ARCTIC PRODUCTIVITY
Changes in phytoplankton concentration now drive
increasedArctic Ocean primary production
K. M. Lewis, G. L. van Dijken, K. R. Arrigo*
Historically, sea ice loss in the Arctic Ocean has promoted increased phytoplankton primary production
because of the greater open water area and a longer growing season. However, debate remains about
whether primary production will continue to rise should sea ice decline further. Using an ocean color
algorithm parameterized for the Arctic Ocean, we show that primary production increased by 57%
between 1998 and 2018. Surprisingly, whereas increases were due to widespread sea ice loss during
the first decade, the subsequent rise in primary production was driven primarily by increased
phytoplankton biomass, which was likely sustained by an influx of new nutrients. This suggests a future
Arctic Ocean that can support higher trophic-level production and additional carbon export.
I
n response to anthropogenic climate change,
the Arctic is warming faster than any other
region, with the majority of the warming
centered over the Arctic Ocean (AO) ( 1 ).
Sea ice has radically decreased in concen-
tration, volume, and duration, with summer
sea ice predicted to disappear completely by
mid-century ( 1 ). Correspondingly, annual
phytoplankton net primary production (NPP)
has significantly increased owing to a longer
growing season and an expanded area of open
water (OW) ( 2 – 5 ). However, scientists debate
how continued sea ice declines will affect AO
NPP in the future ( 6 , 7 ). Greater freshwater
flux through precipitation, ice melt, and river
outflow could intensify surface ocean stratifi-
cation and inhibit the mixing of deep nutrients
into surface waters, thus reducing AO NPP
( 8 , 9 ). Alternatively, greater OW area and more
frequent storms ( 5 )mayincreaseNPPbypro-
moting the upward delivery of new nutrients
to the depleted euphotic zone through enhanced
wind mixing ( 10 ), internal waves ( 11 ), and shelf
break upwelling ( 12 , 13 ). Here, we present a
two-decade-long time series of NPP in the AO
that we parameterized using the largest and
most complete dataset of in situ optics and
phytoplankton biomass and physiology ever
assembled for these waters to assess the cur-
rent trajectory of NPP in response to ongoing
changes in Arctic climate.
Satellite-derived estimates of chlorophyll
a(Chla), sea surface temperature (SST), and sea
ice concentration were used as input to an
AO NPP algorithm ( 2 , 3 , 14 ) to evaluate trends
from 1998 to 2018. We used a modified version of
the standard empirical NASA–Chlaalgorithm
to better account for the distinct bio-optical
properties of the AO, which differ notably from
the global ocean because of higher pigmentpackaging and chromophoric dissolved organic
matter (CDOM) concentrations ( 15 , 16 ). The
updated Chlaalgorithm ( 17 ) was developed
by using 501 coincident measurements of in
situ remote sensing reflectance and Chlafrom
25 different cruises that captured the spatial
heterogeneity across the AO. Time series trends
for mean surface phytoplankton biomass con-
centration (Chla, milligrams per cubic meter),
spatially integrated NPP (teragrams of car-
bon per year), SST (degrees Celsius), OW area
(square kilometers), and OW duration (days)
were statistically evaluated for the entire AO
and 10 subregions (Fig. 1A) for the 21-year time
period.
OW area (<50% sea ice cover) has increased
by27%intheAObetween1998and2018,with
~59,000 km^2 of OW added each year (Table 1).
Subregions that experienced significant in-
creases in OW area (24 to 123%) included the
Basin, Kara, Siberian, Barents, and Chukchi
(Table 1). Increases in OW area in the Laptev
and Beaufort subregions were nonsignificant,
and changes in the outflow shelves of the
Nordic, Canada, and Baffin subregions were
negligible (Table 1). However, the rate of OW
increase in the AO and all subregions, except
the Nordic, has slowed considerably since
2009 (Fig. 2A and Table 1).
At the same time, AO Chlaconcentration
increased significantly (22%) between 1998 and
2018 (Table 1), with almost all of the increase
occurring since 2009 (Fig. 2B and Table 1).
These changes were largely restricted to the198 10 JULY 2020•VOL 369 ISSUE 6500 sciencemag.org SCIENCE
Department of Earth System Science, Stanford University,
Stanford,CA 94305, USA.
*Corresponding author. Email: [email protected]Fig. 1. Regions of interest and changes in phytoplankton biomass.(A)The AO with its shelf seas and basin. Subregions are bounded by black lines by using the
1000-m isobath and categorized as inflow (green), interior (orange), or outflow (purple) shelves. The 200-m isobath is shown in gray. Inflow and outflow currents are
depicted as green and purple arrows, respectively. (B) The rate of change in Chla(milligrams per cubic meter per year) between 1998 and 2018. Subregions are
delineated by gray lines. Black pixels indicate no data.
RESEARCH | REPORTS