Limnol. Oceanogr., 58(3), 2013, 1035–1047
2013, by the Association for the Sciences of Limnology and Oceanography, Inc.
doi:10.4319/lo.2013.58.3.1035
E
Wintertime controls on summer stratification and productivity at the western
Antarctic Peninsula
Hugh J. Venables,* Andrew Clarke, and Michael P. Meredith
British Antarctic Survey, Madingley Road, Cambridge, United Kingdom
Abstract
We report results collected year-round since 1998 in northern Marguerite Bay, just inside the Antarctic Circle.
The magnitude of the spring phytoplankton bloom is much reduced following winters with reduced sea-ice cover.
In years with little winter sea-ice the exposed sea surface leads to deep mixed layers in winter, and reduced watercolumn stratification the following spring. Summer mixed-layer depths are similar, however, so the change is not
in overall light availability but toward a less stable water column with greater vertical mixing and increased
variability in the light conditions experienced by phytoplankton. Macronutrient concentrations are replete at all
times, but the increased vertical mixing likely reduces iron availability. The timing of bloom initiation is similar
between heavy and light ice years, occurring soon after light returns in early spring, at a mixed-layer averaged
light level of , 1 mol photon m22 d21. Ongoing regional climate change in the WAP area, and notably the
ongoing loss of winter sea-ice, is likely to drive a downward trend in the magnitude of phytoplankton blooms in
this region of the Antarctic Peninsula.
Within the Southern Hemisphere, the region that is
changing most rapidly is the western Antarctic Peninsula
(WAP), which has experienced one of the strongest
warmings of anywhere in the world (King 1994; Vaughan
et al. 2003). The annual mean atmospheric warming at the
WAP is 3.7 6 1.6uC century-1 unweighted, or 3.4uC
century-1 when weighted by length of record (Vaughan
et al. 2003). This warming is most pronounced in austral
autumn and winter (Turner et al. 2005). Although the full
mechanisms controlling the warming have not yet been
elucidated, it has been linked to atmospheric circulation
changes. Warm winters at the WAP are associated with
more cyclonic circulation and increased warm air advection. Increased precipitation at the WAP in recent decades
(Thomas et al. 2008) is consistent with the hypothesis that
the warming trend is accompanied by a shift toward more
cyclonic conditions.
Increased temperatures and shifts in the wind pattern
have affected the length of the sea-ice season along the
WAP. Sea-ice duration has declined significantly to the
west of the Antarctic Peninsula. Changes in duration show
a strong trend toward a later advance in autumn and a
weaker trend toward an earlier retreat in spring (Stammerjohn et al. 2008). Superimposed on this trend are
anomalous years of more extreme conditions that are
linked to variations in coupled modes of climate variability,
such as the Southern Annular Mode (SAM) and the El
Nino–Southern Oscillation (ENSO) phenomenon (Meredith et al. 2004; Massom et al. 2006). In addition to reduced
sea-ice extent, 80% of glaciers on the Antarctic Peninsula
are retreating, and retreat rates are accelerating (Cook et al.
2005). This retreat includes the Sheldon Glacier that flows
into Ryder Bay at the northern edge of Marguerite Bay
(Fig. 1; Peck et al. 2010).
* Corresponding author: [email protected]
The ocean also affects regional warming via the on-shelf
flow of Circumpolar Deep Water (CDW) from the
Antarctic Circumpolar Current, which lies immediately
adjacent to the continental shelf break (Klinck 1998) at the
WAP. The CDW water mass is warmer than surface water,
being initially < 2uC before modification by mixing with
cooler shelf water. Because of strengthening westerly winds
associated with an increasing trend in the SAM, it is likely
that the on-shelf flow of warm CDW has increased in
recent decades, with an associated rise in heat flux
(Martinson et al. 2008).
The effects of reducing sea-ice cover on local hydrography and productivity have been studied at various locations
in the Arctic and Antarctic (Clarke et al. 2007; Tremblay
et al. 2008). The effect on productivity can be both positive
and negative, depending on the local conditions and the
date of ice breakup. A shift to open water from permanent
ice cover removes the shading effect of ice and can lead to a
significant increase in productivity (Arrigo et al. 2008;
Montes-Hugo et al. 2009). If, however, the presence of ice
early in the growth season increases stratification, then a
reduction in productivity can result (Vernet et al. 2008;
Montes-Hugo et al. 2009). Changes in salinity, from
changes in sea ice or glacial melt, can also affect
phytoplankton community composition (Moline et al.
2004). The relationship between sea-ice dynamics and
biological productivity is thus complex. Changes in ice
cover also have implications for higher trophic levels
(Ducklow et al. 2007; Schofield et al. 2010).
Zooplankton, especially krill, are dependent on sea ice
(Atkinson et al. 2004) and highly sensitive to sea-ice
changes (Wiedenmann et al. 2009). However, consequences
of altered sea-ice conditions on zooplankton populations
are extremely hard to predict with any accuracy. This
potential for change along the Antarctic Peninsula highlights the importance of long-term studies of the phytoplankton community (Schloss et al. 2012) and the export
1035
1036
Venables et al.
Fig. 1. Map of sampling area: (a) the position of Rothera research station on Adelaide Island, Antarctic Peninsula, with the 1000 m
contour marked; and (b) the sampling site in Ryder Bay close to the base.
from the surface layers (Ducklow et al. 2008) because these
may change rapidly, with effects on carbon sequestration
and higher trophic levels.
The extreme seasonal cycle in light availability in northern
Marguerite Bay (Fig. 1), which is within the Antarctic
Circle, provides an opportunity to study bloom initiation in
a system that must be light-limited for the 3 weeks when the
sun does not rise. That is, when light is zero the growth rate
must be less than or equal to zero. Following this period, the
growth rate must increase through zero, taken to be the
point of bloom initiation. Classically, bloom initiation
occurs when shoaling mixed-layer depths and increasing
irradiance lead to phytoplankton photosynthesis exceeding
total system respiration, including zooplankton grazing of
phytoplankton (Sverdrup 1953). This theory leads to the
definition of a ‘critical depth,’ such that net phytoplankton
growth occurs only if the mixed-layer depth is shallower than
the critical depth. The derivation of the critical depth
includes the incoming irradiance and therefore varies
seasonally with changing light levels. This view is modified
by the fact that a drop in mixed-layer turbulence must
precede restratification (Townsend et al. 1992; Huisman
et al. 1999). Accordingly, processes that act to increase early
season stratification are predicted to lead to an earlier start
to a phytoplankton bloom (Long et al. 2012).
An alternative view of bloom initiation has been
proposed that considers seasonal changes in phytoplankton
losses to be more significant than light-driven increases in
phytoplankton-specific growth rates (i.e., cell divisions per
d) in setting the net phytoplankton growth rate (Behrenfeld
2010). According to this view, referred to as the ‘dilutionrecoupling hypothesis,’ physical processes in early winter
(in particular, mixed-layer deepening) disrupt the balance
between phytoplankton growth and losses, thereby allowing net biomass accumulation despite maximum mixing
depths and low light. During subsequent mixed-layer
stratification, phytoplankton growth and loss rates become
increasingly similar (i.e., ‘recoupled’), such that springtime
increases in net population growth rates are not proportional to light-driven increases in specific growth rates
(Behrenfeld 2010; Boss and Behrenfeld 2010).
In the open Southern Ocean, high-nutrient low-chlorophyll (HNLC) conditions prevail in most areas due to iron
limitation (De Baar et al. 2005; Jickells et al. 2005). These
HNLC conditions likely favor phytoplankton of small cell
size that are more easily grazed by zooplankton (Morel
et al. 1991). Chlorophyll concentrations in these areas are
consistently below 0.5 mg m23. It is less clear which factors
control the upper limit to bloom magnitude (i.e., maximum
phytoplankton concentration) in areas where iron limitation is lifted, such as downstream of some Sub-Antarctic
islands or during iron addition experiments. Possible
controlling factors include light limitation through selfshading (De Baar et al. 2005), photoinhibition (Alderkamp
et al. 2010), or the return of iron stress due to phytoplankton uptake (Venables and Moore 2010).
In this paper, we use a 14 yr record of water-column
measurements and sea-ice observations to investigate the
processes linking sea-ice cover, stratification, and summer
chlorophyll at a WAP site (Fig. 1). The importance of
wintertime sampling is that it allows full investigation of
the preconditioning processes that can influence the timing
and magnitude of the subsequent spring bloom, including
assessment of the effects of light limitation and mixed-layer
processes on phytoplankton.
Methods
Oceanographic sampling has been undertaken from the
British Antarctic Survey’s Rothera Research Station
Wintertime controls on spring bloom
(Fig. 1) since 1998, as part of the Rothera oceanographic
and biological time series (RaTS; Clarke et al. 2008). The
primary RaTS site at 67.570uS 68.225uW (Fig. 1) is located
in Ryder Bay, < 4 km from Rothera base. Water depth at
the RaTS site is 520 m. The south-facing aspect of the bay
makes it particularly sensitive to northerly winds blowing
ice away from the coast (Meredith et al. 2004, 2008, 2010).
Sampling is undertaken from a small boat, or through a
hole cut in fast ice. Sampling is targeted to be twice weekly
in summer and weekly in winter, although weather
conditions sometimes prevent this being achieved. There
are two large gaps in the 14 yr time series. The first major
gap in late 2000 was caused by persistent unworkable ice
(too heavy to allow passage by boat, but not passable on
foot). This problem is also responsible for a number of
other smaller gaps in the time series, unfortunately often in
early spring. The second major gap occurred when
equipment was lost in a fire in September 2001. All gaps
of . 30 d duration are considered here as breaks in the time
series. When the standard RaTS site cannot be occupied
(e.g., due to partial sea-ice cover), a secondary site is used
that has a depth of 300 m and is located at 67.581uS
68.156uW.
During each RaTS sampling time point, vertical
profiling measurements are made using a SeaBird 19+
conductivity temperature depth instrument (CTD), a
WetLabs in-line fluorometer, and a LiCor photosynthetically available radiation (PAR) sensor. Two replicate
profiling packages are used to support the time-series
measurements, with one package being deployed in the
field while the other is being serviced and recalibrated.
Prior to January 2003, a Chelsea Instruments Aquapak
CTD and fluorometer were used. During this period,
profiling was limited to the top 200 m by the pressure rating
of the instrument.
Vertical profile sampling consists of a full-depth CTD
and discrete water sample collection at 15 m. Continuous
profile data are collected for temperature, conductivity,
pressure, fluorescence and PAR. An annual joint cast is
also made with the RaTS profiling system and the SeaBird
911+ CTD package on the R/V Laurence M. Gould during
the Palmer Long Term Ecological Research (LTER) grid
survey (Ducklow et al. 2007). Instrument offsets identified
during these joint casts are applied to the RaTS salinity
data to reconcile them with the LTER 142 data and our
water samples. Given the range of offsets found and the
accuracy of the SeaBird 911+ CTD package, we estimate
that the temperature is accurate to 0.002uC and salinity to
0.005. The discrete water samples collected at 15 m depth
are used for salinity calibration and measurement of
chlorophyll, macronutrients, and a range of other properties (Clarke et al. 2008).
Chlorophyll data presented in the current paper are
derived from the CTD fluorescence sensor data calibrated
against the 15 m extracted chlorophyll sample values. These
chlorophyll profile data have been reprocessed due to the
discovery of a nonlinear sensor response at high values.
Thus, the current data are slightly different from those
presented previously (Clarke et al. 2008), but exhibit the
same qualitative patterns. Additional details regarding
1037
RaTS sampling are described in Clarke et al. (2008).
Relevant to the current study, the RaTS site is always
replete in macronutrients and seasonally persistent high
chlorophyll concentrations observed in some years strongly
suggest a seasonally persistent iron supply. A likely source
for this enhanced iron is from glacial melt, as observed in
the Amundsen Sea (Alderkamp et al. 2012; Gerringa et al.
2012).
Mixed-layer depths, unless otherwise stated, are taken as
the depth where the density difference (Ds) relative to the
surface is 0.05 kg m23. Stratification is quantified here as
the potential energy required to be added to homogenize
the water column from the surface to a given depth or
between two depths. This metric of stratification is the
negative of the potential energy anomaly (Simpson et al.
1978) and has units of joules.
Sea-ice coverage and type is characterized each day in
the three bays around the Rothera station. Sea-ice coverage
has been previously discussed in relation to assessing
freshwater inputs into Ryder Bay (Meredith et al. 2008,
2010). The characterization of sea ice is undertaken by a
series of marine assistants and thus is, inevitably, somewhat
subjective. However, the seasonal and interannual changes
in sea-ice coverage are far greater than the inter-investigator differences in assessment and the Rothera records
agrees well with wider scale satellite-derived trends
(Wallace 2007).
Daily PAR data for our 14 yr time series are estimated
from a Bentham spectroradiometer, with major data gaps
during 2002, 2012, and most of 2003. Days with partial
data are excluded. Bentham spectroradiometer data are
only for the wavelength interval 400–600 nm and are
recorded in W m22. Irradiance in units of mol photons
m22 d21 was calculated for each wavelength (0.5 nm
resolution) and 30 min resolution data integrated to create
daily values. To achieve PAR estimates for the full 400–
700 nm range, the 600–700 nm photon flux was estimated
from the measured flux at 500–600 nm. The sunlight energy
spectrum is relatively flat between 500 and 700 nm
(Gueymard et al. 2003); therefore, this was done by
multiplying the 500–600 nm range by 1.18 to account for
the reduction in energy per photon at higher wavelengths.
This value was then added to create an estimate for the full
400–700 nm photon flux. Downwelling diffuse attenuation
was calculated as the coefficient that gave the same depthintegrated PAR as was observed in the profile. This largely
(though not completely) accounts for the through-ice
profiles, where ice shading is significant. Any light that
passes through the deployment hole cut in the ice is
included, and this leads to water-column light estimates
being slightly higher than they would be without the hole.
When sampling is from a boat, this method agrees very
closely with the standard method of estimating attenuation
from fitting a regression line to the logarithm of the PAR
profile.
From daily PAR data and estimated values for the mean
diffuse attenuation coefficient for the water column, the
vertically averaged light availability for the surface (Ī S) was
calculated for each cast. For the calculation of Ī S, depthresolved light values were averaged over the greater of
1038
Venables et al.
either the mixed-layer depth or the top 20 m, because the
mixed layer is often very shallow but enhanced chlorophyll
concentrations always extent to 20 m deep, and normally
deeper. This process makes no difference compared with
just using the mixed-layer depth during bloom initiation
because mixed-layer depths are . 20 m, but makes the
surface light availability more comparable later in the
season when the mixed-layer depth can be completely
absent.
Growth rates were calculated largely following Behrenfeld (2010), although no attempt was made to convert to
carbon units. To calculate the net rate of chlorophyll
accumulation, the rate of change of either vertically
integrated chlorophyll (when the maximum of the mixedlayer depth [MLD] and photic zone depth [PZD] deepens)
or surface chlorophyll concentration (when the mixed-layer
shallows). This approach accounts for dilution of the
phytoplankton population over an increasing water volume
during mixing events and for phytoplankton being left
below the mixed layer when stratification increases. The
vertical integration is over the maximum of the MLD and
PZD. The PZD was defined as the depth that received a
daily averaged light of 0.2 mol photons m22 d21 because
this was the best match to the vertical extent of
phytoplankton in stratified periods, but the results are
very insensitive to this number. Our data are taken during
daylight and are therefore affected by quenching of the
fluorescence, so the surface value used is the 10–20 m mean
chlorophyll concentration. This process calculates the net
rate of chlorophyll accumulation, r, which is the difference
between the specific growth rate (cell divisions per d) and
the specific loss rate (phytoplankton respiration plus
grazing and viral losses).
The vertical integration depth for both profiles was the
depth from the second cast. This depth is, by construction,
deeper (if it was shallower, then the chlorophyll concentration would be used rather than the depth-integrated
chlorophyll) and so accounts for entrainment of deep
chlorophyll from the first cast. It is particularly important
when there is a shallow mixed layer that then deepens again
into the remains of the previous deeper mixed layer that
still has enhanced chlorophyll concentrations.
The effects of advection of spatially varying chlorophyll
concentrations past the sampling site contribute to noise in
r. This effect is an inevitable feature of all time-series sites,
but the length of the series means that the positive and
negative effects should largely compensate.
Results
Physical changes—Figure 2a shows the time series of ice
cover in Ryder Bay together with the MLD. Figure 3a
shows that reduced duration of fast ice for some part of the
winter (especially yr 1998, 2003, 2008–2011) leads to deep
mixed-layer depths during the period the bay is ice-free.
The reduction of the ice cover is discussed in detail by
Meredith et al. (2010), who showed that this related largely
to an increase in northerly winds blowing ice out of the
south-facing bay. This change in the winds is partly linked
to the intensification of the SAM, which has a mainly zonal
structure over the Southern Ocean but includes nonzonally
symmetric components, including in the region of the WAP
(Thompson and Solomon 2002; Lefebvre et al. 2004).
ENSO is also noted to have a marked effect on meridional
winds in this region (Meredith et al. 2010).
On the local scale, persistent northerly winds lead to the
area close to Rothera taking the characteristics of a small
polynya, with the surface of the water exposed to cold air
temperatures and strong winds. This increased air–sea flux
of heat and mechanical energy leads to the mixed layer
deepening through three interlinked processes at the sea
surface: loss of buoyancy through cooling, brine rejection
from ice formation, and mechanical mixing due to the wind
stress (Wallace et al. 2008; Meredith et al. 2010).
Figure 2b,c show the full time series of potential
temperature and salinity at the RaTS site down to 200 m.
The effect of the deep mixed layers and enhanced watercolumn cooling are evident. Surface salinity is lower in
high-ice years because the brine released from sea-ice
formation is distributed vertically over a large depth range,
where the subsequent ice melt is retained initially in a
shallow stratified layer at the surface (Meredith et al. 2008).
Figure 2b,c show the full time series of potential
temperature and salinity at the RaTS site down to 200 m.
The deep mixing events can be seen because they cool the
water down to depths of 150–200 m and also homogenize the
salinity in these depths as well. Salinity dominates density at
these low temperatures, so the increased mixing of salinity
significantly reduces the stratification in the upper 200 m
during the winter. Stratification is returned to the system in
the spring and summer through melting sea ice, glacial
freshwater runoff (Meredith et al. 2008), and surface warming
(Fig. 2b). The spring increase in stratification is from a lower
winter baseline, so there is reduced stratification during the
spring and summer following deep winter mixing (Fig. 3b).
In order to analyze the time series, it has been split into
deep winter mixing (DWM) years, with maximum winter
MLD . 100 m, and shallow winter mixing (SWM) years.
The deep winter mixed layers are due to significantly
reduced ice extent (Fig. 3a), and so categorizing years on
the basis of the depth of mixing also largely categorizes
them on the basis of ice extent in spring. One partial
exception to this is 2004, which had very deep mixed-layer
depths in midwinter, but during which fast ice and
shallower mixed-layer depths had returned by early spring.
In this analysis, 2004 is included in the DWM years, but
the following results are qualitatively unchanged if it is
included with SWM years.
The annual cycle of stratification (Fig. 4) shows the
lower winter stratification (mean 5 19J for d of yr 150–250)
in years with DWM, due to reduced ice cover. This is
significantly different from years with SWM and extended
periods of fast ice in winter (mean 5 84 J for d of yr 150–
250, t 5 4.5, p , 0.0005). The increase in stratification in
November and December (d of yr 305–365) is similar (3.3
6 1.2 J d21 in DWM yr, 2.7 6 2.2 J d21 in SWM yr),
though from the different winter levels, showing that
changes in winter strongly affect the summer conditions.
There is a noticeable, though not statistically significant,
divergence in the SWM and DWM stratification in January
Wintertime controls on spring bloom
Fig. 2.
1039
Time series of (a) ice (not including brash ice) and mixed-layer depth, (b) potential temperature, and (c) salinity in Ryder Bay.
(3.7 6 4.2 J d21 in DWM yr, 10.7 6 7.3 J d21 in SWM yr).
Reduced stratification leads to the near-surface layer being
preconditioned toward greater vertical mixing for the same
energy input from wind stress and tides.
Bloom initiation—The Rothera time series, with frequent
measurements through the winter from within a polar
circle, provides a unique opportunity to assess the annual
cycle of phytoplankton net growth rate, r, in a region that
must be light-limited for some period during the year. The
two regimes of winter MLD (deep and shallow) provide a
further opportunity to assess the role of the mixed-layer
depth changes in the start and early phases of the
phytoplankton bloom. DWM years have maximum winter
mixed layer depths . 100 m and mean winter mixed-layer
depths . 60 m, significantly deeper than the vertical extent
of summertime enhanced chlorophyll. The MLD then
shoals through spring. This is similar to open-ocean
conditions. In contrast, years with fast ice cover and
SWM have winter mixed-layer depths of 10–50 m, approximately the same depth as the depth of enhanced
chlorophyll in the summer. However, despite the shallow
mixed-layer depths, shading by the ice reduces light
availability so that mixed-layer averaged light availability
is very similar to DWM years.
Figure 5b shows that r becomes positive in very early
spring. Both DWM and SWM years show a very similar
date (d of yr 228; 16 August) at Ī S < 0.6 mol photon
m22 d21. We note that the mixed layers are close to
maximal in both high and low ice regimes at this point, but
the increase in incoming irradiance is sufficient to increase
light availability.
The data are unfortunately most sparse at this point
in the year (Fig. 5a) because of the difficult ice conditions
1040
Venables et al.
Fig. 3. (a) Relationship between winter fast ice and winter MLD. The 2000 and 2001 seasons
are plotted as open circles and not included in the regression because there was no water-column
sampling between late August and early December in either year (n 5 12, r2 5 0.87, p , 0.001).
(b) Relationship between winter MLD and subsequent summer stratification (strat.; n 5 14,
r2 5 0.51, p , 0.01).
often encountered. The linear trend in r gives extra
confidence in the timing found, but there are few data
until d of yr 253. On this date, the light levels, Ī S, for SWM
and DWM were 1.0 and 1.8 mol photon m22 d21,
respectively. These levels are an upper bound to the light
levels that lead to bloom initiation.
Chlorophyll concentrations are similar between DWM
and SWM years for < 3 months after r become positive. In
this period, the deep mixed layers in DWM years mean that
the depth-integrated chlorophyll concentrations are higher
than in SWM years (Fig. 6, top panel). There is, however,
no statistically significant difference in r between regimes at
this time.
It has been suggested that bloom initiation will occur
when turbulent overturning in the mixed layer declines,
which must precede water-column stratification (Townsend
et al. 1992; Huisman et al. 1999). Light estimates from the
MLD may therefore underestimate the effective light
experienced by the phytoplankton at the time of bloom
initiation. Inspection of the time series and each individual
profile suggested that the mixed-layer depth used was often
a good match for the depth of enhanced chlorophyll. For
some casts, using Ds 5 0.02 kg m23 as a MLD criterion led
to a closer match between MLD and the vertical extent of
enhanced chlorophyll. Using these shallower MLDs, the
light availability at the time of initiation increased slightly,
to Ī S < 0.75 mol photon m22 d21.
The growth rate r continues to rise with increasing light
and water-column stability for about 6 weeks after bloom
initiation. This is consistent with the increased specific
growth rate possible from increased incoming irradiance,
which initially more than compensates for any increase in
loss terms due to the shallowing mixed layers and seasonal
increase in zooplankton grazing.
Bloom magnitude—Figure 7 shows the time series for
15 m chlorophyll. It is clear that during recent summers
(2008–2010 especially) and the early part of 1998 show
much lower chlorophyll concentrations than other years.
Figure 6 (bottom panel) shows this, split between the
DWM and SWM years, and Fig. 6 (top panel) shows a
similar pattern in depth-integrated chlorophyll from d of yr
Fig. 4. Annual cycle of near-surface stratification split between years with deep and shallow
winter mixing. The lines are 30 d running means.
Wintertime controls on spring bloom
1041
Fig. 5. Seasonal cycle of (a) net growth rate, r, at Ryder Bay, calculated from chlorophyll
measurements, and (b) mixed-layer light availability (Ī S), growth rate, and mixed-layer depth with
a 30 d running mean. In both plots, data are split between seasons preceded by either deep or
shallow winter mixing.
340 onward. It therefore seems very likely that either the
change in ice cover or MLDs has affected the phytoplankton in Ryder Bay (Fig. 8).
Ice can affect phytoplankton in a range of ways. It can
hold a population of ice algae, which could seed the bay in
spring (Ackley and Sullivan 1994). Loss of the ice could
therefore lead to a reduction in this process. This would
not, however, explain the prolonged seasonal shift,
especially given that there are some short-term increases
in chlorophyll that could act in a similar manner. It is also
not consistent with the observed shifts in dominant species
within a season (Annett et al. 2010).
Another possibility is that the presence or absence of ice
significantly changes the light availability. The presence of
ice cover in spring initially reduces Ī S, through shading of
the water column (Fig. 5b), and then increases Ī S through
creating a shallow mixed layer on melting. Due to the
reduced self-shading by phytoplankton, the light availability, Ī S, in low chlorophyll years is higher than in high
chlorophyll years (Fig. 9a). The lower chlorophyll years
have higher light; therefore, low light availability cannot
explain the low chlorophyll conditions.
The coastal nature of the site means that there is a supply
of glacial meltwater throughout the season, which acts to
add at least as much freshwater as sea-ice melt (Meredith
et al. 2008). This means that the MLDs quickly reduce in
the spring in all years to the point where they are much
shallower than the photic zone. Indeed, in many summer
casts, there is no mixed layer readily observable at the
surface of the water column.
1042
Venables et al.
Fig. 6. Seasonal cycle of chlorophyll at RaTS split between years with deep winter mixing (limited ice cover) and shallow winter
mixing (prolonged ice cover). (top panel) Vertically integrated chlorophyll, over the deeper of the mixed-layer and photic zone depths,
and mixed-layer depth; (bottom panel) near-surface chlorophyll concentrations and photic zone depth.
Despite the effect of meltwater, there is a prolonged
effect of deep winter mixed layer into the summer.
Figure 9b shows the stratification in the upper 30 m and
the chlorophyll concentrations. A reduction in stratification leads to a more unstable water column, preconditioned
toward greater vertical mixing for the same energy input.
Greater vertical mixing will have two significant effects.
The iron supply is likely to have a significant component
entering directly into the photic zone, from both glacial
meltwater (Alderkamp et al. 2012; Gerringa et al. 2012) and
Fig. 7.
the nearshore zone (Planquette et al. 2007). Increased
mixing will dilute this supply and transport some of it
downward, out of the photic zone. The decreased watercolumn stability also increases the variability in the depth
of phytoplankton and so reduces their ability to fully adapt
to the light level they are exposed to. Overall, there is a
significant relationship of increasing summer chlorophyll
with increasing stratification (Fig. 10). Figure 6 (top panel)
shows that the low chlorophyll years have the highest
short-term net growth rates, but also the greatest net loss
Time series of 15 m chlorophyll and mixed-layer depth.
Wintertime controls on spring bloom
Fig. 8. Relationship between winter mixed-layer depth and
summer chlorophyll concentrations. 2000 and 2001 are shown as
open circles (and excluded from the regression) because there were
only two winter casts in each year, all in August. (n 5 12,
r2 5 0.62, p , 0.005).
rates. The variances of r are significantly different between
DWM and SWM years (p , 0.05) but the means are not
significantly different. This greater spread of net growth
rates is indicative of the more unstable and variable
conditions experienced.
1043
Fig. 10. Relationship between summer stratification and
summer chlorophyll (n 5 14, r2 5 0.59, p , 0.005).
The higher integrated chlorophyll in DWM years may
also affect the subsequent bloom magnitude. Assuming
that the increased chlorophyll is indicative of increased
biomass, there is the potential for an earlier increase in
zooplankton populations within the mixed layer. This
allows extra time for zooplankton growth; and through the
spring, the zooplankton (capable of vertical movement) will
Fig. 9. Time series of 15 m chlorophyll concentrations with (a) mixed-layer average light availability, and (b) stratification strength
in the top 30 m.
1044
Venables et al.
concentrate in the shallowing mixed-layer depth and thus
lead to an increased grazing pressure (Boss and Behrenfeld
2010). This effect would be of greater importance if iron
levels were also reduced, because this may also lead to
smaller average cell sizes and thus greater potential for
grazing loss (Morel et al. 1991).
The low r observed around midsummer in high
chlorophyll years (Fig. 6, bottom panel) masks significant
cellular growth within the community structure. There are
many shifts in the dominant species during this time
(Annett et al. 2010), showing that some species must be
growing while others are declining due to grazing or viral
losses or naturally due to their reproductive cycle.
Discussion
Through the 14 yr of sampling at RaTS, there have been
significant changes in the duration of fast sea ice in the winter.
Low sea-ice years lead to deep winter mixed-layer depths,
which are in turn followed by summers with lower nearsurface stratification. The variations in the physical conditions observed through the time series, combined with the
extreme seasonality due to the location within the Antarctic
Circle, provide a valuable test of the effect of mixed-layer
depth, light availability, and dilution on bloom initiation.
In contrast to the theory (Sverdrup 1953) that net
phytoplankton growth rates are controlled in spring by
increasing light availability increasing the specific growth
rate, it has been proposed recently that bloom initiation is
driven more by changes in phytoplankton loss rates
(Behrenfeld 2010; Boss and Behrenfeld 2010). This study,
based on data from the North Atlantic, builds on the work
of (Evans and Parslow 1985) in focusing on the effects that
mixed-layer deepening have on the concentration of
plankton and thus grazing losses relative to the effects of
light availability on specific growth rates.
The data show that bloom initiation occurs at approximately the same time and light levels in each of the two
winter regimes, despite the very different mixed-layer
depths. The increase in light is also driven mainly by the
increase in incoming irradiance rather than changes in
mixed-layer depth within a year or between years with
significantly different mixed-layer depth cycles. The light
availability is thus not strongly connected to changing loss
terms, as suggested in (Behrenfeld 2010). Overall the
classical model for bloom initiation being driven by
increased light availability (Sverdrup 1953) appears to hold
in this circumstance. In this extremely seasonal location, it
is the incoming irradiance that drives a deepening of the
critical depth with time until it is deeper than the mixedlayer depths. The mixed-layer depths themselves stay
relatively constant at the time of bloom initiation, but then
shallow with the input of meltwater. Areas with higher
midwinter light availability or more active frontal structures (Huisman et al. 1999; Taylor and Ferrari 2011a) may
have different controls.
The light level associated with an increase in chlorophyll
is very low, suggesting that there is no significant delay
from other processes limiting the response of phytoplankton to increased irradiance. It is lower than the lowest light
level recorded in the North Atlantic (48–52uN; Boss and
Behrenfeld 2010) but higher than midwinter light levels
recorded in the some parts of the Southern Ocean in
midwinter (Venables and Moore 2010). It is also about half
the value found in a study of the North Atlantic bloom
from satellite data (Siegel et al. 2002), though these
comparisons with other locations are limited by lack of
knowledge about potential differences in loss terms
between these location. There is enhanced chlorophyll
throughout the mixed layer in early spring, even when
mixed layers are deep (. 80 m). This suggests that
reduction in turbulence within the mixed layer or minor
restratification events below the mixed-layer depth density
threshold are not key for setting the timing of bloom
initiation in this location (Taylor and Ferrari 2011b).
It should be noted that it is chlorophyll that has been
measured and there can be significant changes in chlorophyll : carbon ratios in phytoplankton (Landry et al. 2000;
De Baar et al. 2005); hence, the early positive net growth
rates calculated may actually reflect (at least partly)
synthesis of new chlorophyll from internal reserves within
phytoplankton cells rather than a net uptake of energy.
Chlorophyll can presumably serve little or no function to a
phytoplankton cell in the middle of a polar winter when the
sun does not rise.
The more stable environments are closer to optimal
conditions for phytoplankton growth, leading to the very high
concentrations and significant reduction in light availability
through self-shading. The less stable conditions, preconditioned to a greater extent of vertical mixing for the same
energy input, are similar to open-ocean conditions and the
‘low’ chlorophyll concentrations in this study are comparable
with ‘high’ concentrations found in iron-fertilized regions of
the Southern Ocean such as South Georgia (Park et al. 2010),
Kerguelen (Mongin et al. 2008), or Crozet (Venables et al.
2007). The limitation of bloom magnitude discussed here is
therefore qualitatively different from the limitation of
phytoplankton to the HNLC conditions prevalent across
most of the Southern Ocean (Platt et al. 2003; Alderkamp
et al. 2010; Venables and Moore 2010). For coastal waters, the
question is thus what processes set an upper limit to
chlorophyll concentrations in significant blooms (Vernet
et al. 2008). The links to spring mixed-layer depth and
summer chlorophyll have been demonstrated here and further
research is underway to assess the contribution of each to
setting the bloom magnitude.
A further question relating to seasonal cycle is the cause
of the decline of the phytoplankton bloom in autumn. The
greatest loss rates are while the MLD is deepening to about
30 m. This depth is similar to the depth of the enhanced
chlorophyll through the summer, and mixed-layer averaged
light levels are considerably higher than levels observed at
the start of the spring bloom. This suggests the declines,
which are similar in chlorophyll concentrations and in the
depth-averaged chlorophyll (Fig. 6) cannot be explained
simply through dilution or light limitation. Possibilities for
the decline include seasonally changing grazing rates with
maxima late in season before winter diapauses, viral
infections of cells, or a reduction in iron supply due to
the reduced melting and freshwater input. These, and other
Wintertime controls on spring bloom
aspects of the seasonal cycle, will be studied over the
coming seasons.
The year-round sampling of the marine environment
close to the Antarctic Peninsula has allowed a unique
assessment of the full annual cycle of phytoplankton in one
of the most climatically sensitive regions of the world. It
has shown that bloom initiation, defined as the start of a
net positive increase in chlorophyll, occurs at extremely low
light levels very early in the season (0.6 mol photon
m22 d21, 16 Aug on average), and is linked to increasing
light availability after the polar winter.
The long-term sampling has also captured significant
variability in the bloom strength linked to changes in the
ice extent and winter mixed-layer depths. Deep winter
mixing and limited sea-ice melt in spring lead to reduced
water-column stratification in the surface 50 m. This does
not lead to a reduction in light availability because there is
still sufficient glacial meltwater to create very shallow
mixed layers. Instead, it most likely acts by reducing the
stability of the water column and therefore increasing its
propensity to vertical mixing. The increased vertical
mixing will dilute iron added at the surface in glacial
meltwater and reduces the efficiencies of phytoplankton
present by increasing the range of light levels they are
exposed to. In this way, the effects of changes in winter
meteorological conditions persist through the summer.
This is significant because it is winter changes (to ice,
winds, and temperatures) in the WAP region that are most
pronounced. The climatic drivers for these changes
(including an intensification of the SAM) are predicted
to persist and intensify for some time yet, so these changes
are likely to continue.
Acknowledgments
We thank the series of Marine Assistants who spent a year or
two on base collecting these data and the other support staff that
made this possible, especially during the austral winter. We also
thank the anonymous reviewers for their useful input into the final
version of this paper.
References
ACKLEY, S. F., AND C. W. SULLIVAN. 1994. Physical controls on the
development and characteristics of Antarctic sea ice biological
communities—a review and synthesis. Deep-Sea Res. Part I
41: 1583–1604, doi:10.1016/0967-0637(94)90062-0
ALDERKAMP, A.-C., H. J. W. DE BAAR, R. J. W. VISSER, AND K. R.
ARRIGO. 2010. Can photoinhibition control phytoplankton abundance in deeply mixed water columns of the Southern Ocean?
Limnol. Oceanogr. 55: 1248–1264, doi:10.4319/lo.2010.55.3.1248
———, AND OTHERS. 2012. Iron from melting glaciers fuels
phytoplankton blooms in the Amundsen Sea (Southern Ocean):
Phytoplankton characteristics and productivity. Deep-Sea Res.
Part II 71–76: 32–48, doi:10.1016/j.dsr2.2012.03.005
ANNETT, A. L., D. S. CARSON, X. CROSTA, A. CLARKE, AND R. J.
GANESHRAM. 2010. Seasonal progression of diatom assemblages in surface waters of Ryder Bay, Antarctica. Polar Biol.
33: 13–29, doi:10.1007/s00300-009-0681-7
ARRIGO, K. R., G. VAN DIJKEN, AND S. PABI. 2008. Impact of a
shrinking Arctic ice cover on marine primary production.
Geophys. Res. Lett. 35, doi:10.1029/2008GL035028
1045
ATKINSON, A., V. SIEGEL, E. PAKHOMOV, AND P. ROTHERY. 2004.
Long-term decline in krill stock and increase in salps within
the Southern Ocean. Nature 432: 100–103, doi:10.1038/
nature02996
BEHRENFELD, M. 2010. Abandoning Sverdrup’s critical depth
hypothesis on phytoplankton blooms. Ecology 91: 977–989,
doi:10.1890/09-1207.1
BOSS, E., AND M. BEHRENFELD. 2010. In situ evaluation of the
initiation of the North Atlantic phytoplankton bloom.
Geophys. Res. Lett. 37, doi:10.1029/2010GL044174
CLARKE, A., M. P. MEREDITH, M. I. WALLACE, M. A. BRANDON,
AND D. N. THOMAS. 2008. Seasonal and interannual variability
in temperature, chlorophyll and macronutrients in northern
Marguerite Bay, Antarctica. Deep-Sea Res. Part II 55: 1988–
2006, doi:10.1016/j.dsr2.2008.04.035
———, E. J. MURPHY, M. P. MEREDITH, J. C. KING, L. S. PECK, D.
K. A. BARNES, AND R. C. SMITH. 2007. Climate change and the
marine ecosystem of the western Antarctic Peninsula. Philos.
Trans. R. Soc. Lond. B 362: 149–166, doi:10.1098/rstb.2006.
1958
COOK, A. J., A. J. FOX, D. G. VAUGHAN, AND J. G. FERRIGNO.
2005. Retreating glacier fronts on the Antarctic Peninsula
over the past half-century. Science 308: 541–544, doi:10.1126/
science.1104235
DE BAAR, H. J. W., AND OTHERS. 2005. Synthesis of iron
fertilization experiments: From the Iron Age in the Age
of Enlightenment. J. Geophys. Res. 110, doi:10.1029/
2004JC002601
DUCKLOW, H. W., AND OTHERS. 2007. Marine pelagic ecosystems:
The West Antarctic Peninsula. Philos. Trans. R. Soc. Lond. B
362: 67–94, doi:10.1098/rstb.2006.1955
———, AND OTHERS. 2008. Particle export from the upper ocean
over the continental shelf of the west Antarctic Peninsula: A
long-term record, 1992–2007. Deep-Sea Res. Part II 55:
2118–2131, doi:10.1016/j.dsr2.2008.04.028
EVANS, G. T., AND J. S. PARSLOW. 1985. A model of annual
plankton cycles. Biol. Oceanogr. 3: 327–347.
GERRINGA, L. J. A., AND OTHERS. 2012. Iron from melting glaciers
fuels the phytoplankton blooms in Amundsen Sea (Southern
Ocean): Iron biogeochemistry. Deep-Sea Res. Part II 71–76:
16–31, doi:10.1016/j.dsr2.2012.03.007
GUEYMARD, C. A., D. MYERS, AND K. EMERY. 2003. Proposed
reference irradiance spectra for solar energy systems testing.
Solar Energy 73: 443–467, doi:10.1016/S0038-092X(03)00005-7
HUISMAN, J., P. VAN OOSTVEN, AND F. J. WEISSING. 1999. Critical
depth and critical turbulence: Two different mechanisms for
the development of phytoplankton blooms. Limnol. Oceanogr. 44: 1781–1787, doi:10.4319/lo.1999.44.7.1781
JICKELLS, T. D., AND OTHERS. 2005. Global iron connections
between desert dust, ocean biogeochemistry, and climate.
Science 308: 67–71, doi:10.1126/science.1105959
KING, J. C. 1994. Recent climate variability in the vicinity of
the Antarctic Peninsula. Int. J. Climatol. 14: 357–369,
doi:10.1002/joc.3370140402
KLINCK, J. M. 1998. Heat and salt changes on the continental shelf
west of the Antarctic Peninsula between January 1993 and
January 1994. J. Geophys. Res. 103: 7617–7636, doi:10.1029/
98JC00369
LANDRY, M. R., AND OTHERS. 2000. Biological response to iron
fertilization in the eastern equatorial Pacific (IronEx II). I.
Microplankton community abundances and biomass. Mar.
Ecol.-Prog. Ser. 201: 27–42, doi:10.3354/meps201027
LEFEBVRE, W., H. GOOSSE, R. TIMMERMAN, AND T. FICHEFET. 2004.
Influence of the Southern Annular Mode on the sea ice–ocean
system. J. Geophys. Res. 109, doi:10.1029/2004JC002403
1046
Venables et al.
LONG, M. C., L. N. THOMAS, AND R. B. DUNBAR. 2012. Control of
phytoplankton bloom inception in the Ross Sea, Antarctica,
by Ekman restratification. Glob. Biogeochem. Cycles 26,
doi:10.1029/2010GB003982
MARTINSON, D. G., S. E. STAMMERJOHN, R. A. IANNUZZI, R. C.
SMITH, AND M. VERNET. 2008. Western Antarctic Peninsula
physical oceanography and spatio-temporal variability. DeepSea Res. Part II 55: 1964–1987, doi:10.1016/j.dsr2.2008.04.038
MASSOM, R. A., AND OTHERS. 2006. Extreme anomalous atmospheric circulation in the West Antarctic Peninsula region in
austral spring and summer 2001/02, and its profound impact
on sea ice and biota. J. Clim. 19: 3544–3571, doi:10.1175/
JCLI3805.1
MEREDITH, M. P., AND OTHERS. 2008. Variability in the freshwater
balance of northern Marguerite Bay, Antarctic Peninsula:
Results from d18O. Deep-Sea Res. Part II 55: 309–322,
doi:10.1016/j.dsr2.2007.11.005
———, I. A. RENFREW, A. CLARKE, J. C. KING, AND M. A.
BRANDON. 2004. Impact of the 1997/98 ENSO on upper ocean
characteristics in Marguerite Bay, western Antarctic Peninsula. J. Geophys. Res. 109, doi:10.1029/2003JC001784
———, AND OTHERS. 2010. Changes in the freshwater composition
of the upper ocean west of the Antarctic Peninsula during the
first decade of the 21st century. Prog. Oceanogr. 87: 127–143,
doi:10.1016/j.pocean.2010.09.019
MOLINE, M. A., H. CLAUSTRE, T. K. FRAZER, O. SCHOFIELD, AND
M. VERNET. 2004. Alteration of the food web along the
Antarctic Peninsula in response to a regional warming trend.
Glob. Change Biol. 10: 1973–1980, doi:10.1111/j.13652486.2004.00825.x
MONGIN, M., E. MOLINA, AND T. W. TRULL. 2008. Seasonality and
scale of the Kerguelen plateau phytoplankton bloom: A
remote sensing and modeling analysis of the influence of
natural iron fertilization in the Southern Ocean. Deep-Sea
Res. Part II 55: 880–892, doi:10.1016/j.dsr2.2007.12.039
MONTES-HUGO, M., S. C. DONEY, H. W. DUCKLOW, W. FRASER, D.
G. MARTINSON, S. E. STAMMERJOHN, AND O. SCHOFIELD. 2009.
Recent changes in phytoplankton communities associated
with rapid regional climate change along the western
Antarctic Peninsula. Science 323: 1470–1473, doi:10.1126/
science.1164533
MOREL, F. M. M., J. G. RUETER, AND N. M. PRICE. 1991. Iron
nutrition of phytoplankton and its possible importance in the
ecology of ocean regions with high nutrient and low biomass.
Oceanography 4: 56–61, doi:10.5670/oceanog.1991.03
PARK, J., I.-S. OH, H.-C. KIM, AND S. YOO. 2010. Variability of
SeaWiFs chlorophyll a in the southwest Atlantic sector of the
Southern Ocean: Strong topographic effects and weak
seasonality. Deep-Sea Res. Part I 57: 604–620, doi:10.1016/
j.dsr.2010.01.004
PECK, L. S., D. K. A. BARNES, A. J. COOK, A. H. FLEMING, AND
A. CLARKE. 2010. Negative feedback in the cold: Ice retreat
produces new carbon sinks in Antarctica. Global Change
Biology 16: 2614–2623, doi:10.1111/j.1365-2486.2009.02071.x
PLANQUETTE, H., AND OTHERS. 2007. Dissolved iron in the vicinity
of the Crozet Islands, Southern Ocean. Deep-Sea Res. Part II
54: 1999–2019, doi:10.1016/j.dsr2.2007.06.019
PLATT, T., D. S. BROOMHEAD, S. SATHYENDRANATH, A. M.
EDWARDS, AND E. MURPHY. 2003. Phytoplankton biomass
and residual nitrate in the pelagic ecosystem. Proc. R. Soc.
Lond. A 459: 1063–1073, doi:10.1098/rspa.2002.1079
SCHLOSS, I. R., D. ABELE, S. MOREAU, S. DEMERS, A. V. BERS,
O. GONZALEZ, AND G. A. FERREYRA. 2012. Response of
phytoplankton dynamics to 19-year (1991–2009) climate
trends in Potter Cove (Antarctica). J. Mar. Syst. 92: 53–66,
doi:10.1016/j.jmarsys.2011.10.006
SCHOFIELD, O., H. W. DUCKLOW, D. G. MARTINSON, M. P.
MEREDITH, M. A. MOLINE, AND W. FRASER. 2010. How do
polar marine ecosystems respond to rapid climate change?
Science 328: 1520–1523, doi:10.1126/science.1185779
SIEGEL, D. A., S. C. DONEY, AND J. A. YODER. 2002. The North
Atlantic spring phytoplankton bloom and Sverdrup’s critical
depth hypothesis. Science 296: 730–733, doi:10.1126/science.1069174
SIMPSON, J. H., C. M. ALLEN, AND N. C. G. MORRIS. 1978. Fronts
on the continental shelf. J. Geophys. Res. 83: 4607–4614,
doi:10.1029/JC083iC09p04607
STAMMERJOHN, S. E., D. G. MARTINSON, R. C. SMITH, AND R. A.
IANNUZZI. 2008. Sea ice in the western Antarctic Peninsula
region: Spatio-temporal variability from ecological and
climate change perspectives. Deep-Sea Res. Part II 55:
2041–2058, doi:10.1016/j.dsr2.2008.04.026
SVERDRUP, H. U. 1953. On conditions for the vernal blooming of
phytoplankton. J. du Conseil Perm. Int. l’Exploration Mer 18:
287–295.
TAYLOR, J. R., AND R. FERRARI. 2011a. The role of density fronts
in the onset of phytoplankton blooms. Geophys. Res. Lett.
38: L23601, doi:10.1029/2011GL049312
———, AND ———. 2011b. Shutdown of turbulent convection as
a new criterion for the onset of spring phytoplankton blooms.
Limnol. Oceanogr. 56: 2293–2307, doi:10.4319/lo.2011.56.6.
2293
THOMAS, E. R., G. J. MARSHALL, AND J. R. MCCONNELL. 2008. A
doubling in snow accumulation in the western Antarctic
Peninsula since 1850 Geophys. Res. Lett. 35: L01706, doi:
10.1029/2007GL032529
THOMPSON, D. W. J., AND S. SOLOMON. 2002. Interpretation of
recent Southern Hemisphere climate change. Science 296:
895–899, doi:10.1126/science.1069270
TOWNSEND, D. W., M. D. KELLER, M. E. SIERACKI, AND S. G.
ACKLESON. 1992. Spring phytoplankton blooms in the absence
of vertical water column stratification. Nature 360: 59–62,
doi:10.1038/360059a0
TREMBLAY, J.-E., K. SIMPSON, J. MARTIN, L. MILLER, Y. GRATTON,
D. BARBER, AND N. M. PRICE. 2008. Vertical stability and the
annual dynamics of nutrients and chlorophyll fluorescence
in the coastal, southeast Beaufort Sea. J. Geophys. Res. 113,
doi:10.1029/2007JC004547
TURNER, J., AND OTHERS. 2005. Antarctic climate change during
the last 50 years. Int. J. Climatol. 25: 279–294, doi:10.1002/
joc.1130
VAUGHAN, D. G., AND OTHERS. 2003. Recent rapid regional climate
warming on the Antarctic Peninsula. Climatic Change 60:
243–274, doi:10.1023/A:1026021217991
VENABLES, H. J., AND C. M. MOORE. 2010. Phytoplankton and
light in the Southern Ocean: Learning from high nutrient high
chlorophyll areas. J. Geophys. Res. 115: C02015, doi:10.1029/
2009JC005361
———, R. T. POLLARD, AND E. E. POPOVA. 2007. Physical
conditions controlling the early development of a regular
phytoplankton bloom north of the Crozet Plateau, Southern
Ocean. Deep-Sea Res. Part II 54: 1949–1965, doi:10.1016/
j.dsr2.2007.06.014
VERNET, M., AND OTHERS. 2008. Primary production within the
sea-ice zone west of the Antarctic Peninsula: I—Sea ice,
summer mixed layer, and irradiance. Deep-Sea Res. Part II
55: 2068–2085, doi:10.1016/j.dsr2.2008.05.021
WALLACE, D. W. R., M. P. MEREDITH, M. A. BRANDON, T. J.
SHERWIN, A. DALE, AND A. CLARKE. 2008. On the characteristics of internal tides and coastal upwelling behaviour in
Marguerite Bay, west Antarctic Peninsula. Deep-Sea Res.
Part II 55: 2023–2040, doi:10.1016/j.dsr2.2008.04.033
Wintertime controls on spring bloom
WALLACE, M. I. 2007. Ocean circulation in Marguerite Bay. Ph.D.
thesis. The Open Univ., UK.
WIEDENMANN, J., K. A. CRESSWELL, AND M. MANGEL. 2009.
Connecting recruitment of Antarctic krill and sea ice. Limnol.
Oceanogr. 54: 799–811, doi:10.4319/lo.2009.54.3.0799
1047
Associate editor: Heidi M. Sosik
Received: 08 March 2012
Accepted: 17 December 2012
Amended: 12 December 2012