station 5, D for Station 35), with Pmax increasing 1.75-fold and 1.56-fold, respectively. Pmax was lowest for surface water incubations from Station 35, where NO3 was depleted by day 4. When experiments were analyzed together, the Fe effect on Pmax was significant (Table 5, one-way ANOVA, p < 0.05). The original sample depth did not affect Pmax in the incubations for Stations 5, 13, and 35, and there was no interaction between Fe and depth (Table 5). Pmax increased slightly with higher light, although this effect was not significant; there was no interaction between Fe and light effects (Table 5).

doi: 10.12952/journal.elementa.000043.f008.
Figure 8.  

Photosynthesis versus irradiance parameters in incubation experiments.

Photosynthesis versus irradiance (P-E) parameters at day 4 of the incubation experiments are shown for incubations of Fe additions (dark bars) and unamended controls (white bars). Maximum photosynthesis rates (Pmax) are shown for surface (S) and deeper subsurface (D) incubations from Stations 5, 13 and 35 (A) and for different levels of incident irradiance (1%, black; 10%, grey; 50%, light grey) for Stations 57 and 66 (B). Initial slope of the P-E curve () is shown for the surface and subsurface depths (C) and levels of incident irradiance (D), as is photoacclimation parameter (Ek) (E & F).

The was high in all incubations (> 0.024 µg C µg−1 Chl a h−1 [µmol photons m−2 s−1]−1, Figures 8C, 8D) and generally followed the trends in Pmax. Fe addition increased in almost all experiments (Figures 8C, 8D), on average by 1.42-fold. This increase was greatest in the incubations where Fe additions also affected phytoplankton growth (S for Station 5, D for Station 35), where increased 1.89-fold and 1.95-fold, respectively. Thus, in Fe-limited phytoplankton, the relative effect of Fe on was greater than that on Pmax. When all experiments were analyzed together, the Fe effect on was significant (Table 5). The original sample depth did not affect in the incubations at Stations 5, 13, and 35, and there was no interaction between Fe and sample depth (Table 5). The decreased at higher light, but there was no interaction between Fe and light effects (Table 5).

The photoacclimation parameter (Ek) was high (53–104 µmol photons m−2 s−1) in all incubations at 10% and 50% irradiance and somewhat lower at 1% irradiance (42–46 µmol photons m−2 s−1) (Figures 8E, 8F). Fe addition did not affect Ek because both α and Pmax increased relative to the controls. When all experiments were analyzed together, Ek was not significantly affected by either Fe, depth, or the interaction between Fe and depth (Table 5). On the other hand, Ek increased at higher light and there was an interaction between Fe and light effects (Table 5), where Fe addition at high light (50% irradiance) decreased Ek, but there was no effect at lower light.

The in the incubations varied between 0.009 and 0.017 m−2 mg−1 Chl a (Table 4) and was similar to initial of phytoplankton in the ASP (Table 2B). Fe addition did not affect in any of the incubations (Table 4). Moreover, when experiments were analyzed together, there was no effect of either Fe, sample depth, or light on , and there were no interactive effects (Table 5).

The m in the incubations varied between 0.051 and 0.173 mol C mol−1 photons (Table 4) and was similar to initial m of phytoplankton in the ASP (Table 2B). Fe addition increased m in all incubations, on average 1.51-fold (Table 4, Table 5). There was no effect of original sample depth on the m in the incubations and there were no interactions between Fe and depth effects (Table 5). The m was higher at low light (Table 4, Table 5) and there was an interaction between Fe and light effects (Table 5), where the Fe effect was stronger at high light.

Cellular composition of phytoplankton

The POC/PON ratios in the incubations showed considerable variation by day 4, ranging from 3.5 to 8.5 (Table 4). The highest POC/PON ratio was found in the Fe incubations of surface waters from Station 35 that were NO3-limited. Fe addition did not affect the POC/PON ratio in any of the experiments (Table 4), and the lack of Fe effect was confirmed when experiments were analyzed together (Table 5).

The POC/Chl a ratios in the incubations ranged from 36 to 229 g g−1 (Table 4). Fe addition increased the POC/Chl a ratio in the 10% irradiance incubation at Station 66, but did not affect the POC/Chl a ratio in any other incubations (Table 4). When all experiments were analyzed together, Fe additions enhanced POC/Chl a ratios slightly (1.14-fold) compared to the controls, but this difference was not significant (Table 5). Initial sampling depth did not affect POC/Chl a ratios in the incubations, and there was no interaction between depth and Fe effects (Table 5). On the other hand, the POC/Chl a decreased at low light incubations, but there was no interaction between Fe and light effects on POC/Chl a (Table 5).

Bacterial productivity

Bacterial productivity in the incubations by day 4 ranged from 36 to 289 pmol Leu uptake L−1 h−1 (Figure 9). It increased 2 to 17-fold over the course of 4 days in all incubations, except for Station 57 at 1% irradiance, where bacterial productivity dropped slightly from the initial value (compare Figure 9 to Table 2B). Fe addition did not affect bacterial productivity in any of the incubations (Figure 9), including the incubations where Fe addition enhanced phytoplankton growth (Station 5, S; Station 35, D; Station 57, 10%), suggesting that there were no secondary effects of bacterial productivity on Fe-enhanced phytoplankton growth and photosynthesis. Moreover, original sample depth did not significantly affect bacterial productivity in the incubations, and there was no interaction between Fe and depth effects (Table 5). Light levels in the incubations did not affect bacterial productivity either, despite the big difference in phytoplankton biomass between the 1% irradiance incubations and those at higher light (10% and 50% irradiance) (Table 5).

doi: 10.12952/journal.elementa.000043.f009.
Figure 9.  

Bacterial productivity in incubation experiments.

Mean and standard deviation of bacterial productivity by day 4 of the incubation experiments are shown for triplicate incubations of Fe additions (dark bars) and unamended controls (white bars). Rates are shown for surface (S) and deeper subsurface (D) incubations from Stations 5, 13, and 35 (A) and for different levels of incident irradiance (1%, black; 10%, grey; 50%, light grey) for Stations 57 and 66 (B).

Discussion

Phytoplankton response to Fe additions: carrying capacity versus rate limitation

Phytoplankton productivity in Antarctic polynyas is often assumed to be seasonally Fe limited, with a “winter reserve” of DFe that is gradually depleted over the growing season. The amount of phytoplankton biomass that is supported by this DFe has been referred to as the carrying capacity (Hopkinson et al., 2013). Since all available NO3 was drawn down in both the controls and Fe addition incubation experiments with sufficient light (the 10% and 50% light incubations), the results from the incubation experiments suggest that Fe availability is not limiting the carrying capacity of waters in the ASP. Moreover, previous bioassay experiments conducted later in the growing season showed no response of the phytoplankton biomass to Fe additions in the ASP in February 2009 (Mills et al., 2012). These experiments were performed after the peak in phytoplankton bloom for that season, suggesting that Fe limitation was not the cause of the bloom demise.

In contrast, Fe addition increased phytoplankton growth rates at several stations where we performed bioassay experiments (Station 5, surface water; Station 35, subsurface water; Station 57, 10% light; Figure 6; Table 4) during the build-up of the phytoplankton spring bloom, and enhanced photosynthesis rates in almost all experiments (Figure 8). Thus, even though Fe availability is not limiting the carrying capacity of the ASP (Mills et al., 2012), the Fe effects in the bioassay experiments suggest that Fe availability may limit phytoplankton growth rates in several regions of the ASP. We could not discern physical mechanisms (e.g., MLD, water mass properties) that could explain the spatial variability of Fe effects on phytoplankton growth. The timing of the Fe effects in mid-December to early January was early in the growing season, well before the peak of the phytoplankton bloom in the ASP, which is generally in the middle of January (Arrigo et al., 2012). These findings suggest that the notion of a “winter reserve” of DFe that is gradually depleted over the growing season is an oversimplification, concurring with findings by Sedwick et al. (2011) who showed early season depletion of DFe in the Ross Sea Polynya. Instead, bioavailable Fe must be supplied throughout the growing season to support productive phytoplankton blooms such as those observed in the Amundsen and Ross Sea Polynyas.

Bacterial productivity in the incubation experiments was similar to that reported in the high productivity stations in the central ASP, which was high compared to other Antarctic polynyas (see Williams et al., 2015, for a full description). In contrast to phytoplankton growth and photosynthesis, bacterial productivity did not respond to Fe addition, suggesting that bacteria are not Fe-limited during the early season in the ASP. The Fe demand of bacteria is generally lower than that of phytoplankton, as heterotrophic bacteria lack the photosynthetic apparatus that is rich in Fe (Raven et al., 1990). Moreover, bacteria are smaller, with a larger surface to volume ratio that optimizes nutrient uptake. In addition, despite the increase in bacterial productivity over the four days of incubation, there was no secondary response of bacterial productivity to either the Fe-enhanced phytoplankton growth or photosynthesis rates. Bacterial productivity in surface waters of the ASP showed a positive relationship with phytoplankton biomass (Williams et al., 2015), suggesting bacterial productivity was coupled to phytoplankton biomass and productivity. Likely, the Fe effects on phytoplankton biomass, cellular composition, and photosynthesis rates in our experiments were not sufficient to yield a detectable response in bacterial productivity during the time scale of the incubations.

Fe availability as the main driver for phytoplankton photosynthesis rates in the ASP

The Fe addition bioassay experiments showed that greater Fe availability increased all photosynthesis parameters, including Fv/Fm, Pmax, and , similar to results from culture experiments on Phaeocystis antarctica (Strzepek et al., 2012; Alderkamp et al., 2012b; Van Leeuwe and Stefels, 2007). On the other hand, increased light availability in the bioassay experiments increased Pmax, but decreased and Fv/Fm. Typically, photoacclimation to high light increases carbon-fixing activity, such as electron transport and the Calvin cycle, resulting in a higher cellular C/Chl a ratio, whereas it decreases the photon capture and photosynthetic efficiency at low light (Falkowski and LaRoche, 1991; MacIntyre et al., 2002). The spatial distribution of P-E parameters of surface phytoplankton (upper 10 m) in the ASP showed a positive relationship between Fv/Fm, Pmax, and , indicating that phytoplankton acclimation to Fe availability, rather than light availability, is the main driver of photosynthesis rates in the ASP during the build-up of the phytoplankton bloom.

Surface phytoplankton Fv/Fm, Pmax, as well as , were highest in the southern ASP near the ice shelves (Stations 5 and 57), suggesting that the WW near the GIS, as well as meltwater-laden mCDW from the DIS, are major Fe sources to phytoplankton. On the other hand, Fv/Fm, Pmax and were all lower in the central polynya, suggesting that Fe availability limited photosynthesis rates where the phytoplankton biomass was highest. The Fv/Fm, Pmax and in the low salinity AASW strongly affected by recent sea ice melt (Stations 34 and 66) were intermediate, suggesting that melting sea ice may be an Fe source (Sedwick and DiTullio, 1997; Lannuzel et al., 2010), albeit not as pronounced as the ice shelves. The Pmax at Stations 5 and 57 close to the ice shelves (3.6 g C g−1 Chl a h−1) was 1.6-fold higher than in the central polynya (2.2 g C g−1 Chl a h−1), whereas close to the ice shelves (0.067 g C g−1 Chl a h−1 [µmol photons m−2 s−1]−1) was 1.9-fold higher than in the central polynya (0.035 g C g−1 Chl a h−1 [µmol photons m−2 s−1]−1). Both Pmax and close to the ice shelves exceeded values in the controls of the bioassay experiments, and matched those of the Fe treatments, as well as those in P. antarctica cultures growing under Fe-replete and saturating light conditions (Arrigo et al., 2010; Mills et al., 2010).

Curiously, there was no direct relationship between DFe concentrations in surface waters and any P-E parameter. High Pmax (> 3 g C g−1 Chl a h−1) and high (> 0.05 g C g−1 Chl a h−1 [µmol photons m−2 s−1]−1) were observed in waters with DFe as low as 0.11 nmol L−1. Similarly, DFe concentrations in the initial waters of the bioassay experiments did not alter the effects of Fe addition on phytoplankton growth. For instance, Fe addition did not affect phytoplankton growth in waters with the lowest initial DFe in this study (surface waters of Station 13, 0.09 nmol L−1 DFe), whereas initial DFe in the incubations that did show an Fe effect (Station 5 surface water, Station 35 subsurface water) were approximately twice as high at 0.18 and 0.22 nmol L−1 DFe, respectively. Thus, DFe concentrations do not appear to be a good measure of Fe availability for phytoplankton. A lack of direct DFe effects on phytoplankton photosynthesis and growth suggests that either not all DFe was bioavailable to the phytoplankton (Visser et al., 2003) or other sources of bioavailable Fe were present besides DFe (e.g., PFe). In addition, phytoplankton cells may have stored bioavailable DFe in excess of their immediate needs, resulting in Fe-replete cells in Fe-depleted waters.

Bioavailability of at least a fraction of PFe is suggested by the positive relationship between PFe and both Pmax and , indicating that PFe contributes to the pool of bioavailable Fe, or becomes bioavailable at high enough rates to support high photosynthesis rates. A small fraction of this PFe may be internal or attached to phytoplankton cells (Twining and Baines, 2013). Several studies show that Fe:C ratios of Fe-replete phytoplankton are a factor of 2 to 10 higher than those of Fe-limited phytoplankton (Twining et al., 2004; Hassler and Schoemann, 2009; Strzepek et al., 2012), suggesting that phytoplankton have the ability to take up DFe in excess of their immediate requirements and store it. Studies with P. antarctica suggest that metals may be stored inside P. antarctica colonies (Lubbers et al., 1990; Schoemann et al., 2001). The fraction of PFe internal or attached to cells is likely small, given that the highest PFe was observed at the face of the DIS where phytoplankton biomass was low.

Alternatively, extracellular, inorganic PFe in the ASP, such as crustal particles and Fe hydroxides (Hazarin et al., 2014), may be made biologically available through dissolution, photoreduction, and/or biological processing (Barbeau et al., 2003; Boyd and Ellwood, 2010; Rijkenberg et al., 2006, 2008; Boyd and Ellwood, 2010). Since EUML was relatively high throughout the ASP (Figure 3D), photoreduction is likely an active process in surface waters of the ASP. Photoreduction can convert bound Fe(III) species to Fe(II) via ligand to metal charge transfer (Barbeau et al., 2003; Rijkenberg et al., 2006). The P. antarctica bloom is associated with an abundance of relatively unsaturated organic Fe-binding ligands (Thuróczy et al., 2012) that may bind the Fe (II), keep it in solution (Rijkenberg et al., 2006, 2008), and make it available to P. antarctica. Specifically, polysaccharides enhance Fe bioavailability to both P. antarctica and diatoms (Hassler et al., 2011); as the main constituent of the matrix of Phaeocystis colonies (Alderkamp et al., 2007), polysaccharides may thus make PFe available to P. antarctica.

Effects of Fe limitation on phytoplankton photosynthesis at different light intensities

Phytoplankton acclimation to low Fe availability in the ASP will affect their photophysiology because of the high Fe requirements for chlorophyll biosynthesis and the high Fe content of the photosynthetic apparatus and electron transport pathways (Raven, 1990; Maldonado et al., 1999; Strzepek and Price, 2000). In general, acclimation to low Fe decreases the potential to maximize photon capture (Raven, 1990; Greene et al., 1992). Despite this interaction, the incubation experiments showed no interactive effects of Fe and light availability on photosynthesis rates. Two possible reasons for this lack of interaction may be that Antarctic phytoplankton do not increase their Fe quota to photoacclimate to low light (Strzepek et al., 2012), or that light-limited phytoplankton have lower growth rates and therefore require less Fe per unit time, i.e., a lower Fe flux.

For the first explanation for the lack of interaction between Fe and light availability on phytoplankton photosynthesis in the bioassay experiments, we consider photoacclimation strategies. Photoacclimation to low light may be achieved by either increasing the number of PSII reaction centers (RCIIs) or increasing the size of the photosynthetic pigment-containing antenna associated with the RCIIs; both strategies increase the cellular photosynthetic pigment concentrations. Increasing the number of RCIIs, however, increases the Fe requirement because the RCII and downstream electron transport chain are particularly Fe-rich: photosystem II (PSII) contains two or three Fe atoms, the cytochrome b6f complex (Cyt b6f) contains five Fe atoms, and PSI contains 12 Fe atoms (Raven, 1990). The PSII is the functional absorption cross section of PSII and can be used to distinguish between the two different photoacclimation strategies. The PSII is the product of absorption by the PSII antenna pigments and the probability that an exciton within the antenna will cause a photochemical reaction (Mauzerall and Greenbaum, 1989). The amount of pigment associated with each RCII will therefore determine much of the variability in PSII (Kolber et al., 1988; Suggett et al., 2004). Acclimation responses that alter the ratio of pigment:RCII will change the PSII, whereas acclimation responses that alter the number of RCIIs per cell but not the amount of pigments associated with it will not affect the PSII (Moore et al., 2006).

Light availability did not affect PSII in either the incubation experiments (Figure 7D) or the water column (Figure 5C), suggesting that the P. antarctica-dominated phytoplankton community did not change its antenna size during photoacclimation to low light, but rather increased the number of RCIIs, which would increase the cellular Fe requirement, contradicting the notion that Antarctic phytoplankton do not increase their Fe quota at low light (Strzepek et al. 2012). Our results match field results of PSII from the Ross Sea, where PSII did not increase with depth at P. antarctica-dominated stations (Smith et al., 2013). In contrast, P. antarctica did increase PSII under lower growth irradiance in culture experiments (Strzepek et al., 2012). Whether different strains of P. antarctica vary in their ability to adjust PSII or the timescales of photoacclimation in the field are too long to observe changes in PSII during four-day incubation is unknown.

Similar to light availability, Fe availability did not affect PSII consistently when control and Fe treatments were compared in the bioassay experiments. Moreover, there was no spatial pattern in PSII of surface phytoplankton in the ASP, despite the low Fe availability in the central ASP. These observations suggest that P. antarctica in the ASP does not increase its antenna size under Fe limitation. These results contrast with the increase in PSII under Fe limitation reported in culture experiments, where the PSII of Fe-limited P. antarctica was two-fold higher than Fe-replete P. antarctica under similar growth irradiance (Strzepek et al., 2012). The maximum response of PSII to Fe limitation in our incubation experiments was more subtle, with a 1.23-fold increase in control versus Fe treatments. Again, these differences between field observations and culture experiments could be due to (unknown) variability among different strains of P. antarctica in their ability to adjust PSII, or to timescales of acclimation.

The second explanation for the lack of interactions between Fe and light availability on phytoplankton photosynthesis rates is that light-limited phytoplankton have lower growth rates and therefore require less Fe per unit time. For our incubation experiments, this reduced requirement would mean that there was enough DFe in the original waters sampled to support slow-growing phytoplankton until day four of the incubation, but not enough to support fast-growing phytoplankton. In the field, if the DFe supply to a phytoplankton bloom is considered in terms of Fe flux, light-limited phytoplankton with low growth rates would require a lower Fe flux than phytoplankton with high growth rates under high light. Thus, if the DFe flux remains constant, phytoplankton under high light availability, as in the central ASP, may experience more Fe stress than light-limited phytoplankton that are growing more slowly, even if high light phytoplankton would have a lower Fe quota.

Water column productivity in the ASP driven by both Fe and light

Despite Fe limitation of photosynthesis rates, phytoplankton biomass and water column productivity was high throughout the ASP (2.1–6.5 g C m−2 d−1). These rates were similar to primary productivity rates reported by Lee et al. (2012) and Yager et al. (2014) during the same period in the ASP, and similar to rates measured later in the season in a previous year (Alderkamp et al., 2012a). Moreover, the high phytoplankton biomass and productivity match the high primary productivity rates calculated from satellite data in the ASP (Arrigo and Van Dijken, 2003; Arrigo et al., 2012; Yager et al., 2012). In addition, these high water column productivity rates were similar to those in the Pine Island Polynya in the eastern Amundsen Sea (Alderkamp et al., 2012a).

Stations with the highest water column productivity were all located in the moderately fresh AASW layer in the central ASP, where sea ice melt signatures were found enhancing stratification (MLD < 30 m) and increasing light availability. Although we could not find any direct relationship between water column productivity and parameters affecting light availability, such as MLD and EUML, autonomous glider observations at high spatial resolution within the central ASP found a positive relationship between phytoplankton biomass and enhanced stratification (shallower MLD) (Schofield et al., 2015), suggesting an important role for light availability in bloom formation. Stratification, resulting in enhanced light availability to phytoplankton, is especially important for bloom development early in the season (i.e., October–November) (Long et al., 2012), when incident irradiance is relatively low due to lower solar elevation and shorter days compared to the sampling period of this study (i.e. December–January).

The high water column productivity in the central ASP coincided with the stations where Pmax and were limited by Fe availability. In order to estimate the effect of Fe-replete photosynthetic parameters on water column productivity in the central ASP, we used the enhanced P-E parameters close to the ice shelves (Stations 5 and 57) for water column productivity calculations in the central polynya. This calculation showed that enhanced P-E parameters under high Fe availability have the potential to increase water column productivity in the central polynya on average by 1.7 ±0.01-fold, assuming all other factors stay the same (e.g., light availability, temperature, grazing, phytoplankton species composition). Thus, increased Fe availability during the build-up of the bloom would significantly increase the already high rates of phytoplankton productivity in the ASP.

Many Antarctic ice shelves are thinning, including the Dotson Ice Shelf, as the glaciers that feed these ice shelves are accelerating (Pritchard et al., 2009; Rignot et al., 2013). An increase in basal ice shelf melt would most likely be driven by an increase in the amount of mCDW flowing onto the continental shelf and/or an increase in the heat content of mCDW (Jacobs et al., 2011). An increase in meltwater-laden mCDW exiting the ice shelve cavities would destabilize the water column, potentially increasing MLD and decreasing light availability. Early in the growing season when incident irradiance is relatively low, this scenario would potentially delay bloom formation and decrease water column productivity, especially near the ice shelf. However, if the DIS is a source of bioavailable Fe for the ASP, similar to that observed for the Pine Island ice shelf in the PIP (Gerringa et al., 2012), then horizontal diffusivity (Geringa et al., 2012), advective eddy transport (e.g., Årthun et al., 2013), mixing along the Dotson trough (e.g., St-Laurent et al., 2013), and wind- and iceberg-induced mixing (Randall-Goodwin et al., 2015) would ensure a bioavailable Fe flux from the DIS or neighboring ice shelves to the high phytoplankton biomass in the central ASP, where the water column is stabilized by sea ice melt water and higher Fe availability would increase the phytoplankton photosynthesis and water column productivity substantially.

The presence of mCDW has been detected regularly in the troughs of the ASP since observations began (Jacobs et al., 2012; Arneborg et al., 2012; Wåhlin et al., 2013; Ha et al., 2014), suggesting this water mass has regular access to the DIS. However, any seasonal (Thoma et al., 2008; Jacobs et al., 2012; Randall-Goodwin et al., 2015) or annual (Jacobs et al., 2013; Dutrieux et al., 2014) variability in mCDW on the continental shelf will likewise affect the quantity of meltwater-laden mCDW from the DIS, and hence the DFe flux to the ASP. Variability in the DFe flux would affect phytoplankton photosynthetic rates directly according to our results. How phytoplankton productivity across the ASP would be affected by variability in mCDW access to the DIS is unknown, but represents a key question for future investigations.

Data accessibility statement

Data are available in the BCO-DMO database: http://www.bco-dmo.org/dataset/546372.

Copyright

© 2015 Alderkamp et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Contributions

Substantial contributions to conception and design: ACA, GLD, PLY, KRA

Acquisition of data: ACA, GLD, KEL, TLC, ML, RMS, CH, ER, SES

Analysis and interpretation of data: ACA, GLD, KEL, TLC, RMS, OS, SES, PLY, KRA

Drafting the article or revising it critically for important intellectual content: all authors

Final approval of the version to be published: all authors

Competing interests

Authors declare that there are no competing interests.

Funding information

This research was supported by the National Science Foundation Office of Polar Programs, Antarctic Organisms (ANT-0944727 to KRA, ANT-0839069 to PY, ANT-0838995 to RMS and OS, and ANT-0838975 to SS).