Timing of blooms, algal food quality and Calanus
glacialis reproduction and growth in a changing Arctic
JANNE E. SØREIDE
*
, EVA LEUw , JØRGEN BERGE
*
, MARTIN GRAEVEz and
STIG FALK-PETERSENw
*
The University Centre in Svalbard, PO Box 156, N-9171 Longyearbyen, Norway, wNorwegian Polar Institute, N-9296 Troms,
Norway, zAlfred-Wegner-Institute fu
¨
r Polar-und Meeresforschung, Am Handelshafen 12, 27570 Bremerhaven, Germany
Abstract
The Arctic bloom consists of two distinct categories of primary producers, ice algae growing within and on the
underside of the sea ice, and phytoplankton growing in open waters. Long chain omega-3 fatty acids, a subgroup of
polyunsaturated fatty acids (PUFAs) produced exclusively by these algae, are essential to all marine organisms for
successful reproduction, growth, and development. During an extensive field study in the Arctic shelf seas, we
followed the seasonal biomass development of ice algae and phytoplankton and their food quality in terms of their
relative PUFA content. The first PUFA-peak occurred in late April during solid ice cover at the onset of the ice algal
bloom, and the second PUFA-peak occurred in early July just after the ice break-up at the onset of the phytoplankton
bloom. The reproduction and growth of the key Arctic grazer Calanus glacialis perfectly coincided with these two
bloom events. Females of C. glacialis utilized the high-quality ice algal bloom to fuel early maturation and
reproduction, whereas the resulting offspring had access to ample high-quality food during the phytoplankton
bloom 2 months later. Reduction in sea ice thickness and coverage area will alter the current primary production
regime due to earlier ice break-up and onset of the phytoplankton bloom. A potential mismatch between the two
primary production peaks of high-quality food and the reproductive cycle of key Arctic grazers may have negative
consequences for the entire lipid-driven Arctic marine ecosystem.
Keywords: Calanus glacialis, climate change, food quality, ice algae, lipids, mismatch-hypothesis, phytoplankton, PUFAs
Received 19 August 2009 and accepted 20 November 2009
Introduction
Although the dramatic loss of Arctic sea ice during the
last decade is indisputable (Smetacek & Nicol, 2005;
Stroeve et al., 2007; Comiso et al., 2008), the conse-
quences of this loss on key biological processes remain
largely unknown. Of the studies addressing potential
impacts of climate change on polar marine ecosystems,
few have focused on the biochemical aspects of trophic
interactions (e.g., food quality and transfer) (but see
(Falk-Petersen et al., 2007; Kaartvedt, 2008).
Sea ice plays a dual role for primary production in
polar seas (Smetacek & Nicol, 2005), both providing a
habitat for ice algae and regulating the available light
for primary production. Ice algae begin growing in low
light levels in March and continue growing until their
sea ice substratum melts (Hegseth, 1998). In contrast,
phytoplankton production starts after the ice break-up,
giving a temporal discontinuity between sea ice
and open-water production. As the window of oppor-
tunity for primary production becomes narrower at
higher latitudes, the timing and availability of essential
omega-3 fatty acids become increasingly crucial for all
marine organisms. The long-chain eicosapentaenoic
acid (EPA) and docosahexaenoic acid (DHA) are
omega-3 fatty acids produced exclusively by marine
algae. These polyunsaturated fatty acids (PUFAs) play a
key role in reproduction, growth, and physiology for all
organisms in marine ecosystems (Ackman, 1989), as
well as for human health (Riediger et al., 2009). The
importance of omega-3 fatty acids for copepod egg
production, egg hatching, and zooplankton growth
has been well documented in field (Pond et al., 1996;
Swadling et al., 2000; Jonasdottir et al., 2005) and experi-
mental studies (Breteler et al., 2005; Jonasdottir et al.,
2009), and has furthermore been proven to be essential
for proper fish development (Watanabe et al., 1983).
Among the zooplankton in the arctic shelf seas, the
arctic grazer Calanus glacialis accounts for up to 80% of
the biomass (Tremblay et al., 2006; Blachowiak-Samolyk
et al., 2008; Sreide et al., 2008) and plays a key role in
the pelagic lipid-based arctic food web (Falk-Petersen
et al., 1990). C. glacialis accumulates essential PUFAs
from its algal diet, and converts the low-energy carbo-
hydrates and proteins in algae into high-energy wax
ester lipids (Lee et al., 2006; Falk-Petersen et al., 2009).
These lipids make it an extremely energy-rich food
Correspondence: Janne E. Sreide, tel. 1 47 79023300,
fax 1 47 79023301, e-mail: Janne.Soreide@unis.no
Global Change Biology (2010), doi: 10.1111/j.1365-2486.2010.02175.x
r 2010 Blackwell Publishing Ltd 1
(470% lipids of dry weight) for higher trophic levels
(Falk-Petersen et al., 1990).
C. glacialis has a 1–3 year life cycle, depending on the
temperature and food regime. The life-cycle includes six
nauplii and six copepodite stages that follow a pro-
nounced seasonal migration pattern. C. glacialis devel-
ops through the various stages mainly during summer.
In autumn, it accumulates lipids before it descends
towards the deep and enters a diapause to survive the
long and dark food-poor winter. The main overwinter-
ing stages are copepodite stage IV (CIV) and V (CV)
(Falk-Petersen et al., 2009). Overwintering CV indivi-
duals develop into females in mid-winter and ascend to
surface waters in spring to feed and reproduce (Koso-
bokova, 1999). The evolutionary success of C. glacialis
depends on its ability to synchronize its seasonal mi-
gration, reproduction, and growth to the primary pro-
duction regime in Arctic shelf seas (Falk-Petersen et al.,
2009). As sea ice becomes thinner and has less coverage,
the underwater light climate will change significantly
(Tremblay et al., 2006; Pabi et al., 2008). This change will
alter the onset, duration, and magnitude of the sea ice
algal and phytoplankton blooms. Because the peak
growing season for ice algae is confined to consolidated
ice, the qualitative and quantitative importance of sea
ice algae for the development of key Arctic grazers
remain poorly studied.
To predict ecological consequences of climate change
on the algal blooms and PUFA production, we carried
out an extensive field study during the International
Polar Year (IPY, 2007) in the seasonally ice-covered
Rijpfjord in northern (4801N) Svalbard. We followed
the seasonal development of ice algae and phytoplank-
ton, including biomass variation and food quality (i.e.,
the proportion of PUFAs) simultaneously with popula-
tion development of the key grazer C. glacialis. This
study aimed at unravelling the intimate coupling be-
tween the solar cycle, food quality peaks, and the onset
and duration of primary and secondary production.
Materials and methods
Study area
The study was performed in 2007 in Rijpfjorden, Svalbard
(Appendix S1) as part of the Norwegian IPY-project CLEOPA-
TRA (Climate effects on planktonic food quality and trophic
transfer in Arctic marginal ice zone). Rijpfjorden is a north-
facing, relatively shallow fjord (max. 240 m deep) that opens
towards the Arctic Ocean. It has a wide opening that is in
direct contact with a broad shallow shelf (100–200 m deep),
which extends to the shelf-break of the Polar Basin at 811 N
(Appendix S1). Rijpfjorden is dominated by cold Arctic water
masses and is covered by sea ice up to 9 months a year
(Ambrose et al., 2006; Wallace et al., 2010). The zooplankton
community in Rijpfjorden is dominated by Arctic species, with
C. glacialis representing up to 90% of the zooplankton biomass
(Daase & Eiane, 2007; Blachowiak-Samolyk et al., 2008).
Physical data
Temperature, salinity, and in situ fluorescence were measured at
hourly intervals by a mooring placed in close vicinity of our
main sampling station (Stn. SH; Appendix S2). The mooring was
equipped with SBE microcats (model: 37-SM MicroCAT; Sea-
Bird Electronics, Bellevue, WA, USA) at 27 and 205 m that
recorded temperature, conductivity, and pressure. In addition,
Vemco temperature mini loggers (Minilog12; accuracy 0.11C)
were located at 3–30 m intervals between 20 and 200 m. At 17 m
deep, a Seapoint Chlorophyll Fluorometer (Seapoint Sensors
Inc., Exeter, NH, USA) was attached to the mooring. Only
approximate chlorophyll a data were available from this fluo-
rometer due to absence of suitable water samples for exact
calibration, but for identifying the approximate timing of the
phytoplankton bloom is was useful. Samples for exact chloro-
phyll a concentrations were collected at lower resolution,
monthly from April to August, and in October. The mooring
was deployed in September 2006 and recovered in August 2007.
For more detailed information about the mooring, see Berge et al.
(2009) and Wallace et al.(2010).
We measured the photosynthetically active radiation (PAR;
400–700 nm) hourly with a PAR LITE Kipp & Zonen Quantum
Sensor (Campell Scientific, Edmonton, AB, Canada), mounted
4 m above sea level at the Rijpfjorden field station. Sea ice
thickness and snow depth were determined for all sites from
which ice cores for ice algae analyses were taken.
Primary producers
Samples of primary producers were collected monthly from
March to August and in October (Appendix S2). We took core
samples of ice algae using a SIPRE type corer (12.5 cm dia-
meter). For each sampling spot, we took three core replicates
(50–100 cm apart). We sawed off the lowest part (5–8 cm) of the
core that contained visible amounts of ice algae. This core
section was protected against light exposure and immediately
transported to the field station. In the field laboratory, we
slowly (24–36 h) thawed the cores in the dark in 500 mL of GF/
F filtered seawater. After the samples were completely thawed,
subsamples were filtered through precombusted GF/F-filters
(1 h at 450 1C) to estimate the concentration of chlorophyll a,
the total particulate carbon (C), and the fatty acid composition
of the total lipids. Pelagic algae (phytoplankton) were sampled
with a 10 L Niskin-bottle (Ocean Test Equipment Inc., Fort
Lauderdale, FL, USA) at six depths between 0 and 50 (80) m
(Appendix S2). At each depth, samples were measured in
triplicate (0.5–3 L depending on the algae concentrations)
and filtered on precombusted GF/F-filters (1 h at 450 1C) to
estimate the concentration of chlorophyll a, the total particulate
C, and the fatty acid composition of the total lipids.
Chlorophyll a concentration was determined by high-perfor-
mance liquid chromatography (HPLC). The pigments were
2 J. E. SØREIDE et al.
r 2010 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2010.02175.x
extracted from the filters with 1.6 mL methanol. The extract was
sonicated for 30 s using a Vibra-cell sonicator (Sonics and
Materials Inc., Danbury, CT, USA) equipped with a 3 mm
diameter probe. The extraction and HPLC analysis continued
accordingtoWright&Jeffrey(1997),usinganabsorbancediode-
array detector (Spectraphysics UV600LP, Newport Corp., Ther-
mo Fisher Scientific, Waltham, MA, USA). The column was a
C18 Phenomenex Ultracarb (Torrance, CA, USA) 3 mmODS(20)
(150 3.20mm). The HPLC system was calibrated with pigment
standards from DHI, Water and Environment, Denmark.
For particulate C analyses, samples of 20–100 mL for ice algae,
and 250–2000 mL for phytoplankton, depending on biomass
density, were filtered. All filter samples were frozen (20 1C)
until analysis. Particulate C was analyzed on a Thermo Finnigan
FlashEA 1112 elemental analyser (Waltham, MA, USA).
Secondary producers
Zooplankton was sampled monthly by WP2 closing nets with a
0.225 m
2
opening, vertically at four standard depths: 0–20, 20–50,
50–100, and 100 m bottom (Appendix S2). To gather data on
Calanus copepodites and nauplii, we used WP2 nets with mesh
size 200 mm. To estimate Calanus egg abundance, we used
modified WP2 nets with mesh size 63 mm. In August and
October, Calanus copepodites and nauplii were sampled with a
multiple plankton sampler (MPS; Hydro-Bios, Kiel, Germany)
consisting of five closing nets with the same opening diameter
and mesh size (200 mm) as the WP2 closing net. Calanus speci-
mens were identified to the species level based on morphology
and prosome lengths of individual copepodite stages (Kwas-
niewski et al., 2003). Eggs and nauplii were identified to Calanus
genus level. Although sampling depth ranged from 130 m in July
and October to 186 m in September, Calanus density (ind. m
2
)
was calculated for 0–140 m in all months (Appendix S2).
To estimate egg production rates, we incubated females for
24 h at near to in situ temperatures, obtained by placing the
incubator chambers in a large cooling box filled with sea water
and sea ice. Each female was placed alone in a 200 mL chamber
with a false bottom of 500 mm mesh. Incubations started within
2–3 h of sampling in prescreened (60 mm mesh size) surface sea
water collected from the same site as the females.
Fatty acid analysis
Fatty acids of particulate organic matter (POM) were analyzed
at Unilab (Troms, Norway), whereas fatty acid and fatty
alcohol of C. glacialis were analyzed at Alfred-Wegner-Institute
(Bremerhaven, Germany).
For POM, triplicate samples of 100–200 mL from each ice
core and 0.5 to 3 L from each water depth (Niskin samples)
were filtered onto precombusted glass fibre filters (GF/F). The
filters were transferred to glass vials with Teflon-lined caps
and 8 mL dichloromethane–methanol (2 : 1, v/v) was added.
The vials were stored at 80 1C until analyzed. Total lipid was
extracted according to the procedure described in Folch et al.
(1957). A known amount of heneicosanoic acid (21 : 0) was
added as internal standard, and an acid-catalysed transester-
ification was carried out with 1% sulfuric acid in methanol
(Christie, 1982). The extract was then cleaned using a silica
column (Christie, 1982). The relative composition of the fatty
acid methyl esters (FAME) was determined in an Agilent 6890
N (Agilent Technologies Deutschland GmbH & Co. KG,
Waldbronn, Germany) gas chromatograph, equipped with a
fused silica, wall-coated capillary column (50 m 0.25 mm i.d.,
Varian Select FAME, Agilent Technologies Deutschland GmbH
& Co. KG) with an oven thermal gradient from an initial 60 to
150 1Cat301Cmin
1
, and then to a final temperature of 230 1C
at 1.5 1Cmin
1
. Individual components were identified by
comparison with two known standards and were quantified
using HPChemStation software (Hewlett-Packard, Agilent
Technologies Deutschland GmbH & Co. KG).
For Calanus, 10–30 individuals were pooled and transferred to
glass vials with Teflon-lined caps and 8 mL dichloromethane–
methanol (2 : 1, v/v) was added. These vials were stored at
80 1C until analyzed. Specimens from two discrete layers, the
surface (0–50 m) and bottom layer (4100 m), were analyzed
separately. Calanus specimens were homogenized and lipids
were extracted according to Folch et al. (1957). Methyl esters of
fatty acids and free fatty alcohols were prepared by transester-
ification of the lipid extract with 3% concentrated sulfuric acid in
methanol for 4 h at 80 1C under nitrogen atmosphere. FAME and
free alcohols were then simultaneously analyzed with a gas
liquid chromatograph (HP 6890N, Agilent Technologies Deutsch-
land GmbH & Co. KG) on a 30 m 0.25 mm i.d. wall-coated
open tubular column (film thickness: 0.25 mm; liquid phase: DB-
FFAP), equipped with split/splitless injector (250 1C) and flame
ionization detector (280 1C) using temperature programming as
described above. Fatty acids and fatty alcohols were quantified
with an internal 19 : 0 fatty acid standard added to the sample
before the extraction. Individual components were identified by
comparisons to standards or, if necessary, by additional GC-mass
spectrometry runs. The samples were quantified using ChemSta-
tion software (Agilent, Agilent Technologies Deutschland GmbH
& Co. KG). Total lipid composition was calculated as sum of total
fatty acids and fatty alcohols.
All identified PUFAs and omega-3 fatty acids (Appendix S4)
were included when calculating the proportions of PUFAs and
omega-3 fatty acids in algae and C. glacialis.
Statistical analyses
Statistical tests were performed using STATISTICA 7.0 (StatSoft Inc.,
Tulsa, OK, USA): t-tests were used when comparing two inde-
pendent groups, and one-way
ANOVA followed by the post hoc
tests Tukey’s honestly significantly different (HSD) and unequal
Tukey’s HSD were used when comparing multiple groups with
similar or unequal number of replicates per group, respectively
(Winer et al., 1991). If the variance between independent groups
was unequal (i.e., Levene’s T-test P 0.05), we used the Mann–
Whitney U-test (MWU-test) and Kruskal–Wallis multiple com-
parisons of mean ranks for all groups (Siegel & Castellan, 1988).
The significance level was set to P 0.05 in all tests.
Results
Physical properties: hydrography, sea ice, and light
Rijpfjorden froze solid February 2, 2007 (J. Berge, per-
sonal observations). During the ice covered period from
TIMING OF PRODUCTION IN A CHANGING ARCTIC 3
r 2010 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2010.02175.x
February to end of June there was a cold (1.8 1C)
homogenous water mass from surface to bottom
(Fig. 1). The sea ice thickness was on average 0.5 m in
March, and around 1 m thick from April to June (mean
0.9 0.1 m). By the end of June, the sea ice started to
break up, and on 12 July, the fjord was ice free (J. E.
Sreide, personal observations). In Rijpfjorden (80.271N
and 22.291E), the sun appeared for the first time in late
February (22 February) (http://www.esrl.noaa.gov/
gmd/grad/solcalc/). The mean daily light intensities
increased rapidly the following months (Appendix S3),
and the midnight sun appeared from 11 April to 31
August. The 4 months of the long polar night period
began when the sun disappeared on 21 October.
Primary producers: ice algae and phytoplankton
Between March and October there were two distinct
algal blooms, corresponding to the two peaks in PUFA-
production. The earlier PUFA-peak was associated with
the ice algal bloom in late April, and the later PUFA-
peak corresponded to the phytoplankton bloom just
after ice break-up in early July (Figs 1 and 2). Omega-
3 fatty acids accounted for most of the PUFAs in both ice
algae (65%–74%) and phytoplankton (57%–83%) (Table
1, Appendix S4). Ice algae were present as early as
March, but biomass began to build up in April and
lasted until June. Similarly high biomass in terms of
particulate carbon was found in April and June, but the
Fig. 1 The temperature profile measured from September 2006 to September 2007 in Rijpfjorden by a mooring equipped with
temperature loggers spaced through the water column. Timing of sea ice and ice algae are indicated by drawings at the plot, whereas
phytoplankton are chlorophyll a (Chl a) measurements from a fluorometer placed at the mooring at 17 m depth. Peak biomass of ice algae
occurred from mid-April to approx. mid-June. The phytoplankton chlorophyll a values are only approximate values due to lack of
suitable water samples for proper calibration.
Fig. 2 The relative polyunsaturated fatty acids (PUFA) content (as percentage of total fatty acids; mean SD) in algae and females,
copepodite stage V (CV) and stage IV (CIV) of Calanus glacialis from March to October 2007 in Rijpfjorden. Only ice algae were present
from April to June, whereas from July to October only phytoplankton was available for grazers. Hatched lines were drawn when data
between monthly points were missing. For algae average values per month are shown, based upon three to five independent station
measurements for ice algae (with three replicate ice cores each), and two to four stations with six sampling depths for phytoplankton. For
C. glacialis average values based upon three to nine samples of 10–30 individuals per sample are shown per month.
4 J. E. SØREIDE et al.
r 2010 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2010.02175.x
ice algal PUFA content decreased from 37% in April to
22% in June (t-test, P 5 0.001) (Table 1, Fig. 2). The
phytoplankton biomass was negligible during the ice-
covered period from February to June (Fig. 1). The ice
break up in June/July was followed by a phytoplankton
bloom in early July (Table 1, Fig. 1). The quality of the
phytoplankton as food also peaked during this bloom
when the phytoplankton contained up to 40% PUFAs
(Table 1, Fig. 2). The pelagic POM (P-POM) had a low
PUFA content in April (o15%). In June, P-POM con-
sisted primarily of ice algae sloughed off the bottom of
the sea ice, so in June, ice algae and P-POM had a
similar PUFA content (t-test, P 5 0.667).
Secondary producers: C. glacialis
During April and May, females of C. glacialis had a
pronounced increase in their relative PUFA content
(Fig. 2; Appendix S5a). In contrast, CIV and CV indivi-
duals in these months had no significant increases in
relative PUFA content (Fig. 2, Appendix S5b and c). C.
glacialis females produced eggs during the ice algal
bloom (Table 2), which mirrored an increase in egg
abundance in the net samples from April to June
(Fig. 3). We could not estimate egg production after June
due to very low female abundance in this period
(o1 ind. m
3
in the upper 50 m). Similarly, we found
very low egg abundance from July to October (Fig. 3).
The total lipid content in surface dwelling females
dropped in April at the onset of spawning, but re-
mained stable during the spawning period from April
to June (Kruskal–Wallis median test, P 5 0.0563) (Fig. 5).
In contrast, the total PUFA content in females slightly
increased from March to June (Fig. 5).
The peak abundance of C. glacialis nauplii and young
copepodites coincided with the pelagic bloom in July,
which provided the offspring with excellent food (Figs 3
and 4). Young nauplii stages dominated during the ice-
covered period, whereas older feeding nauplii stages
( NIII) dominated at the onset of the phytoplankton
bloom (Fig. 4b). In July, the youngest copepodite stages
(CI–CIII) accounted for most of the population (70%),
whereas the overwintering stages CIV and CV domi-
nated from August (Fig. 4a). Females (40%–69%) and
CIV individuals (17%–46%) were the most common C.
glacialis stages from March to June, followed by CIV
(17%–46%) (Fig. 4a). C. glacialis males were absent from
March to September, but started to appear below 100 m
depth (0.5 ind. m
3
) in October.
By early March, 23% of the female population had
already migrated to the upper 50 m (Fig. 5). In contrast,
during the same time, only 1% of C. glacialis CIV
Table 1 Integrated (0–50 m) total carbon (C) and chlorophyll a (Chl a) biomass, and relative amount of polyunsaturated fatty acids
(PUFAs) of total lipids and the relative amount of omega-3 fatty acids of total PUFAs (mean SD; nd, not determined) in ice- and
pelagic-particulate organic matter (POM) in Rijpfjorden 2007
Total C (g m
2
) Chl a (mg m
2
) PUFAs/total lipids (%) Omega-3/PUFAs (%)
Ice-POM
April 0.2 0.1 22.4 15.3 36.7 1.4 65.0 1.5
June 0.2 0.1 9.2 8.8 22.3 2.8 73.7 2.6
Pelagic-POM
April nd 3.5 0.6 14.3 56.5
June 4.4 0.2
*
15.9 2.2
*
20.2 2.8
*
75.9 1.1
*
July 21.6 4.8 77.5 30.5 40.3 3.1 83.4 4.1
August 12.7 3.2 12.7 3.2w 32.5 4.5 82.4 2.6
October 9.7 3.1 23.1 9.1 23.9 5.4 80.4 3.1
*
Mainly ice algae sloughed off from the bottom ice.
wApproximate values estimated from fluorometer readings.
Table 2 Calanus glacialis egg production measurements from Rijpfjorden in 2007
3–4 March 25–26 April 1–2 May 5–6 June 7–8 June
Number of females incubated 30 29 29 33 30
Incubation temperature ( 1C) 1.7 1.6 1.5 0 0.5
Prosme length in mm (mean SD) 3.3 0.3 3.4 0.3 3.5 0.3 3.5 0.2 3.7 0.4
% Egg laying females 0 27.6 13.8 27.3 63.3
Egg production day
1
(mean SE)
*
0 7.4 2.8 5.7 4.7 14.8 5.2 17.4 3.6
Max egg clutch size day
1
016 154659
*
Per egg laying female.
TIMING OF PRODUCTION IN A CHANGING ARCTIC 5
r 2010 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2010.02175.x
![Fig. 7 Current primary production regime in Arctic shelf seas (a) with highest food quality [highest poly unsaturated fatty acid (PUFA) content] during the ice algal and phytoplankton blooms. Calanus glacialis efficiently uses the high-quality ice algal food in early spring to fuel reproduction, which allows the offspring (nauplii and copepodites) to fully exploit the high food quality in the later occurring phytoplankton bloom. This perfect primary producer–grazer match ensures high population biomass of C. glacialis. Future primary production regime (b) with shorter growth season for ice algae due to earlier ice break up, will lead to shorter time between the two PUFA-peaks associated with the ice algal and phytoplankton blooms. This decrease may lead to a mismatch between primary producers and the ontogenetic development of the offspring. Because C. glacialis requires roughly 3 weeks to develop to first feeding nauplii stage (NIII) after spawning, it may partially or totally miss the high-quality phytoplankton bloom during its most critical growth phase.](/figures/fig-7-current-primary-production-regime-in-arctic-shelf-seas-182r10j4.webp)
