Cascading effects augment the direct impact of CO2 on phytoplankton growth in a biogeochemical model

Abstract

Atmospheric and oceanic CO2 concentrations are rising at an unprecedented rate. Laboratory studies indicate a positive effect of rising CO2 on phytoplankton growth until an optimum is reached, after which the negative impact of accompanying acidification dominates. Here, we implemented carbonate system sensitivities of phytoplankton growth into our global biogeochemical model FESOM-REcoM and accounted explicitly for coccolithophores as the group most sensitive to CO2. In idealized simulations in which solely the atmospheric CO2 mixing ratio was modified, changes in competitive fitness and biomass are not only caused by the direct effects of CO2, but also by indirect effects via nutrient and light limitation as well as grazing. These cascading effects can both amplify or dampen phytoplankton responses to changing ocean pCO2 levels. For example, coccolithophore growth is negatively affected both directly by future pCO2 and indirectly by changes in light limitation, but these effects are compensated by a weakened nutrient limitation resulting from the decrease in small-phytoplankton biomass. In the Southern Ocean, future pCO2 decreases small-phytoplankton biomass and hereby the preferred prey of zooplankton, which reduces the grazing pressure on diatoms and allows them to proliferate more strongly. In simulations that encompass CO2-driven warming and acidification, our model reveals that recent observed changes in North Atlantic coccolithophore biomass are driven primarily by warming and not by CO2. Our results highlight that CO2 can change the effects of other environmental drivers on phytoplankton growth, and that cascading effects may play an important role in projections of future net primary production.

Full text

RESEARCH ARTICLE
Cascading effects augment the direct impact of
CO
2
on phytoplankton growth in a biogeochemical
model
Miriam Seifert
1,
*, Cara Nissen
1
,Bjo¨rn Rost
1,2
, and Judith Hauck
1
Atmospheric and oceanic CO
2
concentrations are rising at an unprecedented rate. Laboratory studies indicate
a positive effect of rising CO
2
on phytoplankton growth until an optimum is reached, after which the negative
impact of accompanying acidification dominates. Here, we implemented carbonate system sensitivities of
phytoplankton growth into our global biogeochemical model FESOM-REcoM and accounted explicitly for
coccolithophores as the group most sensitive to CO
2
. In idealized simulations in which solely the
atmospheric CO
2
mixing ratio was modified, changes in competitive fitness and biomass are not only
caused by the direct effects of CO
2
, but also by indirect effects via nutrient and light limitation as well
as grazing. These cascading effects can both amplify or dampen phytoplankton responses to changing ocean
p
CO
2
levels. For example, coccolithophore growth is negatively affected both directly by future
p
CO
2
and
indirectly by changes in light limitation, but these effects are compensated by a weakened nutrient limitation
resulting from the decrease in small-phytoplankton biomass. In the Southern Ocean, future
p
CO
2
decreases
small-phytoplankton biomass and hereby the preferred prey of zooplankton, which reduces the grazing
pressure on diatoms and allows them to proliferate more strongly. In simulations that encompass CO
2
-
driven warming and acidification, our model reveals that recent observed changes in North Atlantic
coccolithophore biomass are driven primarily by warming and not by CO
2
. Our results highlight that CO
2
can change the effects of other environmental drivers on phytoplankton growth, and that cascading
effects may play an important role in projections of future net primary production.
Keywords: Carbonate chemistry,Phytoplankton,Growth rate,Calcification,Coccolithophores,Climate
change
1. Introduction
Marine primary production, mostly performed by unicel-
lular algae (Middelburg, 2019), contributes to the global
carbon cycle by the fixation of CO
2
in organic biomass via
photosynthesis in the surface ocean and subsequent sink-
ing and remineralization of biomass at depth (organic
carbon pump). By the process of calcification, on the other
hand, CO
2
is released into the surrounding water (carbon-
ate pump). About half of the pelagic calcification is per-
formed by the phytoplankton group of coccolithophores
(Brownlee and Taylor, 2004; Beardall and Raven, 2013),
which therefore contributes to both the organic carbon
and the carbonate pump. Although their contribution to
global primary production is relatively small (<10%; Poul-
ton et al., 2007), some coccolithophore species can form
large blooms that are observable from satellites (Hopkins
and Balch, 2018) and autonomous ocean profilers (Terrats
et al., 2020). Locally, calcification can thereby considerably
dampen the net oceanic uptake of atmospheric CO
2
(Shu-
tler et al., 2013).
Due to increasing anthropogenic CO
2
emissions, envi-
ronmental conditions in the oceans are currently chang-
ing at an unprecedented rate, resulting in oceans that are
warmer, more acidic, less oxygenated, and poorer in
surface-ocean nutrients (Gruber, 2011; Kwiatkowski
et al., 2020). In response to these changes, a decrease in
global marine primary production (Moore et al., 2018) and
shifts in species composition (e.g., Barton et al., 2016;
Poloczanska et al., 2016) are expected in the future. Effects
on the organic carbon and carbonate pump can already be
observed locally (e.g., Bindoff et al., 2019; Pinkerton et al.,
2021). For instance, observations in the North Atlantic
suggest up to a 10-fold increase in coccolithophore abun-
dance over the last decades (1965-2010, Rivero-Calle et al.,
2015; 1990-2012, Krumhardt et al., 2016). Yet, the under-
lying reasons for this increase are not understood conclu-
sively, with different studies suggesting either warming
(Beaugrand et al., 2013) or a positive effect of dissolved
CO
2
(CO
2(aq)
; Rivero-Calle et al., 2015; Krumhardt et al.,
1
Alfred-Wegener-Institut Helmholtz-Zentrum fu
¨r Polar- und
Meeresforschung, Am Handelshafen, Bremerhaven, Germany
2
Universita¨t Bremen, Bremen, Germany
* Corresponding author:
Email: miriam.seifert@awi.de
Seifert, M, et al. 2022. Cascading effects augment the direct impact of
CO
2
on phytoplankton growth in a biogeochemical model.
Elem Sci Anth
,
10: 1. DOI: https://doi.org/10.1525/elementa.2021.00104

2016) as the primary cause for higher coccolithophore
abundances.
Temperature, nutrient concentrations, and light avail-
ability have long been considered as the most important
environmental drivers that control the magnitude of phy-
toplankton growth and coccolithophore calcification.
However, even the carbonate system itself influences
growth and calcification and becomes increasingly rele-
vant given the current pace of ocean acidification. The
impact of the carbonate system can be described with
an optimum function: CO
2(aq)
and bicarbonate (HCO
3)
are the primary substrates of growth and calcification
(e.g., Kottmeier et al., 2014) and rising substrate levels
(carbonation) can enhance growth and calcification. Rising
CO
2(aq)
concentrations result, however, in a concomitant
increase in proton concentrations and thus a lower pH
(acidification), which in turn eventually dampens growth
and calcification (e.g., Rost and Riebesell, 2004; Gao et al.,
2019). Phytoplankton species and groups differ consider-
ably in their ability to make use of carbonation and in
their sensitivity to acidification (e.g., Rost et al., 2008;
Dutkiewicz et al., 2015). For instance, diatoms are affected
only moderately by changes in CO
2(aq)
and HCO
3concen-
trations due to their exceptionally effective CO
2
concen-
trating mechanisms (CCMs; Reinfelder, 2011; Pierella
Karlusich et al., 2021). By diversely changing the compet-
itive fitness of species and groups, CO
2
has also an indirect
effect on phytoplankton community structure (e.g., Doney
et al., 2009; Paul and Bach, 2020). Calcification, however,
is particularly sensitive to ocean acidification, and labora-
tory experiments show that coccolithophore growth is
also impacted disproportionately by increasing CO
2(aq)
levels (Hoppe et al., 2011; Meyer and Riebesell, 2015;
Zhang et al., 2019; Seifert et al., 2020). The opposing
effects of carbonation and acidification lead to a transition
of beneficial to detrimental impacts with increasing par-
tial pressure of CO
2
(pCO
2
), which has been described
mechanistically by several authors (Bach et al., 2011; Bach
et al., 2015; Gafar et al., 2018; Paul and Bach, 2020).
Marine biogeochemical models are important tools
with predictive capabilities to understand connections
and variations of biological, chemical, and biogeochemical
components in space and time (e.g., Holt et al., 2014). In
these models, temperature, light, and nutrients are tradi-
tionally considered as the environmental drivers to control
phytoplankton growth, and very few models to date
account for the effects of CO
2
on growth and calcification.
Dutkiewicz et al. (2015) assessed the pCO
2
impact on all
phytoplankton functional groups in the ecosystem model
DARWIN from preindustrial to 2100 levels under the high
emission scenario (RCP8.5) using different conceptual
approaches based on linear, Michaelis-Menten type, and
Hill type functions. They found that the picocyanobacteria
Synechococcus and diazotrophs benefit most from increas-
ing pCO
2
, while the biomass of other groups, including
coccolithophores and diatoms, tend to decrease. In Krum-
hardt et al. (2019), only coccolithophore growth was con-
sidered to be limited by low CO
2(aq)
concentrations using
a Monod function, similar to other nutrient limitations.
The ratio of particulate inorganic (PIC) to organic carbon
(POC) production by coccolithophores (PIC:POC)
cocco
and,
thus, the calcite production for a given coccolithophore
biomass, was defined to decrease linearly with increasing
CO
2(aq)
. Globally, model runs yielded a decrease in cocco-
lithophore calcification by 17%, but an increase in cocco-
lithophore growth at 900 μatm atmospheric CO
2
relative
to present-day CO
2
levels. Gangstø et al. (2011) did not
account for CO
2
effects on growth, but described a Michae-
lis-Menten-like dependence of the PIC:POC ratio on the
calcite saturation state. In their model, implicit total cal-
cification decreases by 20–60%in a high-emission sce-
nario by the year 2100. Thus, while the CO
2
impacts on
coccolithophores are not consistent across previous stud-
ies, models generally agree that increasing CO
2
can con-
siderably alter the community structure and lead to
a decrease in calcification.
None of these models, however, explicitly considers the
impact of carbon speciation and the resulting two-sided
carbonation-acidification effect on both growth and calci-
fication of phytoplankton. Further, CO
2
effects on phyto-
plankton growth and calcification are currently largely
omitted in the models of the Coupled Model Intercom-
parison Project (CMIP) Phase 6 (Eyring et al., 2016), which
are used for the Sixth Assessment Report (AR6) of the
IPCC. Besides, due to their considerable global contribu-
tion to the carbonate pump and the particular sensitivity
of calcification to ocean acidification, a thorough assess-
ment of CO
2
effects requires the implementation of coc-
colithophores as a separate phytoplankton group. So far,
they have only been included in a few global and regional
models, mainly representing the bloom-forming species
Emiliania huxleyi (Le Que
´re
´et al., 2005; Follows et al.,
2007; Gregg and Casey, 2007; Dutkiewicz et al., 2015;
Kvale et al., 2015; Nissen et al., 2018; Krumhardt et al.,
2019), whereas most biogeochemical models do not rep-
resent coccolithophores or calcification explicitly.
In the present study, we aim to quantify the responses
of marine phytoplankton biomass and coccolithophore
calcification to different atmospheric CO
2
levels and dis-
tinguish between the direct and the indirect effects of CO
2
using the physical-biogeochemical model FESOM-REcoM.
First, we describe the implementation of coccolithophores
and their calcification into the model. Based on a literature
review of laboratory experiments, we then developed opti-
mum functions describing the sensitivity to carbonation
and acidification of growth of all modelled phytoplankton
groups (coccolithophores, diatoms, small phytoplankton)
and for calcification. These functions were subsequently
applied in idealized simulations at preindustrial, present-
day, and future atmospheric CO
2
levels. Finally, motivated
by the observational studies of Beaugrand et al. (2013),
Rivero-Calle et al. (2015), and Krumhardt et al. (2016), we
have attempted to disentangle the effects of warming and
ocean acidification on changes in global and North Atlan-
tic coccolithophore biomass during the last decades.
2. Materials and methods
2.1. Implementing coccolithophores into REcoM
We used the global Regulated Ecosystem Model version 2
(REcoM-2-M) coupled to the Finite Element Sea Ice-Ocean
Art. 10(1) page 2 of 32 Seifert et al: CO
2
effects on phytoplankton growth

Model (FESOM 1.4, Schourup-Kristensen et al., 2014;
Wang et al., 2014). The ocean model FESOM applies a finite
element method that solves primitive hydrostatic equa-
tions on an unstructured mesh, allowing for flexible
multi-resolution meshes (Sidorenko et al., 2011; Wang
et al., 2014). REcoM-2-M describes the biogeochemical
cycling of carbon, nitrogen, silicon, iron, and oxygen
(Hauck et al., 2013; Karakus¸ et al., 2021). The CO
2
flux
from the atmosphere to the ocean and the 3D carbonate
system are computed by the mocsy 2.0 routine (Orr and
Epitalon, 2015). In the previous version of the model
(Karakus¸ et al., 2021), the ecosystem is described by two
phytoplankton groups (diatoms and small-sized phyto-
plankton), two zooplankton groups (small, fast-growing
zooplankton and slow-growing polar macrozooplankton),
and two detritus groups (small, slow-sinking and large,
fast-sinking particles). While the classification of diatoms
is taxonomic, the small phytoplankton comprises a wide
range of taxa, such as non-silicifying and non-calcifying
haptophytes and green algae. In this study, we have
extended the phytoplankton groups by coccolithophores,
which we consider as a taxonomically separate group,
similar to diatoms.
Changes in biomass result from growth and loss terms.
We describe environmental impacts on growth of all phy-
toplankton groups, with a special focus on the parameter
choice for coccolithophores. Parameters were chosen to lie
within ranges based on qualitative literature reviews, and
further tuned to reproduce observed coccolithophore bio-
mass distributions (MAREDAT dataset; O’Brien et al.,
2013). While the bloom-forming coccolithophore species
E. huxleyi is numerically the most important coccolitho-
phore species (Paasche, 2001) and most frequently used in
laboratory studies, it is not a typical representative of
coccolithophores (Rost and Riebesell, 2004). Therefore,
we aimed to represent the average characteristics of the
diverse group of coccolithophores in our model in the
parameter tuning. Subsequently, we describe the explicit
implementation of coccolithophore calcification and CO
2
-
dependent calcite dissolution.
The growth rate GR of phytoplankton group p(cocco-
lithophores, diatoms, or small phytoplankton) depends on
the group-specific constant maximum growth rate μmax
(Table 1), nutrient and light limitation terms ( fNand
fL), and a temperature function (fT):
GRp¼μmax
pfN
pfT
pfL
p:ð1Þ
According to Margalef’s mandala (Margalef, 1978;
Wyatt, 2014), diatoms usually follow the productive r-
strategy with higher growth rates, and coccolithophores the
efficient K-strategy with lower growth rates. While our
group of small phytoplankton includes various species that
presumably cover a wide range of maximum growth rates,
we assume here that the majority of coccolithophore
Table 1. Values of phytoplankton parameters in REcoM-2-M
Parameter
a
Description Units Cocco
b
Dia
c
Sphy
d
μmax Maximum growth rate constant d
1
2.80 3.50 3.00
kDIN N half-saturation constant mmol N m
3
0.90 1.00 0.55
kDFe Fe half-saturation constant μmol Fe m
3
0.09 0.12 0.04
kDSi Si half-saturation constant mmol Si m
3
—4.00 —
qmax Maximum N:C ratio mol N (mol C)
1
0.15 0.20 0.20
qmin Minimum N:C ratio mol N (mol C)
1
0.04 0.05 0.05
qSi
max Maximum Si:C ratio mol Si (mol C)
1
—0.80 —
qSi
min Minimum Si:C ratio mol Si (mol C)
1
—0.04 —
σDIN Maximum N:C uptake ratio mmol N (mmol C)
1
0.20 0.20 0.20
σDSi Maximum Si:C uptake ratio mmol Si (mmol C)
1
—0.20 —
qChl:N
max Maximum Chl:N ratio mg Chl (mmol N)
1
3.50 4.20 3.15
aInitial slope of photosynthesis-irradiance curve mmol C (mg Chl)
1
(W m
2
d)
1
0.10 0.19 0.14
τzoo1Grazing preferences of first zooplankton group dimensionless 0.666 0.083 0.250
τzoo2Grazing preferences of second zooplankton group dimensionless 0.5 1.0 0.5
dChl Maximum Chl loss rate d
1
0.50 0.50 0.50
ηMaintenance respiration rate d
1
0.01 0.01 0.01
a
Parameters relevant for the computations of growth rates are discussed in Section Implementing coccolithophores into REcoM.
b
Coccolithophores.
c
Diatoms.
d
Small phytoplankton.
Seifert et al: CO
2
effects on phytoplankton growth Art. 10(1) page 3 of 32

species, other than the bloom-forming E. huxleyi,aregrow-
ing rather slowly and are specialized to low nutrient regions
and seasons. Thus, we chose a smaller μmax for coccolitho-
phores than for diatoms and small phytoplankton in the
model (Table 1).
Nutrient limitation fNin Equation 1 is determined by
the most limiting nutrient (dissolved inorganic nitrogen,
DIN; dissolved iron, DFe; and, for diatoms, also dissolved
silicic acid, DSi), and is defined as 0 fN1with
0 denoting strongest limitation and 1 denoting no lim-
itation. Limitations by DIN and DSi (lDIN and lDSi)
depend on the intracellular nitrogen or silicate to carbon
quotas, and the group-specific half-saturation constants
kDIN and kDSi (Table 1) for DIN and DSi, respectively
(Geider et al., 1998; Hauck et al., 2013). Limitation by
DFe (lDFe) is described by a Michaelis-Menten function,
which is dependent on the group-specific half-saturation
constant kDFe (Table 1) and the concentration of DFe in
the watercolumn (i.e., ½DFe=ð½DFeþkDFeÞ). For cocco-
lithophores and small phytoplankton, nutrient limitation
is then defined as:
fN
p¼minðlDIN
p;lDFe
pÞ:ð2Þ
Diatoms are additionally affected by DSi limitation
(lDSi). They have a high uptake rate at high nutrient
concentrations (e.g., luxury consumption) and are less
efficient under low nutrient concentrations (Sarthou
et al., 2005; Litchman et al., 2007; Maran
˜o
´netal.,
2013). Small phytoplankton, by contrast, have the
greatest ability to thrive under low nutrient conditions,
e.g. due to their small cell size (Irwin et al., 2006;
Maran
˜o
´n et al., 2013; Irion et al., 2021). Half-
saturation constants kDIN and kDFe of coccolithophores
were chosen to be smaller than those of diatoms and
higher than those of small phytoplankton because their
luxury consumption for nitrate is low compared to dia-
toms, and they are less competitive under low nutrient
concentrations than other phytoplankton groups (Rieg-
man et al., 2000; Rost and Riebesell, 2004). Literature
values for maximum and minimum nitrate-to-carbon
(N:C) ratios (qmax and qmin) do not show a robust differ-
ence between phytoplankton groups (Finkel et al.,
2009). For diatoms and small phytoplankton, qmax and
qmin (Table 1) translate into C:N ratios between 5 and
20, which are well in line with values reported from
laboratory studies of diatoms (C:N range: 2.7–29.7;
Sarthou et al., 2005). For coccolithophores, lower
values result from model tuning, and qmax and qmin
translate into C:N ratios between 6.6 and 25. For the
temperature sensitivity fTin Equation 1, we used
a power function proposed by Fielding (2013) for coc-
colithophores (fT
cocco), which is based on the experimen-
tal growth rate-temperature relations of E. huxleyi
(Figure S1):
fT
cocco ¼0:1419 T0:8151
deg ;ð3Þ
with Tdeg being the temperature 0C. At temperatures
<0C, the function was set to a small value (2.23 10
16
)
rather than zero for numerical reasons. For diatoms and
small phytoplankton, fTis defined by an Arrhenius func-
tion (Figure S1):
fT
Arrhenius ¼exp 4500 1
TK
1
TK;ref
"# !
;ð4Þ
with T
K
being the temperature in K, and TK;ref being the
reference temperature of 288.15 K (15C). Light limitation
fLin Equation 1 is calculated as a function of the group-
specific maximum light harvesting efficiency a(Table 1),
which represents the initial slope of the photosynthesis-
irradiance curve, μmax (Table 1), photosynthetically active
radiation (PAR), the variable chlorophyll-to-carbon ratio
qChl:C, as well as nutrient limitation fN(Equation 2) and
the temperature function fT(Equations 3 and 4; Geider
et al., 1998):
fL
p¼1exp apqChl:C
pPAR
μmax
pfN
pfT
p
!
:ð5Þ
Consequently, changes in either the chlorophyll synthe-
sis or the biomass production and the resulting alteration
of qChl:Caffect the light limitation term. While coccolitho-
phores are tolerant of high light levels, their ability to
grow under low light conditions is lower compared to
diatoms (Nielsen, 1997; Paasche, 2001; Le Que
´re
´et al.,
2005; Zondervan, 2007). In our model, coccolithophores
therefore have a small maximum light harvesting effi-
ciency aand a lower qChl:N
max than diatoms (Table 1).
Loss terms of phytoplankton biomass are composed
of grazing, aggregation, respiration, and excretion of
dissolved organic matter. Grazing is parametrized with
the Fasham formulation (Fasham et al., 1990; Vallina
et al., 2014) and a variable grazing preference depending
on a group-specific initial preference term and the rela-
tive contribution of each phytoplankton group to total
phytoplankton biomass (Karakus¸ et al., 2021). We
assumed that the calcite shell of coccolithophores does
not serve as grazing protection (Mayers et al., 2020) and
therefore chose a similar effective zooplankton grazing
on coccolithophores as on small phytoplankton. We
acknowledge that a protective function of the calcite
shell is still under debate (e.g., Monteiro et al., 2016)
and that a coccolithophore diet might reduce zooplank-
ton growth rates by reduced digestion rates or increased
swimming efforts due to the mass of the indigestible
coccoliths (Haunost et al., 2021). Because the effective
grazing in the model is dependent on the biomass of the
respective phytoplankton group and because coccolitho-
phore biomass is an order of magnitude smaller than
small-phytoplankton biomass (see Section Representa-
tion of coccolithophores and biogeochemical fluxes in
FESOM-REcoM), we used a higher grazing preference of
the small zooplankton group on coccolithophores than
on small phytoplankton (Table 1) to obtain a similar
grazing loss of both groups. We chose the same grazing
preference of the polar macrozooplankton group on coc-
colithophores as on small phytoplankton as its biomass
is generally highest in areas with low coccolithophore
biomass.
Art. 10(1) page 4 of 32 Seifert et al: CO
2
effects on phytoplankton growth

Deviating from the model version described in Hauck
et al. (2013) in which calcification is performed by a con-
stant fraction of the small phytoplankton with a fixed
ðPIC :POCÞsphy ratio of 1, our model calculates calcifica-
tion (Calccocco) depending on the specific growth rate of
coccolithophores (GRcocco; Equation 1), coccolithophore
biomass (C
cocco
), and a reference PIC:POC ratio,
ðPIC :POCÞref
cocco, that is modified by temperature ( fT
CaCO3)
and DIN limitation ( fN
CaCO3):
Calccocco ¼GRcocco Ccocco ðPIC :POC Þref
cocco fT
CaCO3fN
CaCO3:
ð6Þ
Here, we use a ðPIC :POCÞref
cocco ratio of 1, which is moti-
vated by a compilation of PIC:POC ratios of different
coccolithophore species showing that PIC:POC ratios >1
can occur where highly calcified species like Calcidiscus
leptoporus dominate, and PIC:POC ratios <1 can occur in
the Southern Ocean if the lightly calcified E. huxleyi mor-
photype B/C dominates (Krumhardt et al., 2017). Temper-
ature dependence of calcification fT
CaCO3with decreasing
(PIC:POC)cocco ratio at temperatures <10.6C follows Krum-
hardt et al. (2017):
fT
CaCO3¼0:104 Tdeg 0:108 if Tdeg <10:6C
1ifTdeg 10:6
C
ð7Þ
with Tdeg being the temperature in C. Krumhardt et al.
(2017) describe the modification of the (PIC:POC)cocco ratio
by nutrient availability using an equation of the following
form:
fN
CaCO3¼x½N
½NþkN
cocco
þy

:ð8Þ
In their study, [N] is the PO
4
concentration and kN
cocco is
the coccolithophore half-saturation constant for phos-
phate uptake. In their literature review, Krumhardt et al.
(2017) found a 37%increase in (PIC:POC)cocco from
phosphate-repleted to phosphate-limited condition, from
which they derived x¼–0.48 and y¼1.48 (both unitless).
Because our model does not include phosphate, we
adapted Equation 8 to (PIC:POC)cocco modification under
DIN limitation, with [N] and kN
cocco being the DIN concen-
tration and coccolithophore half-saturation constant for
nitrate uptake (Table 1), respectively. Krumhardt et al.
(2017) also assessed the effect of nitrate limitation on
(PIC:POC)cocco,witha25%higher ratio under nitrate-
limited compared to nitrate-repleted conditions from
which we derived x¼–0.31 and y¼1.31.
Instead of an exponential increase with depth with
a vertical length scale of 3500 m following Yamanaka and
Tajika (1996) as in the previous model version (Hauck
et al., 2013), calcite dissolution λnow depends on
the carbonate ion concentration following Aumont
et al. (2015):
CO2
3 sat ¼kspc
Ca2þ;ð9Þ
λ¼0:197 max 0;1CO2
3
CO2
3 sat

1
;ð10Þ
with Ca
2þ
being the calcium ion concentration, kspc the
stoichiometric solubility product (Mucci, 1983), CO2
3the
carbonate ion concentration, and CO2
3sat the saturated car-
bonate ion concentration. The carbonate system was com-
puted once per week where PAR>1%of surface PAR, and
once per month where PAR1%of surface PAR to increase
computational efficiency. To account for calcium carbonate
dissolution above the carbonate saturation horizon, we intro-
duced an additional dissolution in zooplankton guts. The
acidic environment in guts of starving copepods can dissolve
up to 38%of the calcite taken up by grazing (Pond et al.,
1995; Jansen and Wolf-Gladrow, 2001; White et al., 2018). In
addition, aragonite and high-magnesium forms of calcium
carbonate have a shallower saturation horizon than calcite
and contribute to upper-ocean calcium carbonate dissolu-
tion (Sabine et al., 2002; Feely et al., 2004; Barrett et al.,
2014; Battaglia et al., 2016; Sulpis et al., 2021). Besides, high-
CO
2
microenvironments in aggregates caused by metabolic
activity of microbes can lead to calcium carbonate dissolu-
tion even above the saturation horizon (Milliman et al.,
1999; Sulpis et al., 2021). Hence, we have deliberately over-
estimated gut dissolution with 50%of ingested calcite to
account for aragonite and high-magnesium calcite dissolu-
tion that is not covered by the calcite-based dissolution for-
mulation of Aumont et al. (2015).
2.2. CO
2
dependence for phytoplankton growth and
coccolithophore calcification
For the implementation of CO
2
dependencies, we added
limitation functions fCO2similar to the dependencies of
other environmental drivers to Equations 1 and 6:
GRp¼μmax
pfN
pfT
pfL
pfCO2
pð11Þ
CaCO3cocco ¼GRcocco Ccocco ðPIC :POC Þref
cocco fT
CaCO3
fN
CaCO3fCO2
CaCO3:ð12Þ
Multiple concepts describing growth and calcification
dependencies on the carbonate system exist in the litera-
ture (Bach et al., 2011; Bach, 2015; Bach et al., 2015; Gafar
et al., 2018; Paul and Bach, 2020). Here we used the con-
cept of Bach et al. (2015), which has been developed for
coccolithophore calcification, as it can be well accommo-
dated to our model structure. Furthermore, in contrast to
the concepts of Paul and Bach (2020, modified Monod
equation based on a generic proton to CO
2(aq)
relation-
ship) and Bach et al. (2011, modified Michaelis-Menten
function depending only on fCO
2(aq)
), it allows non-
linear changes between carbon species concentrations.
Because they are less well suited for our model structure,
we refrained from using the concepts of Gafar et al. (2018,
CO
2
term including temperature and light effects and
interactions therein) and Bach (2015, bicarbonate to pro-
ton ratio). The concept of Bach et al. (2015) accounts for
changes in the bicarbonate (HCO
3), CO
2(aq)
, and proton
concentrations in a modified Michaelis-Menten function:
fCO2
p=CaCO3¼aHCO
3
bþHCO
3
expðcCO2ðaqÞÞd10pH;
ð13Þ
Seifert et al: CO
2
effects on phytoplankton growth Art. 10(1) page 5 of 32

with pindicating the CO
2
effect on phytoplankton growth
(diatoms, small phytoplankton, or coccolithophores), and
CaCO
3
indicating the CO
2
effect on coccolithophore calci-
fication. Although the effect of CO
2
on phytoplankton
growth and coccolithophore calcification can be either
limiting (on the carbonation side) or inhibiting (on the
acidification side) depending on the state of the carbonate
system, we use the term “CO
2
limitation” for both in the
remaining text equivalent to nutrient and light limitation.
To derive parameters in Equation 13 we used published
laboratory experiments describing phytoplankton growth
rates and PIC:POC ratios of coccolithophores at different
pCO
2
levels. The collection builds on the literature review
by Paul and Bach (2020), complemented by experimental
data of other publications that fit our selection criteria
(Table 2; for more details see Tables S1–S4). In particular,
the data had to i) belong to one of our modelled phyto-
plankton groups, ii) report the response variable (growth
rate and/or (PIC:POC)cocco ratio) for at least three pCO
2
treatment levels, and iii) indicate the concentrations of all
relevant carbonate system parameters (CO
2(aq)
,HCO
3,
and proton concentrations or pH). By including as many
species, ecotypes, and strains as possible, we acknowledge
intraspecific plasticity, which can alleviate the sensitivities
towards changing environmental conditions (e.g., Wolf
et al., 2018) and, thus, lead to more conservative estimates
of CO
2
-driven changes in phytoplankton growth and coc-
colithophore calcification.
We used the R function nls to fit Equation 13 with
a non-linear least square estimate to the laboratory data
of group-specific phytoplankton growth rates and
(PIC:POC)cocco ratios. We forced the functions through the
origin to constrain the lower end of the curve to reason-
able values (i.e., preventing the fit to yield positive growth
or calcification when concentrations of inorganic carbon
are zero). Prior to that, experimental growth rates and
(PIC:POC)cocco ratios were normalized, with the highest
value within each dataset being scaled to 1. Because these
values are reached over a wide range of carbonate species
concentrations, fitted functions yielding a maximum value
of 1 is not possible mathematically. To obtain functions
that are non-limiting (i.e., reach a value of 1) at ideal
Table 2. Parameter values a,b,c, and dfor phytoplankton growth rates and (PIC:POC)cocco ratios, obtained by
fitting Equation 13 to laboratory data from the references given in the table
Process
a
(dimensionless)
b
(mol kg
1
)
c
(kg mol
1
)
d
(kg mol
1
)RMSE
aReferences
Coccolithophore
growth
1.109 37.67 0.3912 9.4501060.15 d
1
Riebesell et al. (2000), Langer
et al. (2006; 2009), Bach
et al. (2015; 2011), Hoppe
et al. (2011), Krug et al.
(2011), Sett et al. (2014),
Hermoso (2015), Mu
¨ller
et al. (2015), Kottmeier
et al. (2016), and Feng et al.
(2017)
Diatom growth 1.040 28.90 0.8778 2.6401060.12 d
1
Ihnken et al. (2011), Sugie
and Yoshimura (2013;
2016), Tatters et al. (2013),
Trimborn et al. (2013),
Barcelos e Ramos et al.
(2014), Panc
ˇic
´et al. (2015),
Wolf et al. (2018), and Li
et al. (2019)
Small phytoplankton
growth
1.162 48.88 0.2255 1.0231070.16 d
1
Kranz et al. (2009), Garcia
et al. (2013), Kim et al.
(2013), Eichner et al.
(2014), Hennon et al.
(2014), and Hoppe et al.
(2018)
Calcification
(PIC:POC)
1.102 42.38 0.7079 1.3431070.24 (unitless) Riebesell et al. (2000), Langer
et al. (2006; 2009), Bach
et al. (2011; 2015), Hoppe
et al. (2011), Krug et al.
(2011), Sett et al. (2014),
Mu
¨ller et al. (2015), and
Feng et al. (2017)
aThe root mean square error (RMSE) determines to what extent the functions can reproduce the data, with RMSE ¼0 representing
a perfect fit.
Art. 10(1) page 6 of 32 Seifert et al: CO
2
effects on phytoplankton growth

conditions, we forced the functions through the median
that was obtained from the inorganic carbon species con-
centrations at the maximum growth rate or (PIC:POC)cocco
ratio of each dataset. To test whether the fitted functions
would be biased by extreme pCO
2
levels tested in some
experiments (up to 5500 μatm), we also fitted the func-
tion only through data points below 2000 and 1000 μatm,
respectively(FigureS2).Inbothcases,growthand
(PIC:POC)cocco were more sensitive to increasing pCO
2
,
showing that including extreme pCO
2
levels does not
exacerbate but softens the impact of CO
2
on growth and
calcification. Hence, we selected the functions that were
fitted over the entire pCO
2
range. Parameter values result-
ing from the function fits are listed in Table 2. Because
the fits were deducted from experimental carbonate sys-
tem environments, extreme carbonate system environ-
ments in the model, e.g. in the Arctic Ocean with
freshwater influence, resulted in carbonate system sensi-
tivities >1. For the implementation in REcoM-2-M, we
limited the resulting term to be 1 to prevent the car-
bonate system from enhancing growth and calcification to
more than the pre-set maximum value.
Figure 1 displays the function fits for a sample matrix
of bicarbonate, CO
2(aq)
, and proton concentrations. Calci-
fication and diatom growth benefit strongest from
increasing pCO
2
at very low levels due to the carbonation
effect (Figure 1). The steep increase in diatom growth rate
already at low pCO
2
is well in line with the presence of
effective and highly regulated CCMs in the group of dia-
toms (Reinfelder, 2011; Pierella Karlusich et al., 2021). The
relatively stable diatom growth rate over a wide pCO
2
range after the steep initial increase displayed in our func-
tion can mirror the reallocation of energy, which is saved
due to downregulation of CCMs under higher CO
2(aq)
and
bicarbonate availability (Shi et al., 2019). The increment
from preindustrial (280 μatm) to present-day (420 μatm)
pCO
2
levels increases the growth rate of small phytoplank-
ton, whereas the growth rate of the other phytoplankton
groups and the calcification rate decrease in this pCO
2
range. Besides, growth rates of small phytoplankton depict
a relatively broad plateau close to maximum growth rates
between present-day and future (750 μatm) pCO
2
levels
(Figure 1). Studies show that small phytoplankton species
such as Micromonas spp. and single trichomes of Tricho-
desmium spp. benefit from pCO
2
levels higher than prein-
dustrial (Boatman et al., 2018), whereas pCO
2
levels higher
than present-day must not necessarily trigger further stim-
ulation in growth (Maat et al., 2014; Boatman et al., 2018;
White et al., 2020). Calcification decreases most steeply
after reaching an optimum at relatively low pCO
2
in our
function (Figure 1), agreeing with experimental studies
that reported coccolithophore calcification unambigu-
ously to be the process most sensitive to any variability
in the carbonate system, with an overall strong decrease in
(PIC:POC)cocco with increasing pCO
2
(e.g., Findlay et al.,
2011; Meyer and Riebesell, 2015). In natural communities,
this effect is caused by not only a decrease in calcification
per cell, but also a shift from highly to less calcified coc-
colithophore species and morphotypes (Beaufort et al.,
2011). Coccolithophore growth rates in our function ini-
tially decrease more steeply with increasing pCO
2
than
small-phytoplankton growth rates, and are only outpaced
Figure 1. Reaction norms of phytoplankton growth rates and (PIC:POC)cocco ratios. Point markers indicate
normalized phytoplankton growth rates and (PIC:POC)cocco ratios from laboratory experiments (see references in
Tables 2 and S1–S4), and lines indicate fitted functions for growth rates of coccolithophores, diatoms, and small
phytoplankton, as well as (PIC:POC)cocco ratios. They are exemplarily displayed for a carbonate system ranging from
3 to 5500 μatm pCO
2
and a constant alkalinity of 2300 μmol kg
1
(computed with ScarFace web version 1.3.0,
https://raitzsch.shinyapps.io/scarface_web/ as in Gattuso et al. [2019], and Raitzsch and Gattuso [2020], assuming
surface pressure, salinity of 35, temperature of 20C, and zero silicate and phosphate concentrations). Root mean
square errors reveal the reproducibility of the data by the fitted functions. Vertical lines denote pCO
2
levels of the
simulations (280, 420, and 750 μatm). The black inset in the main plot indicates the area enlarged on the right side.
PIC ¼particulate inorganic carbon; POC ¼particulate organic carbon.
Seifert et al: CO
2
effects on phytoplankton growth Art. 10(1) page 7 of 32

by small-phytoplankton growth rates at high pCO
2
levels
(Figure 1). Our fits are in line with experimental studies
which found coccolithophore growth to decrease most
strongly with increasing pCO
2
(Riebesell et al., 2017; Sei-
fert et al., 2020), with the exception of some coccolitho-
phore strains (Meyer and Riebesell, 2015). In summary, we
are confident that our fitted functions based on the con-
cept of Bach et al. (2015) and constrained with laboratory
data simulate the general response patterns of phyto-
plankton groups to different pCO
2
levels reasonably well.
2.3. Model simulations
We performed two sets of model simulations on the
CORE-II mesh with higher resolved dynamic areas (up to
20 km, e.g. coastlines and equatorial belt) and coarser
resolution of less dynamic areas (up to 150 km, open
ocean; Sidorenko et al., 2011; Wang et al., 2014). We used
the JRA-55-do reanalysis data set as atmospheric forcing in
all simulations (version 1.3.1; Tsujino et al., 2018). Temper-
ature and salinity were initialized with hydrographic data
from the Polar Science Center Hydrographic Climatology
(PHC 3.0; updated from Steele et al., 2001), DIN and DSi
from the World Ocean Atlas (Garcia et al., 2013), and DFe
from PISCES output (Aumont et al., 2003) which was cor-
rected using observed profiles (de Baar et al., 1999; Boye
et al., 2001). DIC and alkalinity were initialized from GLO-
DAPv2(Lauvsetetal.,2016),andbiomassfromsmall
values (2.23 10
5
for biomass in nitrogen units and
14.77 10
5
for biomass in carbon units, i.e. converted
with the Redfield ratio).
In the first set of simulations, we aimed to investigate
the effect of different atmospheric CO
2
concentrations on
phytoplankton growth and biomass. We performed in
total four model simulations without and with CO
2
depen-
dence at present-day atmospheric CO
2
levels (PRESENT
and PRESENT_CO2, respectively, 420 μatm; Table 3) and
with CO
2
dependence under preindustrial and future CO
2
levels (PREIND_CO2, 280 μatm, and FUTURE_CO2, 750
μatm; Table 3). The simulations were forced with repeated
year forcing, looping over the year 1961. We initialized DIC
with preindustrial DIC for the PREIND_CO2 simulation,
and with present-day DIC in all others (Lauvset et al.,
2016). Each simulation was run for 32 years, and a mean
over the last five years was analyzed for this study. Two
additional simulations were performed to test the sensi-
tivity of phytoplankton biomass towards the parameter
choice of the CO
2
dependence by increasing and decreas-
ing each parameter by 10%(Figure S3).
In a second set of four simulations, we aimed to disen-
tangle the effects of warming and CO
2
on phytoplankton
biomass over the period 1958 to 2019. In one simulation,
both climatological and atmospheric CO
2
forcing varied
inter-annually (VARCLI_VARCO2; Table 3) which accounts
for both the radiative and the geochemical climate change
effect. Further, a second simulation was forced by repeat-
ing the year 1961 for all atmospheric variables except CO
2
(CONSTCLI_VARCO2; Table 3), while in a third, atmo-
spheric CO
2
was held constant at 280 μatm and all other
variables varied inter-annually (VARCLI_CONSTCO2;
Table 3). With these two simulations, we can separate the
effects of the atmospheric CO
2
increase and other chang-
ing atmospheric forcing fields (air temperature, radiation,
humidity, freshwater fluxes, wind). In a control simulation
(CTRL; Table 3), both repeated year forcing of the year
1961 and constant atmospheric CO
2
were used. All simu-
lations of the second set were started from previously
spun-up model states, where the spin-ups covered the
period 1850 to 1957 using JRA repeated year forcing of
the year 1961 (Tsujino et al., 2018). The spin-up simulation
used to initialize the VARCO2 simulations was forced with
increasing atmospheric CO
2
concentration as in Friedling-
stein et al. (2020), and the spin-up simulation used to
initialize the CONSTCO2 simulations was forced with con-
stant atmospheric CO
2
of 278 ppm. As these spin-ups
Table 3. List of model simulations performed in this study (description in Section Model simulations)
Simulation Name
CO
2
Dependence Atm. CO
2
Level
Other Atm. Forcing
Variablesa
Simulation
Period
Set 1: Testing CO
2
effects on the mean state
PRESENT no 420 μatm RYF 1961b32 years
PRESENT_CO2 yes 420 μatm RYF 1961b32 years
PREIND_CO2 yes 280 μatm RYF 1961b32 years
FUTURE_CO2 yes 750 μatm RYF 1961b32 years
Set 2: Disentangling the effects of CO
2
and warming over the recent past
VARCLI_VARCO2 yes historical interannual varying 1958–2019
VARCLI_CONSTCO2 yes 280 μatm interannual varying 1958–2019
CONSTCLI_VARCO2 yes historical RYF 1961b1958–2019
CTRL yes 280 μatm RYF 1961b1958–2019
aOther atmospheric forcing variables are heat flux (by radiation and temperature), humidity, freshwater fluxes, and wind.
bRYF 1961 ¼repeated year forcing, looping over the year 1961.
Art. 10(1) page 8 of 32 Seifert et al: CO
2
effects on phytoplankton growth

were conducted with the prior model version without
coccolithophores and CO
2
dependence, we reset all bio-
logical tracers (phytoplankton and zooplankton biomass)
to small initialization values, while taking advantage of
the spun-up state for physical fields, carbonate chemistry
and nutrients. This approach was chosen to ensure that
the simulated carbonate chemistry represents the histor-
ical evolution of atmospheric CO
2
. We started our analysis
in year 4 of the simulations when the upper ocean eco-
system reached a quasi-equilibrium.
2.4. Datasets used in the evaluation of
coccolithophores in FESOM-REcoM
Simulated coccolithophore biomass concentrations of the
PRESENT and the PRESENT_CO2 simulations were evalu-
ated with a compilation of observations of global cocco-
lithophore biomass concentrations (MARine Ecosystem
DATa dataset, MAREDAT; O’Brien et al., 2013) and litera-
ture estimates of global phytoplankton groups biomass,
net primary production, and biogeochemical fluxes. Addi-
tionally, we evaluated the diatom representation in our
model with the corresponding MAREDAT observational
dataset (Leblanc et al., 2012). For a quantitative compari-
son of the model with the MAREDAT datasets, the mod-
elled coccolithophore and diatom biomass was
subsampled only for grid points with a corresponding
observational grid point and, thus, does not represent the
entire modelled biomass of each group. Subsequently,
biomass estimates of the model and the MAREDAT data-
sets were separated into oceanic biomes (Fay and McKin-
ley, 2014) to assess the regional representation of
simulated coccolithophores and diatoms compared to
observations. An additional analysis of ecological niche
occupations in the model compared to a statistical model
of observations is described in Text S1.
3. Results
3.1. Representation of phytoplankton biomass and
biogeochemical fluxes in FESOM-REcoM
To evaluate the phytoplankton representation in our
model, we focus on the PRESENT simulation, which does
not include the CO
2
dependence. To assess the impact of
the CO
2
dependence at present-day atmospheric CO
2
levels, we then compare the PRESENT with the PRE-
SENT_CO2 simulation. Simulated annual mean cocco-
lithophore biomass is highest in the North Atlantic, the
equatorial regions, upwelling regions, and the northern
boundary of the Southern Ocean (Figure 2). It thereby
resembles the distribution of diatoms except in the polar
regions, where diatom biomass is highest of all three phy-
toplankton groups, whereas coccolithophore biomass is
very small in high latitudes. Small phytoplankton biomass
is more homogeneously distributed, with slightly lower
biomass in the subtropical gyres and the polar regions
Figure 2. Mean phytoplankton biomass concentrations over the upper ocean (150 m) in the PRESENT model
simulation. The PRESENT simulation does not include CO
2
dependencies of phytoplankton growth rates and
coccolithophore calcification. Note the different colorscales of coccolithophores compared to diatoms and small
phytoplankton.
Seifert et al: CO
2
effects on phytoplankton growth Art. 10(1) page 9 of 32

compared to the remaining ocean. Global coccolithophore
biomass is about one order of magnitude smaller than
that of diatom or small phytoplankton (Figure 2).
The comparison to the MAREDAT dataset (O’Brien et al.,
2013) reveals that our model overestimates coccolitho-
phore biomass in the northern hemisphere and the equa-
torial region, especially in the subtropical seasonally
stratified (ST-SS) north region that covers major parts of
our coccolithophore patch in the North Atlantic (0.3 
10
3
Tg C in the observations versus 5 10
3
TgCinthe
model; Figure 3A). Modelled coccolithophore biomass in
the ST-SS (south) region, covering the coccolithophore
patch at the northern border of the Southern Ocean, is
fairly close to observational data with about 2 10
3
Tg
C. Because the model does not depict the coccolithophore
biomass of 0.1 10
3
Tg C from observations further
south in the subpolar seasonally stratified (SP-SS) region
(Figure 3A), high coccolithophore biomass usually occur-
ring between 40 and 60S in the Southern Ocean (Balch et
al., 2011) is shifted northwards in our model. The cocco-
lithophore niche analysis (see Text S1 and Figure S4)
reveals a good model representation of E. huxleyi and
coccolithophore species in warm regions, while cocco-
lithophore species in cold regions are not represented.
Diatom biomass is in the range of observational data for
most regions, except for an overestimation of modelled
biomass in the ST-SS (north) region and a strong underes-
timation in the equatorial region (Figure 3B; Leblanc et al.,
2012). We attribute this underestimation to a patch with
very high observed diatom biomass in the equatorial Atlan-
tic which is not captured by the model, as well as a mis-
match between high diatom biomass on the model grid
and the MAREDAT grid points for which the model was
sampled. For both coccolithophores and diatoms, trends
in inter-biome biomass differences are similar in the grid
subsampled for MAREDAT datapoints (Figure 3) and the
entire model grid (Figure 2; data not shown as barplots).
Globally integrated annual mean coccolithophore bio-
mass amounts to 0.03 Pg C, corresponding to the upper
end of literature estimates (0.001–0.032 Pg C), and to
2.3%of the total simulated phytoplankton biomass
(Table 4). The contribution of coccolithophores to total
NPP is slightly higher in our model (3.3%of total NPP)
than in literature estimates (2%of total NPP; Table 4).
Diatom biomass (0.49 Pg C), silicate export (84.3 Tmol
yr
1
), and the relative share of diatoms to total NPP
(29.9%) fit well into the fairly wide range of literature
estimates (0.01–0.94 Pg C, 69–185 Tmol yr
1
,and
Figure 3. Representation of model-simulated coccolithophores and diatoms compared to the MAREDAT
dataset. (A) Comparison of coccolithophore biomass from the PRESENT and PRESENT_CO2 simulations and the
MAREDAT dataset (O’Brien et al., 2013), and (B) comparison of diatom biomass from the PRESENT and
PRESENT_CO2 simulations and the MAREDAT data set (Leblanc et al., 2012) in (C) different oceanic biomes (after
Fay and McKinley, 2014). Model data points were selected for available MAREDAT data points, and biomass values were
integrated over the entire water column. Ice biomes (north and south) were omitted in (A) and (B) due to the negligible
amount of coccolithophore biomass and low amounts of diatom biomass in both model and MAREDATdata. Model data
were subsampled for locations where MAREDAT data are available and, thus, do not represent the total coccolithophore
and diatom biomass in the model. SP ¼subpolar; ST ¼subtropic; SS ¼seasonal stratified; PS ¼permanently stratified.
Art. 10(1) page 10 of 32 Seifert et al: CO
2
effects on phytoplankton growth

Table 4. Global estimates of phytoplankton biomass (total and by group), net primary production, and global biogeochemical fluxes compared to literature estimates
Description PRESENT PREIND_CO2 PRESENT_CO2 FUTURE_CO2 Literature References
Biomass
A
O’Brien et al. (2013), observations;
B
Leblanc et al. (2012), observations;
C
Buitenhuis et al. (2013), observations;
D
Gregg and Casey (2007), model study;
E
Behrenfeld and Falkowski (1997), satellite data;
F
Buitenhuis et al. (2013), model study;
G
Sinha et al. (2010), model study;
H
Field et al. (1998), model study;
I
Schneider et al. (2008), model study;
J
Carr et al. (2006), satellite data;
K
Jin et al. (2006), model study;
L
Nelson et al. (1995), model study and observations (reference collection);
M
Findlay et al. (2011), experiments;
N
Zondervan et al. (2001), experiments;
O
Zondervan et al. (2002), experiments;
P
Lee (2001), model study;
Q
Jin et al. (2006), model study;
R
Gangstø et al. (2008), model study;
S
Berelson et al. (2007), model study and observations;
T
Dunne et al. (2007), model study and observations (reference collection);
U
Battaglia et al. (2016), model study;
V
Gehlen et al. (2006), model study;
W
Palevsky and Doney (2018), model study;
X
Henson et al. (2011),
Y
Schlitzer (2000), model study;
Z
Sarmiento et al. (2002), model study;
AA
Tre
´guer and De La Rocha (2013), observations (synthesis);
BB
Tre
´guer et al. (2021), observations (synthesis);
* at different depths.
Phytoplankton biomass
a
in Pg C 1.36 1.34 1.35 1.35 na
b
Cocco in Pg C (%of total) 0.031 (2.3) 0.030 (2.3) 0.030 (2.2) 0.032 (2.4) 0.001–0.032 (0.2–2)
A, C
Dia in Pg C (%of total) 0.49 (35.8) 0.52 (38.4) 0.49 (36.4) 0.50 (36.6) 0.01–0.94 (3–50)
B, C
Sphy in Pg C (%of total) 0.84 (61.9) 0.80 (59.3) 0.83 (61.4) 0.83 (61.0) na
b
(PIC:POC)a
cocco (unitless) 1.2 1.1 1.1 1.1 0.7–1.4
M, N, O
Net primary production
Phytoplankton NPP
a
in Pg C yr
1
30.3 28.6 29.7 29.9 24–56
C, D, E, F, G, H, I, J
Cocco in Pg C yr
1
(%of total) 1.0 (3.3) 0.9 (3.1) 0.9 (3.1) 1.0 (3.5) na
b
(2)
K
Dia in Pg C yr
1
(%of total) 9.0 (29.9) 9.4 (32.9) 9.1 (30.8) 9.2 (30.6) na
b
(15–35)
K, L
Sphy in Pg C yr
1
(%of total) 20.2 (66.8) 18.3 (64.0) 19.7 (66.1) 19.7 (66.0) na
b
Export
Calcite export
a
(Pg C yr
1
) 1.2 1.0 1.1 1.1 0.1–4.7
P, Q, R, S, T, U, V,*
POC export
a
(Pg C yr
1
) 4.8 4.7 4.8 4.8 5–13
T, V, W, X, Y,
*
Rain ratio
a
(unitless) 0.3 0.2 0.2 0.2 0.06–0.4
N, Z,
*
Silicate export
a
(Tmol Si yr
1
) 84.3 84.3 84.0 85.7 69–185
L, T, V, AA, BB,
*
NPP ¼net primary production; Cocco ¼coccolithophores; Dia ¼diatoms; Sphy ¼small phytoplankton; PIC ¼particulate inorganic carbon; POC ¼particulate organic carbon.
a
Biomass and NPP were integrated over the upper 150 m of the water column, export fluxes and rain ratio computed at 100 m depth, and (PIC:POC)cocco calculated in the upper 50 m of the water column
from all grid points with a coccolithophore biomass >1 mg C m
3
. Both detritus groups were used to compute export fluxes and rain ratios.
b
na ¼data not available from literature.

15–35%, respectively; Table 4). With approximately 30 Pg
Cyr
1
, total NPP in our model is at the lower end of
literature estimates (Table 4), which can be attributed at
least partly to an underestimation of small-phytoplankton
productivity, given that both coccolithophore and diatom
NPP are likely at the upper end, and to the absence of
nitrogen fixers in our model that usually increase primary
productivity in the low latitudes of the real ocean. The
(PIC:POC)cocco of 1.2 and the calcite export of 1.2 Pg C yr
1
are well within the range of literature values (0.7–1.4 and
0.1–4.7 Pg C yr
1
, respectively; Table 4), indicating a good
representation of coccolithophore calcification and calcite
dissolution in the model. With 0.3, the rain ratio (PIC:POC
export at 100 m) from the model is well within the very
broad range of the literature estimates (0.06–1.0;
Table 4).
In the PRESENT_CO2 simulation, coccolithophore bio-
mass is lower in all regions compared to the PRESENT
simulation, especially in the ST-SS (north), the subtropical
permanently stratified (north), and the equatorial region
(Figure 3A). This lower amount of biomass results in
a smaller over-estimation of coccolithophore biomass
compared to the MAREDAT dataset and, thus, in a better
agreement with observations. Differences in diatom bio-
mass between the PRESENT and the PRESENT_CO2 simu-
lation are very small (Figure 3B). Global estimates of
biomass, production, and export fluxes decrease slightly
compared to the PRESENT simulation, but are still in the
range of literature estimates (Table 4). Varying the para-
meters of the CO
2
dependencies (Table 2)by±10%can
change phytoplankton biomass locally by up to 10%(Fig-
ure S3). However, our analysis shows that the parameter
set used in the PRESENT_CO2 simulation (Table 2), which
is based on function fits to laboratory data, modifies phy-
toplankton biomass only little (less than 10%) compared
to the PRESENT simulation. Hence, we are confident that
our approach of the parameter estimation results in a good
representation of present-day phytoplankton. Overall, our
model represents phytoplankton biomass, distribution,
and export fluxes reasonably well in both the PRESENT
and the PRESENT_CO2 simulation.
3.2. Changes in phytoplankton growth, biomass,
NPP, calcification, and export at different
CO
2
levels
In this section, we first describe spatial patterns of changes
in carbonate chemistry (Figure 4) and phytoplankton bio-
mass (Figure 5). Direct and indirect CO
2
effects are then
analyzed in four selected regions (Figure 6), namely the
three coccolithophore key regions SP-SS þST-SS (north),
equatorial, and ST-SS (south) because of the high sensitiv-
ity of coccolithophores to CO
2
changes (Figure 1), and the
Southern Ocean region, SP-SS þIce (south) where the
spatial comparison between the simulations reveals a con-
siderable decrease of small phytoplankton and a concom-
itant increase of diatoms (Figure 5B and C). Temperature
effects remain the same in all simulations and, thus, are
not discussed. Global estimates of biomass and NPP are
integrated over the upper 150 m of the water column. We
refer to surface maps and, for the separate regions, to
absolute and relative changes at the surface, noting that
relative changes are mostly in the same order of magni-
tude for depth-integrated properties (see Figure S5 for
depth distribution of phytoplankton biomass). If the rela-
tive changes of surface and depth-integrated biomass dif-
fer significantly, we mention them separately.
3.2.1. Preindustrial versus present-day CO
2
Phytoplankton biomass and NPP at preindustrial pCO
2
(280 μatm) are compared to values at present-day pCO
2
(420 μatm) and related to changes in carbonate chemis-
try. On the global scale, largest changes are observed for
small phytoplankton with 2.1%lower total biomass and
NPP (Table 4), driven primarily by changes in the equa-
torial and temperate zones (Figure 5C). This coincides
with the regions of strongest reduction in surface HCO
3
concentrations of up to 8%(about 160 mmol m
3
;Fig-
ure 4B). An exception is the South Pacific subtropical
gyre, where considerably lower surface HCO
3concentra-
tions barely affect small-phytoplankton biomass. Total
biomass and NPP of diatoms are 2.0%and 2.1%higher,
respectively, with lower atmospheric CO
2
concentrations
(Table 4), resulting mainly from changes in the same
regions as for small phytoplankton (Figure 5B). Globally,
biomass and NPP of coccolithophores differ only slightly
between preindustrial and present-day pCO
2
(Table 4;
Figure 5A). Surface CO
2(aq)
is by about 5–10%lower in
the preindustrial compared to the present-day simula-
tion, mostly in the Arctic (up to 10 mmol m
3
)andleast
in the equatorial regions, and pH is up to 0.2 units higher
in the polar regions (Figure 4A and C). Alterations in
total phytoplankton biomass largely follow the slightly
different pattern of change in surface HCO
3as
described above, while CO
2(aq)
and pH seemingly are not
the main factors for differences in phytoplankton bio-
mass in the two simulations. The (PIC:POC)cocco ratio is
about the same in both simulations (Table 4)because
spatial changes in calcite concentration follow changes
in coccolithophore biomass (Figure 5D). POC and calcite
export are slightly lower by 0.1 Pg C yr
1
, resulting in an
unaltered rain ratio, and silicate export is 0.3 Tmol yr
1
higher (Table 4).
In the four selected regions, direct growth responses to
lower atmospheric CO
2
concentrations are not sufficient
to explain the rearrangement of phytoplankton groups.
Growth rates of small phytoplankton are as much as 4%
lower in the SP-SS þST-SS (north), the equatorial, and the
ST-SS (south) regions at preindustrial atmospheric CO
2
(Figure 6A, C, and E). This reduction of the growth rates
is caused by a stronger growth limitation by CO
2
and light,
which is partly counteracted by weakened nutrient limi-
tation (Figure 4D and E). Hence, small-phytoplankton
biomass in these regions is 3–4%lower (0.1–0.3 Tg C).
In the same regions, diatom biomass is 2–5%(0.05–0.1 Tg
C) higher. Changes in their growth rates, however, are not
caused by lower atmospheric CO
2
directly (Figure 6A, C,
and E) but by the modulation of other factors. Coccolitho-
phores are affected most in the SP-SS þST-SS (north)
region, where biomass is reduced by 14%(5%integrated
over the entire watercolumn), which does not correspond
Art. 10(1) page 12 of 32 Seifert et al: CO
2
effects on phytoplankton growth

to the comparatively small change in their growth rate
(Figure 6A) and must have been caused by other factors.
In comparison to the SP-SS þST-SS (north) region, cocco-
lithophore growth rates and biomass vary only little in the
equatorial and the ST-SS (south) region (Figure 6C and E).
As calcification is less inhibited by protons at preindustrial
atmospheric CO
2
, it can partly counteract the reduction in
calcite concentrations caused by lower coccolithophore
biomass in the SP-SS þST-SS (north) region (Figure 6A).
In the Southern Ocean, this disproportional effect of higher
pH on calcification and on growth also explains why the
difference in calcite concentration between present and
preindustrial atmospheric CO
2
is stronger than in cocco-
lithophore biomass (Figure 6G). Absolute differences in
coccolithophore biomass and calcite, however, are negligi-
ble in this region due to the low initial concentrations, and
changes in diatom and small-phytoplankton biomass are
small (Figure 6G). Generally, differences in NPP follow dif-
ferences in biomass, with slight discrepancies of up to 5%
caused by changes in loss rates (Figure 6A and E). Alto-
gether, small phytoplankton is most negatively and diatoms
are most positively affected by preindustrial compared to
present-day atmospheric CO
2
concentrations, while the
global effects on coccolithophores and calcite are smaller
and changes are confined to the regional scale.
3.2.2. Future versus present-day CO
2
Between future atmospheric CO
2
levels (750 μatm) and
present-day CO
2
levels (420 μatm), globally averaged phy-
toplankton biomass and NPP change only marginally (less
than 0.4%;Table 4). Except for slightly higher silicate
exports at future atmospheric CO
2
concentrations (from
84 to 85.7 Tmol yr
1
;Table 4), these marginal changes
implyonlyasmallCO
2
effect on globally averaged
quantities.
In contrast to the global averages, CO
2
has pronounced
regional and group-specific effects (Figures 5 and 6B, D,
F,andH). On a regional scale, future coccolithophore
growth rates are lower compared to present-day rates
because of altered sensitivities to CO
2
and light, with up
to 5%stronger limitations in both cases. Weakened nutri-
ent limitation by up to 9%, primarily caused by DIN (Fig-
ure 4D)butalsobyDFe(Figure 4E), as well as the
modulation of other factors (see Section Unraveling the
cascading effects of changing atmospheric CO
2
levels)
regionally compensates the growth-decreasing CO
2
and
light effects (Figure 6B, D, and F). As a result, coccolitho-
phore biomass is 12%and 14%higher (2%and 5%in the
entire water column), respectively, in the equatorial and
the SP-SS þST-SS (north) region, and remains rather
unchanged in the ST-SS (south) region. In all regions, the
Figure 4. Surface inorganic carbonate system variables and nutrients in model simulations. Maps display (A)
CO
2(aq)
,(B)HCO
3, (C) pH, (D) dissolved inorganic nitrogen (DIN), and (E) dissolved iron (DFe) in the PRESENT_CO2
simulation (middle column), and as total difference between the PRESENT_CO2 simulation and the PREIND_CO2
(left) and FUTURE_CO2 simulations (right), respectively (PREIND_CO2 – PRESENT_CO2 and FUTURE_CO2 –
PRESENT_CO2). Note the logarithmic color scale for present-day DIN and DFe concentrations (middle column).
Seifert et al: CO
2
effects on phytoplankton growth Art. 10(1) page 13 of 32

negative impact of CO
2
on calcification increases by 8–
10%(Figure 6B, D, and F). The biomass-related higher
calcification and calcite concentrations are thereby damp-
ened in the SP-SS þST-SS (north) and equatorial regions,
and both calcification and calcite concentrations are 8%
lower in the ST-SS (south) region where coccolithophore
biomass barely changes (Figure 6B, D, and F). Diatoms
and small-phytoplankton biomass are barely affected by
higher atmospheric CO
2
concentrations in the SP-SS þ
ST-SS (north), equatorial, and ST-SS (south) regions
(Figure 6B, D, and F).
In the Southern Ocean (SP-SS þIce (south) region),
changes in CO
2(aq)
and pH are strongest with 10–20%
higher surface CO
2(aq)
(up to 15–20 mmol m
3
) and lower
pH (up to 0.3 units; Figure 4A and C). Here, more growth-
decreasing effects of CO
2
and light, partly counteracted by
weakened nutrient limitation (Figure 4D and E), lower
small-phytoplankton growth rates by about 3%and bio-
mass by about 5%(Figure 6H). CO
2
, light, and nutrient
effects on diatoms and the resulting growth rate changes
are less pronounced than for small phytoplankton, but
have the same sign. Despite that similarity, diatom bio-
mass increases by 8%, implying modulations of other fac-
tors (see Section Unraveling the cascading effects of
changing atmospheric CO
2
levels). Relative changes in coc-
colithophore biomass and calcite in the Southern Ocean
are negligible due to the generally low biomass and calcite
concentrations.
As in the PREIND_CO2 simulations, changes in NPP
between the PRESENT_CO2 and the FUTURE_CO2 simu-
lation follow changes in biomass, but vary slightly in mag-
nitude with deviations up to 5%compared to biomass
differences (Figure 6B). In summary, the most pro-
nounced changes compared to present-day CO
2
can be
seen in the regionally higher coccolithophore biomass and
its impact on calcite, as well as a shift to fewer diatoms
and more small phytoplankton in the Southern Ocean at
future atmospheric CO
2
.
3.3. Unraveling the cascading effects of changing
atmospheric CO
2
levels
We identify three cascading effects triggered by changes in
atmospheric CO
2
, which can modulate or even counteract
the direct CO
2
effect on phytoplankton growth and bio-
mass. CO
2
affects the nutrient and light limitation terms
which feed back on the growth rate. The seasonal variation
of the nutrient and light feedbacks determine the transla-
tion of direct CO
2
effects into biomass. Finally, we found an
additional role of cascading effects via top-down factors. In
the following, we explain each of the cascading effects for
selected phytoplankton groups and regions (Figure 7A).
3.3.1. Nutrient and light feedbacks
Coccolithophore growth rates in the FUTURE_CO2 simu-
lation are strongly modified by alleviated nutrient limita-
tion, stronger light limitation (Figure 6B, D, and F), and
lower chlorophyll-to-carbon ratios (Figure S6) compared
to the PRESENT_CO2 simulation. Nutrient feedbacks are
caused by a shift in the community structure and total
biomass, leading to changes in available nutrients
(Figure 7B).Lightfeedbacksaremorecomplex
(Figure 7C). Firstly, the nutrient limitation term fNmodi-
fies the light limitation term fLdirectly, with a stronger
Figure 5. Surface phytoplankton biomass and coccolithophore calcite in model simulations. Maps display (A–C)
surface phytoplankton biomass and (D) coccolithophore calcite concentration in the PRESENT_CO2 simulation
(middle column), and as total difference between the PRESENT_CO2 simulation and the PREIND_CO2 (left) and
FUTURE_CO2 (right) simulations, respectively (PREIND_CO2 – PRESENT_CO2 and FUTURE_CO2 – PRESENT_CO2).
Art. 10(1) page 14 of 32 Seifert et al: CO
2
effects on phytoplankton growth

light limitation under more replete nutrient conditions
(Equation 5). Secondly, fNmodifies the chlorophyll-to-
carbon ratio qChl:Cin fL(Figure S6), because chlorophyll
synthesis is dependent on biomass production of a phyto-
plankton group, but also varies with nitrogen assimilation
and PAR (Hauck et al., 2013) which causes non-
proportional changes in cellular chlorophyll and carbon
concentrations. Phytoplankton can make less use of a pre-
vailing light condition with a lower qChl:C. Finally, PAR has
a direct impact on fL, and is itself modified by the pre-
vailing chlorophyll concentration in the watercolumn.
Thus, lower nutrient limitation comes at the expense of
stronger light limitation. Hence,thenegativedirect
impact of high CO
2
levels on coccolithophores is modu-
lated by these two feedbacks.
3.3.2. Seasonal variations of feedbacks
The nutrient and light feedbacks imposed by CO
2
depen-
dencies vary seasonally, and manifest themselves most
strongly in changes in biomass and NPP during the grow-
ing season (Figure 7B and C). For instance, while the
negative effect of CO
2
and light on coccolithophore
growth in the FUTURE_CO2 simulation can be counter-
acted partly by decreasing nutrient limitation in the
annual mean, this effect cannot fully explain the consid-
erably higher coccolithophore biomass in the SP-SS þST-
SS (north) region (Figure 6B). Monthly differences
between the FUTURE_CO2 and the PRESENT_CO2 simu-
lations show that coccolithophore growth rates are up to
5%higher between May and September at high atmo-
spheric CO
2
(Figure 8A), caused by lower light and
Figure 6. Variation in phytoplankton measures between preindustrial, present-day, and future simulations.
Barplots indicate relative changes in surface CO
2
, light and nutrient limitation terms, phytoplankton growth rates,
biomass, calcite concentrations, net primary production (NPP), and calcification between the PRESENT_CO2
simulation and the PREIND_CO2 (A, C, E, G) and FUTURE_CO2 (B, D, F, H) simulations, respectively, in four
biomes denoted on the maps. Annual means for all parameters, spatial means for the limitation terms and growth
rates, and spatial integrals for biomasses and NPP are presented. For calcification, we display the impact of nutrients
and CO
2
on the (PIC:POC)cocco ratio, as well as the change in integrated calcite mass; light is not affecting the
(PIC:POC)cocco ratio. A relative increase in a limitation term means that it is less limiting and therefore causes
a higher growth rate and vice versa. SP ¼subpolar; ST ¼subtropic; SS ¼seasonal stratified; PS ¼permanently
stratified, lim. ¼limitation; GR ¼growth rate; Calcif. ¼calcification.
Seifert et al: CO
2
effects on phytoplankton growth Art. 10(1) page 15 of 32

nutrient limitation. As this period roughly coincides with
the growing season with high initial biomass, it causes an
annual mean increase in biomass and NPP. Lower annual
mean growth rates at future compared to present-day CO
2
levels are caused by lower growth rates in winter (10–12%
between October and April) due to a stronger light limi-
tation and a relatively smaller alleviation in nutrient lim-
itation compared to the remaining year, a time where
initial biomass levels are low and changes in biomass and
NPP are negligible. Annual means of growth rates are
therefore often insufficient to explain changes in NPP and
biomass, and the seasonality of cascading effects needs to
be considered.
3.3.3. Grazing feedbacks
Finally, we identified modifications in the grazing pressure
as another indirect CO
2
effect on phytoplankton biomass.
Due to the variable grazing preference in our model (Fas-
ham et al., 1990; Vallina et al., 2014), grazing fluxes
change non-linearly with variations in the share of each
phytoplankton group to total biomass as well as with
changes in the total biomass of zooplankton (Figure
7D). Despite the higher diatom biomass in the
FUTURE_CO2 compared to the PRESENT_CO2 simulation
(about 7%) due to increasing growth rates during summer
in the SP-SS þIce (south) region and the corresponding
higher prey availability, grazing on diatoms decreases by
about 6%intherespectivemonths(Figure 8B). This
decrease is caused by a lower zooplankton biomass,
related to the decreasing small-phytoplankton biomass
(Figure 6H), a food source that cannot be replaced
entirely by diatoms (note grazing preferences in Section
Implementing coccolithophores into REcoM). Lower grazing
on diatoms in austral winter (June–August; Figure 8B)is
a result of less diatom biomass. Grazing feedbacks also
explain higher diatom biomass concentrations in the PRE-
IND_CO2 simulation in all regions except from the South-
ern Ocean (Figure 6A, C, and E). Another example from
the preindustrial SP-SS þST-SS (north) region is the
increasing grazing pressure on coccolithophores and their
consequently lower biomass, which results from a combi-
nation of relatively high coccolithophore biomass in that
region and biomass-decreasing effects of lower CO
2
on
small phytoplankton which are then less available as prey.
Other loss terms of phytoplankton biomass (respiration,
excretion, aggregation) depend directly on the total
Figure 7. Cascading effects of environmental drivers and loss rates that affect the biomass of a single
phytoplankton group in our model. (A) Overview of the processes affecting phytoplankton productivity and
biomass. Besides direct impacts, modifications in CO
2
limitations can induce biomass changes through multiple
feedback loops, i.e. via changes in nutrient and light limitation and in grazing. The temperature sensitivity remains
unchanged at varying CO
2
levels. (B–D) Examples for nonlinear cascading CO
2
effects: (B and C) pathways of changes
in nutrient and light limitations, respectively, for example, for future coccolithophores in all regions except from the
Southern Ocean (SO); and (D) top-down effect by changes in the grazing pressure, for example, on future diatoms on
the SP-SS þIce (south) region.
Art. 10(1) page 16 of 32 Seifert et al: CO
2
effects on phytoplankton growth

phytoplankton and detritus biomass and/or biomass of
the respective phytoplankton group and are not or negli-
gibly affected by shifts in the share between phytoplank-
ton groups. Hence, grazing on a phytoplankton group can
be alleviated if the biomass of a zooplankton group
decreases because its preferred prey is not present in suf-
ficient abundance. In summary, cascading effects of CO
2
on light and nutrient limitations, on their seasonality, and
on the grazing pressure are pivotal to the biomass and
NPP responses of the phytoplankton groups.
3.4. Disentangling effects of historical warming
and CO
2
increases on coccolithophore biomass
changes
In our second set of simulations (Table 3) we aimed to
disentangle the effects of warming and CO
2
on historical
changes in phytoplankton biomass. Globally, depth-
integrated coccolithophore biomass levels in the VARCLI_
VARCO2 simulation are highly variable in the simulated
period (between 24 and 27 Tg C) with lower biomass in
the 1990s–2000s compared to the 1960s–1970s (24–25
Tg C and 26–27 Tg C, respectively; Figure 9A). Over this
time period, temperatures and CO
2(aq)
concentrations
increase by about 0.5C and 0.003 mmol m
3
, respectively
(Figure 9B). Fluctuations in coccolithophore biomass
levels are small at constant climatological forcing (CON-
STCLI_VARCO2 simulation), implying that interannual var-
iability of coccolithophore biomass is driven mainly by
climate variability and, thus, combined alterations in
temperature and CO
2(aq)
concentrations. The lower
CO
2(aq)
concentration in the VARCLI_CONSTCO2 simula-
tion compared to the VARCLI_VARCO2 simulation
Figure 8. Simulated monthly variations in phytoplankton measures. Barplots indicate relative monthly changes
(January–December) in surface light, nutrient and CO
2
limitation (lim.) terms, phytoplankton growth rates, biomasses,
NPP, and grazing rates between the PRESENT_CO2 and the FUTURE_CO2 simulation for (A) coccolithophores in the
SP-SS þST-SS (north) region, and (B) diatoms in the SP-SS þIce (south) region. Both panels also show the relative
change in zooplankton (Zoo) biomass between the PRESENT_CO2 and the FUTURE_CO2 simulation (orange bars) in
the respective region. Green shadings in (A) indicate the period of higher growth rates in the FUTURE_CO2 simulation
compared to the PRESENT_CO2 simulation (May–September). Orange shadings in (B) indicate the months of
decreasing grazing rate due to lower zooplankton biomass (October–February), and blue shadings indicate the
months of decreasing grazing rate due to lower diatom productivity (June–August) in the FUTURE_CO2
simulation compared to the PRESENT_CO2 simulation.
Seifert et al: CO
2
effects on phytoplankton growth Art. 10(1) page 17 of 32

(–0.001 mmol m
3
at the beginning and increasing over
the course of the time series) increases the coccolitho-
phore biomass by almost 1 Tg C in the entire time period.
However, increasing CO
2(aq)
concentrations over time at
constant temperature have almost no effect on the inte-
grated coccolithophore biomass (CONSTCLI_VARCO2 sim-
ulation). Hence, according to our simulations, CO
2(aq)
is
not the main driver for global biomass changes since the
1960s, and variability in global coccolithophore biomass is
driven mainly by climate variability.
In the North Atlantic, mean coccolithophore biomass
levels in the 2000s are 4%higher than biomass levels in
the 1960s (2.4 and 2.3 Tg C, respectively; Figure 9C). Our
model therefore simulates a much smaller increase in
coccolithophore biomass in the North Atlantic between
1960 and 2010 than the up to 10-fold increase in cocco-
lithophore occurrence in observations (Beaugrand et al.,
2013; Rivero-Calle et al., 2015; Krumhardt et al., 2016).
After 2010, biomass levels decrease quickly by 0.2 Tg C,
reaching biomass levels of 2.2 Tg C in 2019. Temperatures
increase rapidly from about 11.5C before the 1990s to
about 12.25C by the end of the time series with strong
interannual variability, whereas CO
2(aq)
concentrations
increase steadily from about 12 10
3
mmol m
3
in the
1960s to about 15.5 10
3
mmol m
3
in the 2010s
(Figure 9D). Thus, just as on a global scale, interannual
biomass variability is caused by climate variability. In con-
trast to what is simulated on a global scale, however,
biomass levels are increasingly higher over the course of
the time series at increasing CO
2(aq)
concentrations (VAR-
CLI_VARCO2 simulation) than at constant CO
2(aq)
concen-
trations (VARCLI_CONSTCO2 simulation, up to 0.1 Tg C in
2019; Figure 9C). This result is most likely due to a growth
enhancing effect of higher CO
2(aq)
concentrations, which
are about 6 10
3
mmol m
3
higher in the VARCLI_
VARCO2 compared to the VARCLI_CONSTCO2 simulation
at the beginning of the time series, with increasing differ-
ence towards the end of the time series (Figure 9D).
However, biomass levels barely change with increasing
CO
2(aq)
concentrations and constant temperature (CON-
STCLI_VARCO2 simulation), pointing towards a small
effect of CO
2(aq)
on coccolithophore biomass in the period
of the time series. We therefore attribute biomass changes
predominantly to climate variability.
4. Discussion
Our model simulations show that adding a CO
2
depen-
dence to phytoplankton growth leads to a cascade of
Figure 9. Time series of drift-corrected coccolithophore biomass, surface CO
2(aq)
concentration, and surface
temperature. Values are displayed (A and B) globally and (C and D) in the North Atlantic for the VARCLI_VARCO2
(solid line), VARCLI_CONSTCO2 (dashed line), and CONSTCLI_VARCO2 (dotted line) simulations (Table 3). Grey line (A
and C) indicates the 5-year moving average of coccolithophore biomass in the VARCLI_VARCO2 simulation.
Differences in CO
2(aq)
(B and D) are caused mostly by different prescribed levels of atmospheric CO
2
,andthe
effects of varying climate on CO
2(aq)
are small (compare solid and dotted lines). Temperatures in the
VARCLI_CONSTCO2 and the VARCLI_VARCO2 simulation are equal and therefore not distinguishable in panels B
and D. Biomass concentrations were integrated over the entire water column. The first three years (1958–1960) were
omitted in all simulations. For drift correction, the deviation between the control (CTRL) time-series and its initial
value in 1961 was subtracted from the other simulations.
Art. 10(1) page 18 of 32 Seifert et al: CO
2
effects on phytoplankton growth

effects on (i) the growth rate of the individual phytoplank-
ton groups by feedbacks on light and nutrient limitations
and the seasonality therein (bottom-up effects), and (ii)
the interaction between phytoplankton and zooplankton
(top-down effects). In the preindustrial global ocean, small
phytoplankton are affected most by direct and indirect
CO
2
effects. The lower biomass of small phytoplankton
in low to mid-latitudes causes lower zooplankton biomass
concentrations and a shift of the grazing pressure to coc-
colithophores, which results regionally in up to 14%lower
coccolithophore biomass. Diatoms, in contrast, benefit
from lower zooplankton biomass rather than suffering
from higher grazing pressure due to chosen grazing pre-
ferences, but effects are rather localized. In the future
global ocean, coccolithophores are affected most by direct
and indirect CO
2
effects, resulting in higher coccolitho-
phore biomass. Changes in coccolithophore biomass affect
the total phytoplankton biomass only marginally and
therefore do not feed back on zooplankton biomass and
grazing rates. In high latitudes, bottom-up effects cause
lower small-phytoplankton biomass with consequently
lower zooplankton biomass, which decreases the grazing
pressure on diatoms. In low to mid-latitudes, bottom-up
effects trigger slightly higher growth rates and biomass
concentrations of small phytoplankton.
These results highlight two important aspects that
shape top-down effects. Firstly, the standing stock of
a group determines whether changes therein matter for
the overall zooplankton prey availability (e.g., small phy-
toplankton versus coccolithophores). Secondly, the grazing
preference determines if and how zooplankton can switch
to a different prey (e.g., small phytoplankton and cocco-
lithophores versus diatoms). The small zooplankton group
in our model grazes preferentially on small phytoplankton
and coccolithophores, and the polar macrozooplankton
grazes mainly on diatoms. While the spatial distribution
of polar macrozooplankton is confined to south of 50S
and north of 50N, this group out-competes here the
small zooplankton group due to its higher grazing effi-
ciency. In the remaining ocean, grazing of the small-
phytoplankton group dominates. A variety of grazing for-
mulations are used in biogeochemical models which differ
in how they describe prey switching and total grazing
pressure, suggesting that using a different grazing param-
etrization can change the simulated grazing pressure on
individual phytoplankton groups (Vallina et al., 2014).
Nonetheless, given the good agreement of the simulated
biomass fields in our simulations with observational data
(Figures 3A and Band S4), we are confident that the
qualitative dynamics of the feedbacks triggered by the
addition of CO
2
dependencies also hold across grazing
parameterizations. The complexity of modelled trophic
interactions with partially unexpected feedbacks was also
highlighted by Dutkiewicz et al. (2021), who showed that
a small reduction in the abundance of one species can
increase the total phytoplankton standing stock due to
shifts to slower growing grazers and the freeing up of
limiting nutrients, which is comparable to the cascading
effects in our model. In the following, we discuss regional
changes in phytoplankton biomass concentrations.
4.1. Cascading effects cause fewer small
phytoplankton in the preindustrial ocean and more
diatoms in the future Southern Ocean
In the preindustrial global ocean, small-phytoplankton
biomass is lower compared to present-day CO
2
levels.
Increasing levels of small-phytoplankton biomass over the
last decades have also been shown in field and laboratory
studies, with the projection of further increasing biomass
with ongoing climate change (Daufresne et al., 2009; Peter
and Sommer, 2012; Pinkerton et al., 2021). This beneficial
effect has been associated with joint warming and increas-
ing CO
2
(Hare et al., 2007; Feng et al., 2009). However, our
model results suggest that even the carbonation effect
alone, mainly caused by the large growth increase from
preindustrial to present-day CO
2
levels, can lead to
increasing biomass of small phytoplankton.
In the future Southern Ocean our model reveals a shift
to less small phytoplankton and more diatom biomass
compared to present-day CO
2
levels (Figures 5B and C
and 6H). This advantage of diatoms over small phyto-
plankton is driven mainly by decreasing grazing pressure
on diatoms (Figure 8B). In nature, the advantage of dia-
toms at future oceanic conditions may be further
enhanced by alterations in the diatom species composi-
tion. Laboratory and mesocosm studies reveal that under
ocean acidification, diatoms likely shift towards larger spe-
cies (Tortell et al., 2008; Wu et al., 2014; Bach et al., 2019),
changing community size structures and composition and
having implications on the food web and biogeochemical
cycling (Finkel et al., 2009; Alvarez-Fernandez et al., 2018).
As currently we do not include cell size structure within
a phytoplankton group in our model, accounting for shifts
in this important trait (Dutkiewicz et al., 2015) could be
a step forward to assess implications on export fluxes,
with POC export likely increasing due to the higher bio-
mass and ballasting of large diatom cells. While in parts of
the Southern Ocean, a deepening of the mixed layer depth
has been observed and modelled for recent years (Hauck
et al., 2015; Panassa et al., 2018; Salle
´e et al., 2021) and
future projections (Hauck et al., 2015), other regions may
experience a temperature-driven stronger stratification
(Constable et al., 2014). Small phytoplankton, because of
their lower nutrient requirement, typically follow South-
ern Ocean diatom blooms that have drawn down nutrient
concentrations (Irion et al., 2021). Similarly, a shoaling of
the mixed layer would be more advantageous to small
phytoplankton than for diatoms (Dutkiewicz et al., 2015;
Petrou et al., 2016). Our results indicate, however, that
increasing CO
2
levels may dampen the nutrient-driven
advantage of small phytoplankton over diatoms in the
future Southern Ocean (Figure 6H).
4.2. Alleviated nutrient limitation outbalances CO
2
-
induced decrease of coccolithophore growth in
a high-CO
2
ocean
While laboratory CO
2
manipulation experiments tend to
show either unaffected or decreasing coccolithophore
growth rates between present-day and future CO
2
levels
(e.g., Bach et al., 2013; Seifert et al., 2020; but see Meyer
and Riebesell, 2015, for the diversity of responses),
Seifert et al: CO
2
effects on phytoplankton growth Art. 10(1) page 19 of 32

coccolithophore biomass in our model is increasing due to
the unexpected cascading effects that offset the negative
effects of CO
2
on growth (Figures 5A and 6B, D, and F).
Interactive effects between warming and high CO
2
condi-
tions indicate a shift of the turning point between carbon-
ation and acidification towards higher CO
2
levels at
warmer temperatures (e.g., Sett et al., 2014; Gao et al.,
2018), which may reduce the negative CO
2
effect on coc-
colithophores in the future scenario and ultimately lead to
higher coccolithophore biomass, just as in our CO
2
-only
simulation. However, while the CO
2
limitation in our
model showed only modest seasonal variations, meso-
cosm experiments have shown that a more severe CO
2
-
induced growth reduction at the beginning of the growing
season can impede bloom formation and thereby reduce
the overall ability to maintain sufficiently large seed popu-
lations for the ensuing growth seasons (Riebesell et al.,
2017). Besides, a climate change-related increase in upper
ocean stratification and the concomitant decrease in nutri-
ent supply could increase the competitive fitness of small
phytoplankton, as explained in Section Cascading effects
cause fewer small phytoplankton in the preindustrial ocean
and more diatoms in the future Southern Ocean,which
could outweigh the negative effect of ocean acidification.
Less nutrients are then available for diatoms and cocco-
lithophores, which decreases their competitive fitness, an
effect that cannot be seen in simulations where only
atmospheric CO
2
is changed. Consequently, the negative
CO
2
effect on coccolithophores could become more prev-
alent in a joint warming and high-CO
2
scenario, and
decrease coccolithophore biomass in the future.
Our model suggests that the CO
2
impact on calcifica-
tion, which is positive in the preindustrial and negative in
the future ocean compared to present-day conditions, is
modified and partly compensated by changes in cocco-
lithophore biomass. Higher calcification rates under pre-
industrial CO
2
levels compared to the present-day in our
study (Figure 6C, E, and G) seem to disagree with findings
of Rigual-Herna
´ndez et al. (2020) who found that the
thickness of the coccoliths barely differs between prein-
dustrial samples, originating from sediment cores, and
modern plankton samples. However, coccolith thickness
and the sensitivity to CO
2
is highly species-specific
(Rigual-Herna
´ndez et al., 2020), and shifts in coccolitho-
phore species composition can impact patterns in calcite
concentration and export (Beaufort et al., 2011). The num-
ber of coccoliths per cell is also variable, at least for E.
huxleyi, and can additionally affect calcite concentrations
under changing environmental conditions (Paasche,
2001). Whether or not there was a global trend towards
lower calcification in all coccolithophore species from pre-
industrial to present-day remains unresolved.
In our simulations, calcification is the most sensitive
process towards future CO
2
levels (Figure 1). Calcite con-
centrations increase in some regions (SP-SS þST-SS (north)
and equatorial) only because of increasing coccolitho-
phore biomass (Figures 5A and 6B and D). Based on the
ballasting hypothesis (e.g., Klaas and Archer, 2002; Cram
et al., 2018), this increase would be indicative of an
increase of export production, which is currently not
represented in our model. Furthermore, while on a global
scale, enhanced calcification and the accompanying CO
2
evolution affects the anthropogenic increase in atmo-
spheric CO
2
only little (Gangstø et al., 2011), it can shape
the regional air-sea CO
2
flux (Shutler et al., 2013).
4.3. Comparison to other modelling studies that
include CO
2
dependencies
Increasing phytoplankton biomass and decreasing calcifi-
cation at future CO
2
levels in our model compare well
with the outcome of other modelling studies that include
aCO
2
dependence of growth rates and/or calcification
(Gangstø et al., 2011; Dutkiewicz et al., 2015; Krumhardt
et al., 2019). Similar to our model, considerable alterations
in global phytoplankton communities attributable to dif-
ferent sensitivities towards increasing CO
2
levels were sim-
ulated in the ecosystem model DARWIN of Dutkiewicz et
al. (2015), whereas warming alone caused a poleward shift
in phytoplankton dispersal with no significant impact on
the community structure. While we used a mechanistic
approach, Dutkiewicz et al. (2015) used a conceptual
approach with a linear and positive dependence of growth
rates on CO
2
. They showed that decreasing nutrient avail-
ability due to stronger stratification lowers globally inte-
grated primary production by 5%in 2100, which can be
counteracted by CO
2
-enhanced growth rates, leading to
almost unchanged global primary production levels in
2100 compared to present-day levels. Similarly, Krumhardt
et al. (2019) attributed global coccolithophore biomass
increase to a direct CO
2
effect in their model. Contrary
to these two studies which only account for the carbon-
ation effect, we also modelled the acidification effect on
phytoplanktongrowthwherebythedirectCO
2
effect
causes a decrease in future growth rates. However, cascad-
ing effects through nutrient and light feedbacks (Figure
7) can lead to an overall increase in biomass. While we did
not simulate changes in circulation and stratification, we
hypothesize that reduced nutrient availability in a more
stratified ocean could reinforce rather than counteract the
CO
2
-driven decrease in growth. Thus, we rather expect
a future decrease in phytoplankton biomass, which cannot
be balanced by CO
2
, different to the idealized simulations
of Dutkiewicz et al. (2015).
In our model, increasing future coccolithophore bio-
mass due to cascading effects over-compensates for the
pH-driven decrease in calcification (Figure 6B and D),
resulting in a net increase of global calcite concentrations
that corresponds to the findings of Krumhardt et al.
(2019). Both our study and that of Krumhardt et al.
(2019) are in line with Gangstø et al. (2011), who pro-
jected a negative direct CO
2
effect on calcite production
in a high emission scenario of the biogeochemical
Bern3D/PISCES model. Going beyond the study of Gang-
stø et al. (2011) by modelling CO
2
effects on coccolitho-
phore biomass, we have shown that biomass changes can
compensate for decreasing calcite production, which is
also in line with Krumhardt et al. (2019). In summary, our
findings highlight the need to consider both carbonation
and acidification effects on phytoplankton growth and
coccolithophore calcification.
Art. 10(1) page 20 of 32 Seifert et al: CO
2
effects on phytoplankton growth

4.4. Increase in North Atlantic coccolithophore
biomass is driven mainly by warming
According to our model, we attribute observed changes in
North Atlantic coccolithophore biomass in recent decades
primarily to warming and not to CO
2
. Thus, our results
largely comply with the warming hypothesis of Beaugrand
et al. (2013), as changes in the modelled carbonate system
during the 60 years of the time series are too small to have
a pronounced effect on coccolithophore biomass. We
thereby contrast the CO
2
hypotheses of Rivero-Calle
et al. (2015) and Krumhardt et al. (2016), who concluded
that warming alone did not control changes in coccolitho-
phore biomass. Even with more significant shifts in the
carbonate system, the negative impact of CO
2
on cocco-
lithophore growth obtained by our model are over-
compensated by more beneficial cascading effects on
nutrient limitation (Section Unraveling the cascading
effects of changing atmospheric CO
2
levels). Admittedly,
none of the temperature functions in our model includes
decreasing growth rates caused by temperature stress,
which can considerably affect community structures
(e.g., D’Amario et al., 2020), but we suppose that ther-
mally stressed species or ecotypes would be replaced by
more heat-tolerant species or ecotypes of the same phy-
toplankton group, resulting in the same changes in coc-
colithophore biomass as simulated here. We expect
thermal stress to become more relevant in polar regions
and/or for a more intense temperature increase that out-
paces the speed of species migration, but not for the given
temperature increase within 60 years in the North Atlan-
tic. We additionally need to bear in mind that our param-
eterization aims to represent the diverse group of
coccolithophores, while the North Atlantic coccolitho-
phore community is dominated by the fast growing,
bloom-forming E. huxleyi (Haidar and Thierstein, 2001;
Balch et al., 2019), which may respond more strongly to
changing CO
2
conditions than coccolithophores in our
model. Besides, the CO
2
effect on coccolithophore growth
would be different if extreme CO
2
levels of laboratory
experiments had been excluded from the parameteriza-
tion. Based on our model results, however, we conclude
that the direct CO
2
impact on coccolithophore biomass in
the recent decades, globally and in the North Atlantic, is
small compared to the effect of warming, supporting the
hypothesis of Beaugrand et al. (2013).
4.5. Limitations and caveats
Our description of CO
2
dependencies goes well beyond
previous studies by developing mechanistic response func-
tions of phytoplankton functional type growth rates and
coccolithophore calcification. We are confident that it cap-
tures the first-order effects of CO
2
on phytoplankton bio-
mass and NPP. Nevertheless, we acknowledge that
additional processes (i.e., multiple driver interactions, mix-
otrophy, CO
2
effects on silicification and zooplankton,
adaptation and evolution, as well as mesozooplankton
grazing) have the potential to further modify the response
of the marine ecosystem to high-CO
2
levels in simulations
and in the real world.
The response of phytoplankton growth and calcifica-
tion to changing CO
2
levels can be modified significantly
by synergistic and antagonistic interactions with other
environmental drivers such as temperature and light
(e.g., Harvey et al., 2013; Brandenburg et al., 2019; Seifert
et al., 2020), and can result in enhanced or dampened
responses than expected from CO
2
alone. However, most
biogeochemical models, including ours, do not account
for the interaction between environmental drivers. For
instance, acidification affects the response to nutrient
availability of coccolithophores (Zondervan, 2007; Zhang
et al., 2019) and other phytoplankton groups (Li et al.,
2012; Spungin et al., 2014; Li et al., 2017). The beneficial
effect of nutrients on coccolithophore biomass in the
FUTURE_CO2 simulation may thus be weakened by acid-
ification, and therefore dampen the biomass increase
compared to present-day conditions. The implementation
of multiple driver interactions would therefore be a crucial
amendment to phytoplankton modelling.
Mixotrophy, which is the capacity of an organism to live
auto- and heterotrophically, can reduce the vulnerability
of phytoplankton to environmental changes, for example,
under substrate limitation (low carbonation) of photosyn-
thesis. Mixotrophy is very common in species that we
summarize as small phytoplankton (Stoecker et al.,
2017), but some coccolithophore species can live mixotro-
phically as well (Godrijan et al., 2020). While our phyto-
plankton growth rate parameterization is based entirely
on photosynthesis, effects of changing CO
2
levels could be
diminished if we additionally considered the possibility to
switchtomixotrophywhenCO
2
or other conditions
become unfavorable. Besides, not only calcification, but
also silicification by diatoms appears to be CO
2
sensitive
(Petrou et al., 2019). Although this sensitivity could alter
the response of diatoms towards increasing CO
2
condi-
tions and, for instance, possibly hamper the success of
diatoms in the future Southern Ocean, the knowledge
about this relationship is still too limited to be incorpo-
rated into modelling approaches. Similarly, CO
2
effects on
grazers and higher trophic levels could further modify the
ecosystem response (Cripps et al., 2014).
Adaptation and evolution can shape substantially the
response of phytoplankton towards high or low pCO
2
levels (Brennan et al., 2017), processes that are not usually
considered in ocean biogeochemical models. For instance,
coccolithophores can adapt within months or years to
warming and high CO
2
conditions (Schlu
¨ter et al., 2014).
However, with the current pace of environmental changes
caused by climate change and the limits of adaptation,
whether adaptational and evolutionary processes are fast
enough to offset negative effects of environmental driver
changes is questionable. The assessment of abilities and
speed of phytoplankton (and other organism) adaptations
is challenging (Kelly and Griffiths, 2021), and the data
basis remains highly varied among studies, which cur-
rently prevents a robust implementation into biogeo-
chemical models.
While our model represents a polar macrozooplankton
as well as a generic small zooplankton group, it does not
include mesozooplankton, which is globally distributed
Seifert et al: CO
2
effects on phytoplankton growth Art. 10(1) page 21 of 32

(Moriarty and O’Brien, 2013) and an important compo-
nent of trophic interactions and the biological carbon
pump (Buitenhuis et al., 2006). Hence, considering meso-
zooplankton could alter the grazing feedbacks presented
in our study, for instance in the Southern Ocean where the
grazing pressure on diatoms is indirectly lowered due to
the decrease of small-phytoplankton biomass.
4.6. Conclusions
Model projections are the primary tool to understand
future shifts in marine productivity and its impact on
higher trophic levels and food webs. Even if the resilience
of phytoplankton communities to climate change-related
impacts might be relatively high because environmental
niches can be re-occupied by species that are better
adapted to the new environmental conditions (Dutkiewicz
et al., 2021), most models project a future decrease in
global NPP, with a large spread in the magnitude across
different models (Kwiatkowski et al., 2020). Carbonation
and acidification may have a relatively smaller effect on
phytoplankton growth than other environmental drivers
such as warming or changes in nutrient availability, but
these processes will become increasingly important with
ongoing ocean acidification. Our study highlights that
CO
2
, especially due to its cascading effects on light and
nutrient limitation and grazing, can change regional phy-
toplankton community compositions and primary produc-
tion considerably in a simulation with future atmospheric
CO
2
concentrations. Modifications in biomass concentra-
tions can locally account for 10%and more. In addition,
other studies reveal that CO
2
can tip the scale when mul-
tiple driver interactions are involved (e.g., Harvey et al.,
2013; Brandenburg et al., 2019; Seifert et al., 2020).
Accounting for carbonation and acidification effects on
phytoplankton growth and calcification in model projec-
tions adds an important physiological constraint on pri-
mary production and biogeochemical cycling in the future
ocean.
Data accessibility statement
Model data necessary to produce the findings of this study
are archived in Zenodo (https://zenodo.org/; Seifert and
Hauck, 2022). Full model output can be requested from
the corresponding author.
Supplemental files
The supplemental files for this article can be found as
follows:
Figures S1–S6.pdf
Text S1.pdf
Tables S1–S4.pdf
Acknowledgments
We thank Meike Vogt from ETH Zurich, Switzerland, and
Colleen O’Brien for providing gridded data products of
coccolithophore biomass concentrations for our model
evaluation. Furthermore, we thank Fre
´de
´ric Maps and an
anonymous reviewer for their constructive comments and
suggestions.
Funding
This research was supported under the Initiative and Net-
working Fund of the Helmholtz Association (Helmholtz
Young Investigator Group Marine Carbon and Ecosystem
Feedbacks in the Earth System [MarESys], grant number
VH-NG-1301) as well as the European Union’s Horizon
2020 research and innovation program under grant agree-
ment number 869357 (project OceanNETs: Ocean-based
Negative Emission Technologies—analyzing the feasibility,
risks, and co-benefits of ocean-based negative emission
technologies for stabilizing the climate). MS acknowledges
funding from the Helmholtz Graduate School for Polar
and Marine Research POLMAR. CN has received funding
from the European Union’s Horizon 2020 research and
innovation program under grant agreement number
820989 (project COMFORT). The work reflects only the
authors’ views; the European Commission and their exec-
utive agency are not responsible for any use that may be
made of the information the work contains.
Competing interests
The authors have no competing interests, as defined by
Elementa, that might be perceived to influence the
research presented in this article.
Author contributions
Contributed to conception and design: MS, CN, BR, JH.
Contributed to acquisition of data: MS, CN, BR, JH.
Contributed to analysis and interpretation of data: MS,
CN, BR, JH.
Drafted and/or revised the article: MS, CN, BR, JH.
Approved the submitted version for publication: MS,
CN, BR, JH.
References
Alvarez-Fernandez, S,Bach, LT,Taucher, J,Riebesell, U,
Sommer, U,Aberle, N,Brussaard, CPD,Boersma,
M. 2018. Plankton responses to ocean acidification:
The role of nutrient limitation. Progress in Oceanog-
raphy 165: 11–18. DOI: https//dx.doi.org/10.1016/
j.pocean.2018.04.006.
Aumont, O,Ethe
´,C,Tagliabue, A,Bopp, L,Gehlen, M.
2015. PISCES-v2: An ocean biogeochemical model
for carbon and ecosystem studies. Geoscientific
Model Development 8(8): 2465–2513. DOI: http://
dx.doi.org/10.5194/gmd-8-2465-2015.
Aumont, O,Maier-Reimer, E,Blain, S,Monfray, P.
2003. An ecosystem model of the global ocean
including Fe, Si, P colimitations. Global Biogeochem-
ical Cycles 17(2): 1–26. DOI: http://dx.doi.org/10.
1029/2001GB001745.
Bach, LT. 2015. Reconsidering the role of carbonate ion
concentration in calcification by marine organisms.
Biogeosciences 12(16): 4939–4951. DOI: http://dx.
doi.org/10.5194/bg-12-4939-2015.
Bach, LT,Herna
´ndez-Herna
´ndez, N,Taucher, J,Spisla,
C,Sforna, C,Riebesell, U,Arı
´stegui, J. 2019.
Effects of elevated CO
2
on a natural diatom commu-
nity in the subtropical NE Atlantic. Frontiers in
Art. 10(1) page 22 of 32 Seifert et al: CO
2
effects on phytoplankton growth

Marine Science 6: 75. DOI: http://dx.doi.org/10.
3389/fmars.2019.00075.
Bach, LT,Mackinder, LCM,Schulz, KG,Wheeler, G,
Schroeder, DC,Brownlee, C,Riebesell, U. 2013.
Dissecting the impact of CO
2
and pH on the
mechanisms of photosynthesis and calcification in
the coccolithophore Emiliania huxleyi.New Phytolo-
gist 199(1): 121–134. DOI: http://dx.doi.org/10.
1111/nph.12225.
Bach, LT,Riebesell, U,Gutowska, MA,Federwisch, L,
Schulz, KG. 2015. A unifying concept of coccolitho-
phore sensitivity to changing carbonate chemistry
embedded in an ecological framework. Progress in
Oceanography 135: 125–138. DOI: http://dx.doi.
org/10.1016/j.pocean.2015.04.012.
Bach, LT,Riebesell, U,Schulz, KG. 2011. Distinguishing
between the effects of ocean acidification and ocean
carbonation in the coccolithophore Emiliania hux-
leyi.Limnology and Oceanography 56(6):
2040–2050. DOI: http://dx.doi.org/10.4319/lo.
2011.56.6.2040.
Balch, WM,Bowler, BC,Drapeau, DT,Lubelczyk, LC,
Lyczkowski, E,Mitchell, C,Wyeth, A. 2019. Coc-
colithophore distributions of the North and South
Atlantic Ocean. Deep Sea Research Part I: Oceano-
graphic Research Papers 151: 103066. DOI: http://
dx.doi.org/10.1016/j.dsr.2019.06.012.
Balch, WM,Drapeau, DT,Bowler, BC,Lyczskowski, E,
Booth, ES,Alley, D. 2011. The contribution of coc-
colithophores to the optical and inorganic carbon
budgets during the Southern Ocean Gas Exchange
Experiment: New evidence in support of the “Great
Calcite Belt” hypothesis. Journal of Geophysical
Research: Oceans 116(C4): 1–14. DOI: http://dx.
doi.org/10.1029/2011JC006941.
Barcelos e Ramos, J,Schulz, KG,Brownlee, C,Sett, S,
Azevedo, EB. 2014. Effects of increasing seawater
carbon dioxide concentrations on chain formation
of the diatom Asterionellopsis glacialis.PLoS One
9(3): 1–8. DOI: http://dx.doi.org/10.1371/journal.
pone.0090749.
Barrett, PM,Resing, JA,Buck, NJ,Feely, RA,Bullister,
JL,Buck, CS,Landing, WM. 2014. Calcium carbon-
ate dissolution in the upper 1000 m of the eastern
North Atlantic. Global Biogeochemical Cycles 28(4):
386–397. DOI: http://dx.doi.org/10.1002/
2013GB004619.
Barton, AD,Irwin, AJ,Finkel, ZV,Stock, CA. 2016.
Anthropogenic climate change drives shift and shuf-
fle in North Atlantic phytoplankton communities.
Proceedings of the National Academy of Sciences
113(11): 2964–2969. DOI: http://dx.doi.org/10.
1073/pnas.1519080113.
Battaglia, G,Steinacher, M,Joos, F. 2016. A probabilistic
assessment of calcium carbonate export and disso-
lution in the modern ocean. Biogeosciences 13(9):
2823–2848. DOI: http://dx.doi.org/10.5194/bg-
13-2823-2016.
Beardall, J,Raven, JA. 2013. Calcification and ocean acid-
ification: New insights from the coccolithophore
Emiliania huxleyi.New Phytologist 199(1): 1–3. DOI:
http://dx.doi.org/10.1111/nph.12297.
Beaufort, L,Probert, I,de Garidel-Thoron, T,Bendif,
EM,Ruiz-Pino, D,Metzl, N,Goyet, C,Buchet, N,
Coupel, P,Grelaud, M,Rost, B,Rickaby, REM,de
Vargas, C. 2011. Sensitivity of coccolithophores to
carbonate chemistry and ocean acidification. Nature
476(7358): 80–83. DOI: http://dx.doi.org/10.
1038/nature10295.
Beaugrand, G,McQuatters-Gollop, A,Edwards, M,
Goberville, E. 2013. Long-term responses of North
Atlantic calcifying plankton to climate change.
Nature Climate Change 3(3): 263–267. DOI:
http://dx.doi.org/10.1038/nclimate1753.
Behrenfeld, MJ,Falkowski, PG. 1997. Photosynthetic
rates derived from satellite-based chlorophyll con-
centration. Limnology and Oceanography 42(1):
1–20. DOI: http://dx.doi.org/10.4319/lo.1997.42.1.
0001.
Berelson, WM,Balch, WM,Najjar, R,Feely, RA,Sabine,
C,Lee, K. 2007. Relating estimates of CaCO
3
pro-
duction, export, and dissolution in the water col-
umn to measurements of CaCO
3
rain into
sediment traps and dissolution on the sea floor: A
revised global carbonate budget. Global Biogeochem-
ical Cycles 21(1): 1–15. DOI: http://dx.doi.org/10.
1029/2006GB002803.
Bindoff, NL,Cheung, WWL,Kairo, JG,Arı
´stegui, J,
Guinder, VA,Hallberg, R,Hilmi, N,Jiao, N,
Karim, NS,Levin, L,O’Donoghue, S,Purca Cuica-
pusa, SR,Rinkevich, B,Suga, T,Tagliabue, A,Wil-
liamson, P. 2019. Changing ocean, marine
ecosystems, and dependent communities, in Po
¨rt-
ner, HO, Roberts, DC, Masson-Delmotte, V, Zhai, P,
Tignor, M, Poloczanska, E, Mintenbeck, K, Alegrı
´a, A,
Nicolai, M, Okem, A, Petzold, J, Rama, B, Weyer, NM
eds., IPCC special report on the ocean and cryosphere
in a changing climate. New York, NY: Cambridge
University Press: 447–587. DOI: http://dx.doi.org/
10.1017/9781009157964.007.
Boatman, TG,Davey, PA,Lawson, T,Geider, RJ. 2018.
CO
2
modulation of the rates of photosynthesis and
light-dependent O2consumption in Trichodesmium.
Journal of Experimental Botany 70(2): 589–597.
DOI: http://dx.doi.org/10.1093/jxb/ery368.
Boye, M,van den Berg, CMG,de Jong, JTM,Leach, H,
Croot, P,de Baar, HJW. 2001. Organic complexa-
tion of iron in the Southern Ocean. Deep Sea
Research Part I: Oceanographic Research Papers
48(6): 1477–1497. DOI: http://dx.doi.org/10.
1016/S0967-0637(00)00099-6.
Brandenburg, KM,Velthuis, M,Van de Waal, DB. 2019.
Meta-analysis reveals enhanced growth of marine
harmful algae from temperate regions with warm-
ing and elevated CO
2
levels. Global Change Biology
25: 2607–2618. DOI: http://dx.doi.org/10.1111/
gcb.14678.
Brennan, GL,Colegrave, N,Collins, S. 2017. Evolution-
ary consequences of multidriver environmental
change in an aquatic primary producer. Proceedings
Seifert et al: CO
2
effects on phytoplankton growth Art. 10(1) page 23 of 32

of the National Academy of Sciences 114(37):
9930–9935. DOI: http://dx.doi.org/10.1073/pnas.
1703375114.
Brownlee, C,Taylor, A. 2004. Chapter, Calcification in
coccolithophores: A cellular perspective, in Thier-
stein, HR, Young, JR eds., Coccolithophores. Berlin,
Germany: Springer. DOI: http://dx.doi.org/10.
1007/978-3-662-06278-4_2.
Buitenhuis, ET,Hashioka, T,Que
´re
´, CL. 2013. Combined
constraints on global ocean primary production
using observations and models. Global Biogeochem-
ical Cycles 27(3): 847–858. DOI: http://dx.doi.org/
10.1002/gbc.20074.
Buitenhuis, ET,Le Que
´re
´,C,Aumont, O,Beaugrand, G,
Bunker, A,Hirst, A,Ikeda, T,O’Brien, T,Piont-
kovski, S,Straile, D. 2006. Biogeochemical fluxes
through mesozooplankton. Global Biogeochemical
Cycles 20(2): 1–18. DOI: http://dx.doi.org/10.
1029/2005GB002511.
Buitenhuis, ET,Vogt, M,Moriarty, R,Bednars
ˇek, N,
Doney, SC,Leblanc, K,Le Que
´re
´,C,Luo, YW,
O’Brien, C,O’Brien, T,Peloquin, J,Schiebel R,
Swan, C. 2013. MAREDAT: Towards a world atlas
of MARine Ecosystem DATa. Earth System Science
Data 5(2): 227–239. DOI: http://dx.doi.org/10.
5194/essd-5-227-2013.
Carr, ME,Friedrichs, MA,Schmeltz, M,Noguchi Aita,
M,Antoine, D,Arrigo, KR,Asanuma, I,Aumont,
O,Barber, R,Behrenfeld, M,Bidigare, R,Buiten-
huis, ET,Campbell, J,Ciotti, A,Dierssen, H,Dow-
ell, M,Dunne, J,Esaias, W,Gentili, B,Gregg, W,
Groom, S,Hoepffner, N,Ishizaka, J,Kameda, T,
Que
´re
´,CL,Lohrenz, S,Marra, J,Me
´lin, F,Moore,
K,Morel, A,Reddy, TE,Ryan, J,Scardi, M,Smyth,
T,Turpie, K,Tilstone, G,Waters, K,Yamanaka, Y.
2006. A comparison of global estimates of marine
primary production from ocean color. Deep Sea
Research Part II: Topical Studies in Oceanography
53(5): 741–770. DOI: http://dx.doi.org/10.1016/j.
dsr2.2006.01.028.
Constable, AJ,Melbourne-Thomas, J,Corney, SP,
Arrigo, KR,Barbraud, C,Barnes, DKA,Bindoff,
NL,Boyd, PW,Brandt, A,Costa, DP,Davidson, AT,
Ducklow, HW,Emmerson, L,Fukuchi, M,Gutt, J,
Hindell, MA,Hofmann, EE,Hosie, GW,Iida, T,
Jacob, S,Johnston, NM,Kawaguchi, S,Kokubun,
N,Koubbi, P,Lea, M-A,Makhado, A,Massom, RA,
Meiners, K,Meredith, MP,Murphy, EJ,Nicol, S,
Reid, K,Richerson, K,Riddle, MJ,Rintoul, SR,
Smith, WO Jr,Southwell, C,Stark, JS,Sumner,
M,Swadling, KM,Takahashi, KT,Trathan, PN,
Welsford, DC,Weimerskirch, H,Westwood, KJ,
Wienecke, BC,Wolf-Gladrow, D,Wright, SW,
Xavier, JC,Ziegler, P. 2014. Climate change and
Southern Ocean ecosystems I: How changes in phys-
ical habitats directly affect marine biota. Global
Change Biology 20(10): 3004–3025. DOI: http://
dx.doi.org/10.1111/gcb.12623.
Cram, JA,Weber, T,Leung, SW,McDonnell, AMP,Liang,
JH,Deutsch, C. 2018. The role of particle size,
ballast, temperature, and oxygen in the sinking flux
to the deep sea. Global Biogeochemical Cycles 32(5):
858–876. DOI: http://dx.doi.org/10.1029/
2017GB005710.
Cripps, G,Lindeque, P,Flynn, KJ. 2014. Have we been
underestimating the effects of ocean acidification in
zooplankton? Global Change Biology 20(11):
3377–3385. DOI: http://dx.doi.org/10.1111/gcb.
12582.
D’Amario, B,Pe
´rez, C,Grelaud, M,Pitta, P,Krasako-
poulou, E,Ziveri, P. 2020. Coccolithophore com-
munity response to ocean acidification and
warming in the Eastern Mediterranean Sea: Results
from a mesocosm experiment. Scientific Reports
10(1): 1–14. DOI: http://dx.doi.org/10.1038/
s41598-020-69519-5.
Daufresne, M,Lengfellner, K,Sommer, U. 2009. Global
warming benefits the small in aquatic ecosystems.
Proceedings of the National Academy of Sciences
106(31): 12788–12793. DOI: http://dx.doi.org/10.
1073/pnas.0902080106.
de Baar, HJW,de Jong, JTM,Nolting, RF,Timmermans,
KR,van Leeuwe, MA,Bathmann, U,Rutgers van
der Loeff, M,Sildam, J. 1999. Low dissolved Fe and
the absence of diatom blooms in remote Pacific
waters of the Southern Ocean. Marine Chemistry
66(1): 1–34. DOI: http://dx.doi.org/10.1016/
S0304-4203(99)00022-5.
Doney, SC,Fabry, VJ,Feely, RA,Kleypas, JA. 2009.
Ocean acidification: The other CO
2
problem. Annual
Review of Marine Science 1(1): 169–192. DOI:
http://dx.doi.org/10.1146/annurev.marine.010908.
163834.
Dunne, JP,Sarmiento, JL,Gnanadesikan, A.2007.A
synthesis of global particle export from the surface
ocean and cycling through the ocean interior and on
the seafloor. Global Biogeochemical Cycles 21(4):
1–16. DOI: http://dx.doi.org/10.1029/2006GB
002907.
Dutkiewicz, S,Boyd, PW,Riebesell, U. 2021. Exploring
biogeochemical and ecological redundancy in phy-
toplankton communities in the global ocean. Global
Change Biology 27(6): 1196–1213. DOI: http://dx.
doi.org/10.1111/gcb.15493.
Dutkiewicz, S,Morris, JJ,Follows, MJ,Scott, J,Levitan,
O,Dyhrman, ST,Berman-Frank, I. 2015. Impact of
ocean acidification on the structure of future phy-
toplankton communities. Nature Climate Change
5(11): 1002–1006. DOI: http://dx.doi.org/10.
1038/NCLIMATE2722.
Eichner, M,Kranz, SA,Rost, B. 2014. Combined effects
of different CO
2
levels and N sources on the diazo-
trophic cyanobacterium. Trichodesmium. Physiologia
Plantarum 152(2): 316–330. DOI: http://dx.doi.
org/10.1111/ppl.12172.
Eyring, V,Bony, S,Meehl, GA,Senior, CA,Stevens, B,
Stouffer, RJ,Taylor, KE. 2016. Overview of the Cou-
pled Model Intercomparison Project Phase 6
(CMIP6) experimental design and organization.
Art. 10(1) page 24 of 32 Seifert et al: CO
2
effects on phytoplankton growth

Geoscientific Model Development 9(5): 1937–1958.
DOI: http://dx.doi.org/10.5194/gmd-9-1937-2016.
Fasham, MJR,Ducklow, HW,McKelvie, SM. 1990. A
nitrogen-based model of plankton dynamics in the
oceanic mixed layer. Journal of Marine Research
48(3): 591–639. DOI: http://dx.doi.org/10.1357/
002224090784984678.
Fay, AR,McKinley, GA. 2014. Global open-ocean biomes:
Mean and temporal variability. Earth System Science
Data 6(2): 273–284. DOI: http://dx.doi.org/10.
5194/essd-6-273-2014.
Feely, RA,Sabine, CL,Lee, K,Berelson, W,Kleypas, J,
Fabry, VJ,Millero, FJ. 2004. Impact of anthropo-
genic CO
2
on the CaCO
3
system in the oceans. Sci-
ence 305(5682): 362–366. DOI: http://dx.doi.org/
10.1126/science.1097329.
Feng, Y,Hare, CE,Leblanc, K,Rose, JM,Zhang, Y,DiTul-
lio, GR,Lee, PA,Wilhelm, SW,Rowe, JM,Sun, J,
Nemcek, N,Gueguen, C,Passow, U,Benner, I,
Brown, C,Hutchins, DA. 2009. Effects of increased
pCO
2
and temperature on the North Atlantic spring
bloom. I. The phytoplankton community and bio-
geochemical response. Marine Ecology Progress
Series 388:13–25.DOI:http://dx.doi.org/10.
3354/meps08133.
Feng, Y,Roleda, MY,Armstrong, E,Boyd, PW,Hurd, CL.
2017. Environmental controls on the growth, photo-
synthetic and calcification rates of a Southern Hemi-
sphere strain of the coccolithophore. Emiliania
huxleyi. Limnology and Oceanography 62(2):
519–540. DOI: http://dx.doi.org/10.1002/lno.
10442.
Field, CB,Behrenfeld, MJ,Randerson, JT,Falkowski, P.
1998. Primary production of the biosphere: Integrat-
ing terrestrial and oceanic components. Science
281(5374): 237–240. DOI: http://dx.doi.org/10.
1126/science.281.5374.237.
Fielding, SR.2013.Emiliania huxleyi specific growth rate
dependence on temperature. Limnology and Ocean-
ography 58(2): 663–666. DOI: http://dx.doi.org/10.
4319/lo.2013.58.2.0663.
Findlay, HS,Calosi, P,Crawfurd, K. 2011. Determinants
of the PIC:POC response in the coccolithophore Emi-
liania huxleyi under future ocean acidification sce-
narios. Limnology and Oceanography 56(3):
1168–1178. DOI: http://dx.doi.org/10.4319/lo.
2011.56.3.1168.
Finkel, ZV,Beardall, J,Flynn, KJ,Quigg, A,Rees, TAV,
Raven, JA. 2009. Phytoplankton in a changing
world: Cell size and elemental stoichiometry. Journal
of Plankton Research 32(1): 119–137. DOI: http://
dx.doi.org/10.1093/plankt/fbp098.
Follows, MJ,Dutkiewicz, S,Grant, S,Chisholm, SW.
2007. Emergent biogeography of microbial commu-
nities in a model ocean. Science 315(5820):
1843–1846. DOI: http://dx.doi.org/10.1126/
science.1138544.
Friedlingstein, P,O’Sullivan, M,Jones, MW,Andrew,
RM,Hauck, J,Olsen, A,Peters, GP,Peters, W,
Pongratz, J,Sitch, S,Que
´re
´,CL,Canadell, JG,
Ciais, P,Jackson, RB,Alin, S,Araga
˜o, LEOC,
Arneth, A,Arora, V,Bates, NR,Becker, M,
Benoit-Cattin, A,Bittig, HC,Bopp, L,Bultan, S,
Chandra, N,Chevallier, F,Chini, LP,Evans, W,
Florentie, L,Forster, PM,Gasser, T,Gehlen, M,
Gilfillan, D,Gkritzalis, T,Gregor, L,Gruber, N,
Harris, I,Hartung, K,Haverd, V,Houghton, RA,
Ilyina, T,Jain, AK,Joetzjer, E,Kadono, K,Kato, E,
Kitidis, V,Korsbakken, JI,Landschu
¨tzer, P,
Lefe
`vre, N,Lenton, A,Lienert, S,Liu, Z,Lombar-
dozzi, D,Marland, G,Metzl, N,Munro, DR,
Nabel, JEMS,Nakaoka, S-I,Niwa, Y,O’Brien, K,
Ono, T,Palmer, PI,Pierrot, D,Poulter, B,
Resplandy, L,Robertson, E,Ro
¨denbeck, C,
Schwinger, J,Se
´fe
´rian, R,Skjelvan, I,Smith, AJP,
Sutton, AJ,Tanhua, T,Tans, PP,Tian, H,Tilbrook,
B,van der Werf, G,Vuichard, N,Walker, AP,Wan-
ninkhof, R,Watson, AJ,Willis, D,Wiltshire, AJ,
Yuan, W,Yue, X,Zaehle, S. 2020. Global carbon
budget 2020. Earth System Science Data 12(4):
3269–3340. DOI: http://dx.doi.org/10.5194/essd-
12-3269-2020.
Gafar, NA,Eyre, BD,Schulz, KG. 2018. A conceptual
model for projecting coccolithophorid growth, cal-
cification and photosynthetic carbon fixation rates
in response to global ocean change. Frontiers in
Marine Science 4(433): 1–18. DOI: http://dx.doi.
org/10.3389/fmars.2017.00433.
Gangstø, R,Gehlen, M,Schneider, B,Bopp, L,Aumont,
O,Joos, F. 2008. Modeling the marine aragonite
cycle: Changes under rising carbon dioxide and its
role in shallow water CaCO
3
dissolution. Biogeos-
ciences 5(4): 1057–1072. DOI: http://dx.doi.org/
10.5194/bg-5-1057-2008.
Gangstø, R,Joos, F,Gehlen, M. 2011. Sensitivity of
pelagic calcification to ocean acidification. Biogeos-
ciences 8(2): 433–458. DOI: http://dx.doi.org/10.
5194/bg-8-433-2011.
Gao, K,Beardall, J,Ha
¨der, DP,Hall-Spencer, JM,Gao,
G,Hutchins, DA. 2019. Effects of ocean acidifica-
tion on marine photosynthetic organisms under the
concurrent influences of warming, UV radiation and
deoxygenation. Frontiers in Marine Science 6(322):
1–18. DOI: http://dx.doi.org/10.3389/fmars.2019.
00322.
Gao, K,Zhang, Y,Ha
¨der, DP. 2018. Individual and inter-
active effects of ocean acidification, global warming,
and UV radiation on phytoplankton. Journal of
Applied Phycology 30(2): 743–759. DOI: http://dx.
doi.org/10.1007/s10811-017-1329-6.
Garcia, HE,Boyer, TP,Locarnini, RA,Antonov, JI,Mis-
honov, AV,Baranova, OK,Zweng, MM,Reagan,
JR,Johnson, DR,Levitus, S. 2013. World Ocean
Atlas 2013. Volume 3, Dissolved oxygen, apparent
oxygen utilization, and oxygen saturation.Silver
Spring, MD: National Oceanographic Data Center,
Ocean Climate Laboratory. DOI: http://dx.doi.org/
10.7289/V5XG9P2W.
Garcia, NS,Fu, FX,Breene, CL,Yu, EK,Bernhardt, PW,
Mulholland, MR,Hutchins, DA. 2013. Combined
Seifert et al: CO
2
effects on phytoplankton growth Art. 10(1) page 25 of 32

effects of CO
2
and light on large and small isolates
of the unicellular N
2
-fixing cyanobacterium Croco-
sphaera watsonii from the western tropical Atlantic
Ocean. European Journal of Phycology 48(1):
128–139. DOI: http://dx.doi.org/10.1080/09670
262.2013.773383.
Gattuso, JP,Epitalon, JM,Lavigne, H,Orr, J. 2019. sea-
carb: Seawater carbonate chemistry. R-package ver-
sion 3.2.12. Available at http://CRAN.R-project.org/
package¼seacarb. Accessed 11 February 2021.
Gehlen, M,Bopp, L,Emprin, N,Aumont, O,Heinze, C,
Ragueneau, O. 2006. Reconciling surface ocean
productivity, export fluxes and sediment composi-
tion in a global biogeochemical ocean model. Bio-
geosciences 3(4): 521–537. DOI: http://dx.doi.org/
10.5194/bg-3-521-2006.
Geider, RJ,Maclntyre, HL,Kana, TM. 1998. A dynamic
regulatory model of phytoplanktonic acclimation to
light, nutrients, and temperature. Limnology and
Oceanography 43(4): 679–694. DOI: http://dx.doi.
org/10.4319/lo.1998.43.4.0679.
Godrijan, J,Drapeau, D,Balch, WM. 2020. Mixotrophic
uptake of organic compounds by coccolithophores.
Limnology and Oceanography 65(6): 1410–1421.
DOI: http://dx.doi.org/10.1002/lno.11396.
Gregg, WW,Casey, NW. 2007. Modeling coccolitho-
phores in the global oceans. Deep Sea Research Part
II: Topical Studies in Oceanography 54(5): 447–477.
DOI: http://dx.doi.org/10.1016/j.dsr2.2006.12.007.
Gruber, N. 2011. Warming up, turning sour, losing breath:
Ocean biogeochemistry under global change. Philo-
sophical Transactions of the Royal Society A: Mathe-
matical, Physical and Engineering Sciences
369(1943): 1980–1996. DOI: http://dx.doi.org/10.
1098/rsta.2011.0003.
Haidar, AT,Thierstein, HR. 2001. Coccolithophore
dynamics off Bermuda (N. Atlantic). Deep Sea
Research Part II: Topical Studies in Oceanography
48(8): 1925–1956. DOI: http://dx.doi.org/10.
1016/S0967-0645(00)00169-7.
Hare, CE,Leblanc, K,DiTullio, GR,Kudela, RM,Zhang,
Y,Lee, PA,Riseman, S,Hutchins, DA. 2007. Con-
sequences of increased temperature and CO
2
for
phytoplankton community structure in the Bering
Sea. Marine Ecology Progress Series 352: 9–16. DOI:
http://dx.doi.org/10.3354/meps07182.
Harvey, BP,Gwynn-Jones, D,Moore, PJ. 2013. Meta-
analysis reveals complex marine biological responses
to the interactive effects of ocean acidification and
warming. Ecology and Evolution 3(4): 1016–1030.
DOI: http://dx.doi.org/10.1002/ece3.516.
Hauck, J,Vo
¨lker, C,Wang, T,Hoppema, M,Losch, M,
Wolf-Gladrow, DA. 2013. Seasonally different car-
bon flux changes in the Southern Ocean in response
to the southern annular mode. Global Biogeochemi-
cal Cycles 27(4): 1236–1245. DOI: http://dx.doi.
org/10.1002/2013GB004600.
Hauck, J,Vo
¨lker, C,Wolf-Gladrow, DA,Laufko
¨tter, C,
Vogt, M,Aumont, O,Bopp, L,Buitenhuis, ET,
Doney, SC,Dunne, J,Gruber, N,Hashioka, T,
John, J,Que
´re
´,CL,Lima, ID,Nakano, H,Se
´fe
´rian,
R,Totterdell, I. 2015. On the Southern Ocean CO
2
uptake and the role of the biological carbon pump
in the 21st century. Global Biogeochemical Cycles
29(9): 1451–1470. DOI: http://dx.doi.org/10.
1002/2015GB005140.
Haunost, M,Riebesell, U,D’Amore, F,Kelting, O,Bach,
LT. 2021. Influence of the calcium carbonate shell of
coccolithophores on ingestion and growth of a dino-
flagellate predator. Frontiers in Marine Science 8:
738. DOI: http://dx.doi.org/10.3389/fmars.2021.
664269.
Hennon, GMM,Quay, P,Morales, RL,Swanson, LM,
Virginia Armbrust, E. 2014. Acclimation condi-
tions modify physiological response of the diatom
Thalassiosira pseudonana to elevated CO
2
concentra-
tions in a nitrate-limited chemostat. Journal of Phy-
cology 50(2): 243–253. DOI: http://dx.doi.org/10.
1111/jpy.12156.
Henson, SA,Sanders, R,Madsen, E,Morris, PJ,Le
Moigne, F,Quartly, GD. 2011. A reduced estimate
of the strength of the ocean’s biological carbon
pump. Geophysical Research Letters 38(4): 1–5. DOI:
http://dx.doi.org/10.1029/2011GL046735.
Hermoso, M. 2015. Control of ambient pH on growth and
stable isotopes in phytoplanktonic calcifying algae.
Paleoceanography 30(8): 1100–1112. DOI: http://
dx.doi.org/10.1002/2015PA002844.
Holt, J,Allen, JI,Anderson, TR,Brewin, R,Butenscho
¨n,
M,Harle, J,Huse, G,Lehodey, P,Lindemann, C,
Memery, L,Salihoglu, B, Senina, I, Yool, A.2014.
Challenges in integrative approaches to modelling
the marine ecosystems of the North Atlantic: Physics
to fish and coasts to ocean. Progress in Oceanogra-
phy 129: 285–313. DOI: http://dx.doi.org/10.1016/
j.pocean.2014.04.024.
Hopkins, J,Balch, WM. 2018. A new approach to esti-
mating coccolithophore calcification rates from
space. Journal of Geophysical Research: Biogeos-
ciences 123(5): 1447–1459. DOI: http://dx.doi.
org/10.1002/2017JG004235.
Hoppe, CJM,Flintrop, CM,Rost, B.2018.TheArctic
picoeukaryote Micromonas pusilla benefits synergis-
tically from warming and ocean acidification. Bio-
geosciences 15(14): 4353–4365. DOI: http://dx.doi.
org/10.5194/bg-15-4353-2018.
Hoppe, CJM,Langer, G,Rost, B. 2011. Emiliania huxleyi
shows identical responses to elevated pCO
2
in TA
and DIC manipulations. Journal of Experimental
Marine Biology and Ecology 406(1–2): 54–62. DOI:
http://dx.doi.org/10.1016/j.jembe.2011.06.008.
Ihnken, S,Roberts, S,Beardall, J. 2011. Differential
responses of growth and photosynthesis in the
marine diatom Chaetoceros muelleri to CO
2
and light
availability. Phycologia 50(2): 182–193. DOI: http://
dx.doi.org/10.2216/10-11.1.
Irion, S,Christaki, U,Berthelot, H,L’helguen, S,Jar-
dillier, L. 2021. Small phytoplankton contribute
greatly to CO
2
-fixation after the diatom bloom in
Art. 10(1) page 26 of 32 Seifert et al: CO
2
effects on phytoplankton growth

the Southern Ocean. The ISME Journal: 1–14. DOI:
http://dx.doi.org/10.1038/s41396-021-00915-z.
Irwin, AJ,Finkel, ZV,Schofield, OME,Falkowski, PG.
2006. Scaling-up from nutrient physiology to the
size-structure of phytoplankton communities. Jour-
nal of Plankton Research 28(5): 459–471. DOI:
http://dx.doi.org/10.1093/plankt/fbi148.
Jansen, H,Wolf-Gladrow, DA. 2001. Carbonate dissolu-
tion in copepod guts: A numerical model. Marine
Ecology Progress Series 221: 199–207. DOI: http://
dx.doi.org/10.3354/meps221199.
Jin, X,Gruber, N,Dunne, JP,Sarmiento, JL,Armstrong,
RA. 2006. Diagnosing the contribution of phyto-
plankton functional groups to the production and
export of particulate organic carbon, CaCO
3
,and
opal from global nutrient and alkalinity distribu-
tions. Global Biogeochemical Cycles 20(2): 1–17.
DOI: http://dx.doi.org/10.1029/2005GB002532.
Karakus¸, O,Vo
¨lker, C,Iversen, M,Hagen, W,Gladrow,
DW,Fach, B,Hauck, J. 2021. Modeling the impact
of macrozooplankton on carbon export production
in the Southern Ocean. Journal of Geophysical
Research: Oceans 126(12): e2021JC017315. DOI:
http://dx.doi.org/10.1029/2021JC017315.
Kelly, MW,Griffiths, JS. 2021. Selection experiments in
the sea: What can experimental evolution tell us
about how marine life will respond to climate
change? The Biological Bulletin 241: 30–42. DOI:
http://dx.doi.org/10.1086/715109.
Kim, H,Spivack, AJ,Menden-Deuer, S. 2013. pH alters
the swimming behaviors of the raphidophyte Het-
erosigma akashiwo: Implications for bloom forma-
tion in an acidified ocean. Harmful Algae 26:
1–11. DOI: http://dx.doi.org/10.1016/j.hal.2013.
03.004.
Klaas, C,Archer, DE.2002.Associationofsinking
organic matter with various types of mineral ballast
in the deep sea: Implications for the rain ratio.
Global Biogeochemical Cycles 16(4): 63-1–63-14.
DOI: http://dx.doi.org/10.1029/2001GB001765.
Kottmeier, DM,Rokitta, SD,Rost, B. 2016. Acidification,
not carbonation, is the major regulator of carbon
fluxes in the coccolithophore. Emiliania huxleyi. New
Phytologist 211: 126–137. DOI: http://dx.doi.org/
10.1111/nph.13885.
Kottmeier, DM,Rokitta, SD,Tortell, PD,Rost, B.2014.
Strong shift from HCO
3to CO
2
uptake in Emiliania
huxleyi with acidification: New approach unravels
acclimation versus short-term pH effects. Photosyn-
thesis Research 121(2–3): 265–275. DOI: http://dx.
doi.org/10.1007/s11120-014-9984-9.
Kranz, SA,Dieter, S,Richter, KU,Rost, B. 2009. Carbon
acquisition by Trichodesmium:TheeffectofpCO
2
and diurnal changes. Limnology and Oceanography
54(2): 548–559. DOI: http://dx.doi.org/10.4319/lo.
2009.54.2.0548.
Krug, S,Schulz, K,Riebesell, U. 2011. Effects of changes
in carbonate chemistry speciation on Coccolithus
braarudii: A discussion of coccolithophorid
sensitivities. Biogeosciences 8: 771–777. DOI:
http://dx.doi.org/10.5194/bg-8-771-2011.
Krumhardt, KM,Lovenduski, NS,Freeman, NM,Bates,
NR.2016.Apparentincreaseincoccolithophore
abundance in the subtropical North Atlantic from
1990 to 2014. Biogeosciences 13(4): 1163–1177.
DOI: http://dx.doi.org/10.5194/bg-13-1163-2016.
Krumhardt, KM,Lovenduski, NS,Iglesias-Rodriguez,
MD,Kleypas, JA. 2017. Coccolithophore growth
and calcification in a changing ocean. Progress in
Oceanography 159: 276–295. DOI: http://dx.doi.
org/10.1016/j.pocean.2017.10.007.
Krumhardt, KM,Lovenduski, NS,Long, MC,Levy, M,
Lindsay, K,Moore, JK,Nissen, C. 2019. Coccolitho-
phore growth and calcification in an acidified ocean:
Insights from Community Earth System Model simu-
lations. Journal of Advances in Modeling Earth Sys-
tems 11(5): 1418–1437. DOI: http://dx.doi.org/10.
1029/2018MS001483.
Kvale, KF,Meissner, KJ,Keller, DP,Eby, M,Schmittner,
A. 2015. Explicit planktic calcifiers in the University
of Victoria Earth System Climate Model, Version 2.9.
Atmosphere-Ocean 53(3): 332–350. DOI: http://dx.
doi.org/10.1080/07055900.2015.1049112.
Kwiatkowski, L,Torres, O,Bopp, L,Aumont, O,Cham-
berlain, M,Christian, JR,Dunne, JP,Gehlen, M,
Ilyina, T,John, JG,Lenton, A,Li, H,Lovenduski,
NS,Orr, JC,Palmieri, J,Santana-Falco
´n, Y,
Schwinger, J,Se
´fe
´rian, R,Stock, CA,Tagliabue,
A,Takano, Y,Tjiputra, J,Toyama, K,Tsujino, H,
Watanabe, M,Yamamoto, A,Yool, A,Ziehn, T.
2020. Twenty-first century ocean warming, acidifica-
tion, deoxygenation, and upper-ocean nutrient and
primary production decline from CMIP6 model pro-
jections. Biogeosciences 17(13): 3439–3470. DOI:
http://dx.doi.org/10.5194/bg-17-3439-2020.
Langer, G,Geisen, M,Baumann, KH,Kla
¨s, J,Riebesell,
U,Thoms, S,Youn g, JR. 2006. Species-specific
responses of calcifying algae to changing seawater
carbonate chemistry. Geochemistry, Geophysics, Geo-
systems 7(9): 1–12. DOI: http://dx.doi.org/10.1029/
2005GC001227.
Langer, G,Nehrke, G,Probert, I,Ly, J,Ziveri, P. 2009.
Strain-specific responses of Emiliania huxleyi to
changing seawater carbonate chemistry. Biogeos-
ciences 6(11): 2637–2646. DOI: http://dx.doi.org/
10.5194/bg-6-2637-2009.
Lauvset, SK,Key, RM,Olsen, A,van Heuven, S,Velo, A,
Lin, X,Schirnick, C,Kozyr, A,Tanhua, T,Hop-
pema, M,Jutterstro
¨m, S,Steinfeldt, R,Jeansson,
E,Ishii, M,Perez, FF,Suzuki, T,Watelet, S.2016.A
new global interior ocean mapped climatology: The
11GLODAP version 2. Earth System Science
Data 8(2): 325–340. DOI: http://dx.doi.org/10.
5194/essd-8-325-2016.
Le Que
´re
´,C,Harrison, SP,Colin Prentice, I,Buiten-
huis, ET,Aumont, O,Bopp, L,Claustre, H,Cotrim
Da Cunha, L,Geider, R,Giraud, X,Klaas, C,Koh-
feld, KE,Legendre, L,Manizza, M,Platt, T,Rivkin,
RB,Sathyendranath, S,Uitz, J,Watson, AJ,Wolf-
Seifert et al: CO
2
effects on phytoplankton growth Art. 10(1) page 27 of 32

Gladrow, D. 2005. Ecosystem dynamics based on
plankton functional types for global ocean biogeo-
chemistry models. Global Change Biology 11:
2016–2040. DOI: http://dx.doi.org/10.1111/j.
1365-2486.2005.1004.x.
Leblanc, K,Arstegui, J,Armand, L,Assmy, P,Beker, B,
Bode, A,Breton, E,Cornet, V,Gibson, J,Gosselin,
M-P,Kopczynska, E,Marshall, H,Peloquin, J,
Piontkovski, S,Poulton, AJ,Que
´guiner, B,Schie-
bel, R,Shipe, R,Stefels, J,van Leeuwe, MA,Var-
ela, M,Widdicombe, C,Yallop, M. 2012. A global
diatom database—Abundance, biovolume and bio-
mass in the world ocean. Earth System Science Data
4(1): 149–165. DOI: http://dx.doi.org/10.5194/
essd-4-149-2012.
Lee, K. 2001. Global net community production estimated
from the annual cycle of surface water total dis-
solved inorganic carbon. Limnology and Oceanogra-
phy 46(6): 1287–1297. DOI: http://dx.doi.org/10.
4319/lo.2001.46.6.1287.
Li, F,Fan, J,Hu, L,Beardall, J,Xu, J. 2019. Physiological
and biochemical responses of Thalassiosira weissflo-
gii (diatom) to seawater acidification and alkaliza-
tion. ICES Journal of Marine Science 76(6):
1850–1859. DOI: http://dx.doi.org/10.1093/
icesjms/fsz028.
Li, W,Gao, K,Beardall, J.2012.Interactiveeffectsof
ocean acidification and nitrogen-limitation on the
diatom Phaeodactylum tricornutum.PLoS One
7(12): e51590. DOI: http://dx.doi.org/10.1371/
journal.pone.0051590.
Li, W,Yang,Y,Li, Z,Xu, J,Gao, K. 2017. Effects of seawater
acidification on the growth rates of the diatom Tha-
lassiosira (Conticribra) weissflogii under different
nutrient, light, and UV radiation regimes. Journal
of Applied Phycology 29(1): 133–142. DOI: http://
dx.doi.org/10.1007/s10811-016-0944-y.
Litchman, E,Klausmeier, CA,Schofield, OM,Falk-
owski, PG. 2007. The role of functional traits and
trade-offs in structuring phytoplankton communi-
ties: Scaling from cellular to ecosystem level. Ecology
Letters 10(12): 1170–1181. DOI: http://dx.doi.org/
10.1111/j.1461-0248.2007.01117.x.
Maat, DS,Crawfurd, KJ,Timmermans, KR,Brussaard,
CPD,Wommack, KE. 2014. Elevated CO
2
and phos-
phate limitation favor Micromonas pusilla through
stimulated growth and reduced viral impact. Applied
and Environmental Microbiology 80(10): 3119–3127.
DOI: http://dx.doi.org/10.1128/AEM.03639-13.
Maran˜o
´n, E,Cermen˜o, P,Lo
´pez-Sandoval, DC,Rodrı
´-
guez-Ramos, T,Sobrino, C,Huete-Ortega, M,
Blanco, JM,Rodrı
´guez, J. 2013. Unimodal size scal-
ing of phytoplankton growth and the size depen-
dence of nutrient uptake and use. Ecology Letters
16(3): 371–379. DOI: http://dx.doi.org/10.1111/
ele.12052.
Margalef, R. 1978. Life-forms of phytoplankton as sur-
vival alternatives in an unstable environment. Ocea-
nologica Acta 1(4): 493–509.
Mayers, KMJ,Poulton, AJ,Bidle, K,Thamatrakoln, K,
Schieler, B,Giering, SLC,Wells, SR,Tarran GA,
Mayor, D,Johnson, M,Riebesell, U,Larsen, A,
Vardi, A,Harvey, EL. 2020. The possession of coc-
coliths fails to deter microzooplankton grazers.
Frontiers in Marine Science 7: 976. DOI: http://dx.
doi.org/10.3389/fmars.2020.569896.
Meyer, J,Riebesell, U. 2015. Reviews and syntheses:
Responses of coccolithophores to ocean acidifica-
tion: A meta-analysis. Biogeosciences 12(6):
1671–1682. DOI: http://dx.doi.org/10.5194/bg-12-
1671-2015.
Middelburg, JJ.2019.Marine carbon biogeochemistry—A
primer for earth system scientists,inLohmann,G,
Mysak, LA, Notholt, J, Rabassa, J, Unnithan, V eds.,
Springer briefs in earth system sciences. Berlin, Ger-
many: Springer Open. DOI: http://dx.doi.org/10.
1007/978-3-030-10822-9.
Milliman, JD,Troy, PJ,Balch, WM,Adams, AK,Li, YH,
Mackenzie, FT. 1999. Biologically mediated disso-
lution of calcium carbonate above the chemical lyso-
cline? Deep Sea Research Part I: Oceanographic
Research Papers 46(10): 1653–1669. DOI: http://
dx.doi.org/10.1016/S0967-0637(99)00034-5.
Monteiro, FM,Bach, LT,Brownlee, C,Bown, P,Rickaby,
REM,Poulton, AJ,Tyrrell, T,Beaufort, L,Dutkie-
wicz, S,Gibbs, S,Gutowska, MA,Lee, R,Riebe-
sell, U,Young, J,Ridgwell, A. 2016. Why marine
phytoplankton calcify. Science Advances 2(7):
e1501822. DOI: http://dx.doi.org/10.1126/sciadv.
1501822.
Moore, JK,Fu, W,Primeau, F,Britten, GL,Lindsay, K,
Long, M,Doney, SC,Mahowald, N,Hoffman, F,
Randerson, JT. 2018. Sustained climate warming
drives declining marine biological productivity. Sci-
ence 359(6380): 1139–1143. DOI: http://dx.doi.
org/10.1126/science.aao6379.
Moriarty, R,O’Brien, TD. 2013. Distribution of mesozoo-
plankton biomass in the global ocean. Earth System
Science Data 5(1): 45–55. DOI: http://dx.doi.org/10.
5194/essd-5-45-2013.
Mucci, A. 1983. The solubility of calcite and aragonite in
seawater at various salinities, temperatures, and one
atmosphere total pressure. American Journal of Sci-
ence 283(7): 780–799. DOI: http://dx.doi.org/10.
2475/ajs.283.7.780.
Mu
¨ller, MN,Trull, TW,Hallegraeff, GM. 2015. Differing
responses of three Southern Ocean Emiliania huxleyi
ecotypes to changing seawater carbonate chemistry.
Marine Ecology Progress Series 531: 81–90. DOI:
http://dx.doi.org/10.3354/meps11309.
Nelson, DM,Tre
´guer, P,Brzezinski, MA,Leynaert, A,
Que
´guiner, B. 1995. Production and dissolution of
biogenic silica in the ocean: Revised global estimates,
comparison with regional data and relationship to
biogenic sedimentation. Global Biogeochemical Cycles
9(3): 359–372. DOI: http://dx.doi.org/10.1029/
95GB01070.
Nielsen, MV. 1997. Growth, dark respiration and photo-
synthetic parameters of the coccolithophorid
Art. 10(1) page 28 of 32 Seifert et al: CO
2
effects on phytoplankton growth

Emiliania huxleyi (Prymnesiophyceae) acclimated to
different day length-irradiance combinations. Jour-
nal of Phycology 33(5): 818–822. DOI: http://dx.doi.
org/10.1111/j.0022-3646.1997.00818.x.
Nissen, C,Vogt, M,Mu
¨nnich, M,Gruber, N,Haumann,
FA. 2018. Factors controlling coccolithophore bioge-
ography in the Southern Ocean. Biogeosciences
15(22): 6997–7024. DOI: http://dx.doi.org/10.
5194/bg-15-6997-2018.
O’Brien, CJ,Peloquin, JA,Vogt, M,Heinle, M,Gruber, N,
Ajani, P,Andruleit, H,Arı
´stegui, J,Beaufort, L,
Estrada, M,Karentz, D,Kopczyn
´ska, E,Lee, R,
Poulton, AJ,Pritchard, T,Widdicombe, C.2013.
Global marine plankton functional type biomass dis-
tributions: Coccolithophores. Earth System Science
Data 5(2): 259–276. DOI: http://dx.doi.org/10.
5194/essd-5-259-2013.
Orr, J,Epitalon, JM. 2015. Improved routines to model
the ocean carbonate system: Mocsy 2.0. Biogeos-
ciences 8: 485–499. DOI: http://dx.doi.org/10.
5194/gmd-8-485-2015.
Paasche, E. 2001. A review of the coccolithophorid Emi-
liania huxleyi (Prymnesiophyceae), with particular
reference to growth, coccolith formation, and
calcification-photosynthesis interactions. Phycologia
40(6): 503–529. DOI: http://dx.doi.org/10.2216/
i0031-8884-40-6-503.1.
Palevsky, HI,Doney, SC. 2018. How choice of depth hori-
zon influences the estimated spatial patterns and
global magnitude of ocean carbon export flux. Geo-
physical Research Letters 45(9): 4171–4179. DOI:
http://dx.doi.org/10.1029/2017GL076498.
Panassa, E,Santana-Casiano, JM, Gonza
´lez-Da
´vila, M,
Hoppema, M, van Heuven, SMAC, Vo
¨lker, C,
Wolf-Gladrow,D,Hauck,J.2018. Variability of
nutrients and carbon dioxide in the Antarctic Inter-
mediate Water between 1990 and 2014. Ocean
Dynamics 68(3): 295–308. DOI: http://dx.doi.org/
10.1007/s10236-018-1131-2.
Panc
ˇic
´,M,Hansen, PJ,Tammilehto, A,Lundholm, N.
2015. Resilience to temperature and pH changes in
a future climate change scenario in six strains of the
polar diatom Fragilariopsis cylindrus.Biogeosciences
12(14): 4235–4244. DOI: http://dx.doi.org/10.
5194/bg-12-4235-2015.
Paul, AJ,Bach, LT. 2020. Universal response pattern of
phytoplankton growth rates to increasing CO
2
.New
Phytologist 228(6): 1710–1716. DOI: http://dx.doi.
org/10.1111/nph.16806.
Peter, KH,Sommer, U. 2012. Phytoplankton cell size:
Intra- and interspecific effects of warming and graz-
ing. PLoS One 7(11): 1–9. DOI: http://dx.doi.org/10.
1371/journal.pone.0049632.
Petrou, K,Baker, KG,Nielsen, DA,Hancock, AM,
Schulz, KG,Davidson, AT. 2019. Acidification
diminishes diatom silica production in the Southern
Ocean. Nature Climate Change 9(10): 781–786. DOI:
http://dx.doi.org/10.1038/s41558-019-0557-y.
Petrou, K,Kranz, SA,Trimborn, S,Hassler, CS,Amei-
jeiras, SB,Sackett, O,Ralph, PJ,Davidson, AT.
2016. Southern Ocean phytoplankton physiology
in a changing climate. Journal of Plant Physiology
203: 135–150. DOI: http://dx.doi.org/10.1016/j.
jplph.2016.05.004.
Pierella Karlusich, JJ,Bowler, C,Biswas, H. 2021. Car-
bon dioxide concentration mechanisms in natural
populations of marine diatoms: Insights from Tara
Oceans. Frontiers in Plant Science 12:659.DOI:
http://dx.doi.org/10.3389/fpls.2021.657821.
Pinkerton, MH,Boyd, PW,Deppeler, S,Hayward, A,
Ho
¨fer, J,Moreau, S. 2021. Evidence for the impact
of climate change on primary producers in the
Southern Ocean. Frontiers in Ecology and Evolution
9: 134. DOI: http://dx.doi.org/10.3389/fevo.2021.
592027.
Poloczanska, ES,Burrows, MT,Brown, CJ,Garca Moli-
nos, J,Halpern, BS,Hoegh-Guldberg, O,Kappel,
CV,Moore, PJ,Richardson, AJ,Schoeman, DS,
Sydeman, WJ. 2016. Responses of marine organ-
isms to climate change across oceans. Frontiers in
Marine Science 3: 62. DOI: http://dx.doi.org/10.
3389/fmars.2016.00062.
Pond, DW,Harris, RP,Brownlee, C. 1995. A microinjec-
tion technique using a pH-sensitive dye to deter-
mine the gut pH of Calanus helgolandicus.Marine
Biology 123(1): 75–79. DOI: http://dx.doi.org/10.
1007/BF00350325.
Poulton, AJ,Adey, TR,Balch, WM,Holligan, PM. 2007.
Relating coccolithophore calcification rates to phy-
toplankton community dynamics: Regional differ-
ences and implications for carbon export. Deep Sea
Research Part II: Topical Studies in Oceanography
54(5): 538–557. DOI: http://dx.doi.org/10.1016/j.
dsr2.2006.12.003.
Raitzsch, M,Gattuso, JP. 2020. ScarFace—Seacarb calcu-
lations with R Shiny user interface. DOI: http://dx.
doi.org/10.5281/zenodo.4153455.
Reinfelder, JR. 2011. Carbon concentrating mechanisms
in eukaryotic marine phytoplankton. Annual Review
of Marine Science 3(1): 291–315. DOI: http://dx.doi.
org/10.1146/annurev-marine-120709-142720.
Riebesell, U,Bach, LT,Bellerby, RG,Monsalve, JRB,
Boxhammer, T,Czerny, J,Larsen, A,Ludwig, A,
Schulz, KG. 2017. Competitive fitness of a predom-
inant pelagic calcifier impaired by ocean acidifica-
tion. Nature Geoscience 10: 19–24. DOI: http://dx.
doi.org/10.1038/NGEO2854.
Riebesell, U,Zondervan, I,Rost, B,Tortell, PD,Zeebe,
RE,Morel, FMM. 2000. Reduced calcification of
marine plankton in response to increased atmo-
spheric CO
2
.Nature 407(6802): 364. DOI: http://
dx.doi.org/10.1038/35030078.
Riegman, R,Stolte W,Noordeloos, AAM,Slezak, D.
2000. Nutrient uptake and alkaline phosphatase
(EC 3:1:3:1) activity of Emiliania huxleyi (Prymnesio-
phycaea) during growth under N and P limitation in
continuous cultures. Journal of Phycology 36:
87–96. DOI: http://dx.doi.org/10.1046/j.1529-
8817.2000.99023.x.
Seifert et al: CO
2
effects on phytoplankton growth Art. 10(1) page 29 of 32

Rigual-Herna
´ndez, AS,Sa
´nchez-Santos, JM,Eriksen,
R,Moy, AD,Sierro, FJ,Flores, JA,Abrantes, F,
Bostock, H,Nodder, SD,Gonza
´lez-Lanchas, A,
Trull, TW. 2020. Limited variability in the phyto-
plankton Emiliania huxleyi since the pre-industrial
era in the Subantarctic Southern Ocean. Anthropo-
cene 31: 100254. DOI: http://dx.doi.org/10.1016/j.
ancene.2020.100254.
Rivero-Calle, S,Gnanadesikan, A,Del Castillo, CE,
Balch, WM,Guikema, SD. 2015. Multidecadal
increase in North Atlantic coccolithophores and the
potential role of rising CO
2
.Science 350(6267):
1533–1537. DOI: http://dx.doi.org/10.1126/
science.aaa8026.
Rost, B,Riebesell, U. 2004. Coccolithophores and the
biological pump: Responses to environmental
changes, in Thierstein, HR, Young, JR eds., Cocco-
lithophores. Berlin, Germany: Springer: 99–125. DOI:
http://dx.doi.org/10.1007/978-3-662-06278-4_5.
Rost, B,Zondervan, I,Wolf-Gladrow, D. 2008. Sensitiv-
ity of phytoplankton to future changes in ocean
carbonate chemistry: Current knowledge, contradic-
tions and research directions. Marine Ecology Prog-
ress Series 373: 227–237. DOI: http://dx.doi.org/10.
3354/meps07776.
Sabine, CL,Key, RM,Feely, RA,Greeley, D. 2002. Inor-
ganic carbon in the Indian Ocean: Distribution and
dissolution processes. Global Biogeochemical Cycles
16(4): 15-1–15-18. DOI: http://dx.doi.org/10.1029/
2002GB001869.
Salle
´e, JB,Pellichero, V,Akhoudas, C,Pauthenet, E,
Vignes, L,Schmidtko, S,Garabato, AN,Suther-
land, P,Kuusela, M. 2021. Summertime increases
in upper-ocean stratification and mixed-layer depth.
Nature 591(7851): 592–598. DOI: http://dx.doi.
org/10.1038/s41586-021-03303-x.
Sarmiento, JL,Dunne, J,Gnanadesikan, A,Key, RM,
Matsumoto, K,Slater, R. 2002. A new estimate
of the CaCO
3
to organic carbon export ratio. Global
Biogeochemical Cycles 16(4): 54-1–54-12. DOI:
http://dx.doi.org/10.1029/2002GB001919.
Sarthou, G,Timmermans, KR,Blain, S,Tre
´guer, P.
2005. Growth physiology and fate of diatoms in the
ocean: A review. Journal of Sea Research 53(1):
25–42. DOI: http://dx.doi.org/10.1016/j.seares.
2004.01.007.
Schlitzer, R. 2000. Applying the adjoint method for bio-
geochemical modeling: Export of particulate organic
matter in the world ocean. Geophysical Monograph—
American Geophysical Union 114: 107–124.
Schlu
¨ter, L,Lohbeck, KT,Gutowska, MA,Gro
¨ger, JP,
Riebesell, U,Reusch, TBH. 2014. Adaptation of
a globally important coccolithophore to ocean
warming and acidification. Nature Climate Change
4(11): 1024–1030. DOI: http://dx.doi.org/10.1038/
NCLIMATE2379.
Schneider, B,Bopp, L,Gehlen, M,Segschneider, J,Fro
¨-
licher, TL,Cadule, P,Friedlingstein, P,Doney, SC,
Behrenfeld, MJ,Joos, F. 2008. Climate-induced
interannual variability of marine primary and export
production in three global coupled climate carbon
cycle models. Biogeosciences 5(2): 597–614. DOI:
http://dx.doi.org/10.5194/bg-5-597-2008.
Schourup-Kristensen, V,Sidorenko, D,Wolf-Gladrow,
DA,Vo
¨lker, C. 2014. A skill assessment of the bio-
geochemical model REcoM2 coupled to the Finite
Element Sea Ice–Ocean Model (FESOM 1.3). Geos-
cientific Model Development 7(6): 2769–2802. DOI:
http://dx.doi.org/10.5194/gmd-7-2769-2014.
Seifert, M,Hauck, J. 2022. Cascading effects augment
the direct impact of CO
2
on phytoplankton growth
in a biogeochemical model, links to model results.
Zenodo. DOI: http://dx.doi.org/10.5281/zenodo.
6674793.
Seifert, M,Rost, B,Trimborn, S,Hauck, J. 2020. Meta-
analysis of multiple driver effects on marine phyto-
plankton highlights modulating role of pCO
2
.Global
Change Biology 26(12): 6787–6804. DOI: http://dx.
doi.org/10.1111/gcb.15341.
Sett, S,Bach, LT,Schulz, KG,Koch-Klavsen, S,Lebrato,
M,Riebesell, U. 2014. Temperature modulates coc-
colithophorid sensitivity of growth, photosynthesis
and calcification to increasing seawater pCO
2
.PLoS
One 9(2): e88308. DOI: http://dx.doi.org/10.1371/
journal.pone.0088308.
Shi, D,Hong, H,Su, X,Liao, L,Chang, S,Lin, W.2019.
The physiological response of marine diatoms to
ocean acidification: differential roles of seawater
pCO
2
and pH. Journal of Phycology 55(3):
521–533. DOI: http://dx.doi.org/10.1111/jpy.
12855.
Shutler, J,Land, PE,Brown, CW,Findlay, HS,Donlon,
CJ,Medland, M,Snooke, R,Blackford, JC.2013.
Coccolithophore surface distributions in the North
Atlantic and their modulation of the air-sea flux of
CO
2
from 10 years of satellite Earth observation
data. Biogeosciences 10(4): 2699–2709. DOI:
http://dx.doi.org/10.5194/bg-10-2699-2013.
Sidorenko, D,Wang, Q,Danilov, S,Schro
¨ter, J. 2011.
FESOM under coordinated ocean-ice reference
experiment forcing. Ocean Dynamics 61(7):
881–890. DOI: http://dx.doi.org/10.1007/s10236-
011-0406-7.
Sinha, B,Buitenhuis, ET,Que
´re
´,CL,Anderson, TR.
2010. Comparison of the emergent behavior of
a complex ecosystem model in two ocean general
circulation models. Progress in Oceanography 84(3):
204–224. DOI: http://dx.doi.org/10.1016/j.pocean.
2009.10.003.
Spungin, D,Berman-Frank, I,Levitan, O.2014.Tricho-
desmium’s strategies to alleviate phosphorus limita-
tion in the future acidified oceans. Environmental
Microbiology 16(6): 1935–1947. DOI: http://dx.doi.
org/10.1111/1462-2920.12424.
Steele, M,Morley, R,Ermold, W. 2001. PHC: A global
ocean hydrography with a high-quality Arctic Ocean.
Journal of Climate 14(9): 2079–2087. DOI: http://
dx.doi.org/10.1175/1520-0442(2001)014<2079:P
AGOHW>2.0.CO;2.
Art. 10(1) page 30 of 32 Seifert et al: CO
2
effects on phytoplankton growth

Stoecker, DK,Hansen, PJ,Caron, DA,Mitra, A.2017.
Mixotrophy in the Marine Plankton. Annual Review
of Marine Science 9(1): 311–335.
Sugie, K,Yoshimura, T. 2013. Effects of pCO
2
and iron on
the elemental composition and cell geometry of the
marine diatom Pseudo-nitzschia pseudodelicatissima
(Bacillariophyceae). Journal of Phycology 49(3):
475–488. DOI: http://dx.doi.org/10.1111/jpy.
12054.
Sugie, K,Yoshimura, T. 2016. Effects of high CO
2
levels
on the ecophysiology of the diatom Thalassiosira
weissflogii differ depending on the iron nutritional
status. ICES Journal of Marine Science 73(3):
680–692. DOI: http://dx.doi.org/10.1093/icesjms/
fsv259.
Sulpis, O,Jeansson, E,Dinauer, A,Lauvset, SK,Mid-
delburg, JJ. 2021. Calcium carbonate dissolution
patterns in the ocean. Nature Geoscience 14:
423–428. DOI: http://dx.doi.org/10.1038/s41561-
021-00743-y.
Tatters, AO,Roleda, MY,Schnetzer, A,Fu, F,Hurd, CL,
Boyd, PW,Caron, DA,Lie, AAY,Hoffmann, LJ,
Hutchins, DA. 2013. Short- and long-term condi-
tioning of a temperate marine diatom community
to acidification and warming. Philosophical Transac-
tions of the Royal Society B: Biological Sciences
368(1627): 20120437. DOI: http://dx.doi.org/10.
1098/rstb.2012.0437.
Terrats, L,Claustre, H,Cornec, M,Mangin, A,Neuker-
mans, G. 2020. Detection of coccolithophore
blooms with BioGeoChemical-Argo floats. Geophysi-
cal Research Letters 47(23): e2020GL090559. DOI:
http://dx.doi.org/10.1029/2020GL090559.
Tortell, PD,Payne, CD,Li, Y,Trimborn, S,Rost, B,
Smith, WO,Riesselman, C,Dunbar, RB,Sedwick,
P,DiTullio, GR. 2008. CO
2
sensitivity of Southern
Ocean phytoplankton. Geophysical Research Letters
35(4): 1–5. DOI: http://dx.doi.org/10.1029/
2007GL032583.
Tre
´guer, PJ,De La Rocha, CL. 2013. The world ocean
silica cycle. Annual Review of Marine Science 5(1):
477–501. DOI: http://dx.doi.org/10.1146/annurev-
marine-121211-172346.
Tre
´guer, PJ,Sutton, JN,Brzezinski, M,Charette, MA,
Devries, T,Dutkiewicz, S,Ehlert, C,Hawkings, J,
Leynaert, A,Liu, SM,Llopis Monferrer, N,Lo
´pez-
Acosta, M,Maldonado, M,Rahman, S,Ran, L,
Rouxel, O. 2021. Reviews and syntheses: The bio-
geochemical cycle of silicon in the modern ocean.
Biogeosciences 18(4): 1269–1289. DOI: http://dx.
doi.org/10.5194/bg-18-1269-2021.
Trimborn, S,Brenneis, T,Sweet, E,Rost, B. 2013. Sen-
sitivity of Antarctic phytoplankton species to ocean
acidification: Growth, carbon acquisition, and spe-
cies interaction. Limnology and Oceanography
58(3): 997–1007. DOI: http://dx.doi.org/10.4319/
lo.2013.58.3.0997.
Tsujino, H,Urakawa, S,Nakano, H,Small, RJ,Kim,
WM,Yeager, S,Danabasoglu, G, Suzuki, T, Bam-
ber, JL,Bentsen, M, Bo
¨ning, CW,Bozec, A,
Chassignet, EP, Curchitser, E, Dias, FB,Durack,
PJ,Griffies, SM, Harada, Y, Ilicak, M, Josey, SA,
Kobayashi, C, Kobayashi, S, Komuro, Y, Large,
WG,Sommer, JL,Marsland, SJ, Masina, S, Schei-
nert, M, Tomita, H, Valdivieso, M, Yamazaki, D.
2018. JRA-55 based surface dataset for driving
ocean–sea-ice models (JRA55-do). Ocean Modelling
130: 79–139. DOI: http://dx.doi.org/10.1016/j.
ocemod.2018.07.002.
Vallina, SM,Ward, BA,Dutkiewicz, S,Follows, MJ.
2014. Maximal feeding with active prey-switching:
A kill-the-winner functional response and its effect
on global diversity and biogeography. Progress in
Oceanography 120: 93–109. DOI: http://dx.doi.
org/10.1016/j.pocean.2013.08.001.
Wang, Q,Danilov, S,Sidorenko, D,Timmermann, R,
Wekerle, C,Wang, X,Jung, T,Schro
¨ter, J.2014.
The Finite Element Sea Ice-Ocean Model (FESOM) v.
1.4: Formulation of an ocean general circulation
model. Geoscientific Model Development 7(2):
663–693. DOI: http://dx.doi.org/10.5194/gmd-7-
663-2014.
White, E,Hoppe, CJM,Rost, B. 2020. The Arctic picoeu-
karyote Micromonas pusilla benefits from ocean
acidification under constant and dynamic light. Bio-
geosciences 17(3): 635–647. DOI: http://dx.doi.org/
10.5194/bg-17-635-2020.
White, MM,Waller, JD,Lubelczyk, LC,Drapeau, DT,
Bowler, BC,Balch, WM,Fields, DM. 2018. Cocco-
lith dissolution within copepod guts affects fecal
pellet density and sinking rate. Scientific Reports
8(1): 1–6. DOI: http://dx.doi.org/10.1038/s41598-
018-28073-x.
Wolf, KKE,Hoppe, CJM,Rost, B.2018.Resilienceby
diversity: Large intraspecific differences in climate
change responses of an Arctic diatom. Limnology
and Oceanography 63(1): 397–411. DOI: http://dx.
doi.org/10.1002/lno.10639.
Wu, Y,Campbell, DA,Irwin, AJ,Suggett, DJ,Finkel, ZV.
2014. Ocean acidification enhances the growth rate
of larger diatoms. Limnology and Oceanography
59(3): 1027–1034. DOI: http://dx.doi.org/10.
4319/lo.2014.59.3.1027.
Wyatt, T. 2014. Margalef’s mandala and phytoplankton
bloom strategies. Deep Sea Research Part II: Topical
Studies in Oceanography 101: 32–49. DOI: http://
dx.doi.org/10.1016/j.dsr2.2012.12.006.
Yama naka, Y,Tajika, E.1996.Theroleofthevertical
fluxes of particulate organic matter and calcite in
the oceanic carbon cycle: Studies using an ocean
biogeochemical general circulation model. Global
Biogeochemical Cycles 10(2): 361–382. DOI:
http://dx.doi.org/10.1029/96GB00634.
Zhang, Y,Fu, F,Hutchins, DA,Gao, K. 2019. Combined
effects of CO
2
level, light intensity, and nutrient
availability on the coccolithophore Emiliania hux-
leyi.Hydrobiologia 842: 127–141. DOI: http://dx.
doi.org/10.1007/s10750-019-04031-0.
Zondervan, I. 2007. The effects of light, macronutrients,
trace metals and CO
2
on the production of calcium
Seifert et al: CO
2
effects on phytoplankton growth Art. 10(1) page 31 of 32

carbonate and organic carbon in coccolithophores—
A review. Deep Sea Research Part II: Topical Studies in
Oceanography 54(5): 521–537. DOI: http://dx.doi.
org/10.1016/j.dsr2.2006.12.004.
Zondervan, I,Rost, B,Riebesell, U. 2002. Effect of CO
2
concentration on the PIC/POC ratio in the cocco-
lithophore Emiliania huxleyi grown under light-
limiting conditions and different daylengths.
Journal of Experimental Marine Biology and Ecology
272: 55–70. DOI: http://dx.doi.org/10.1016/
S0022-0981(02)00037-0.
Zondervan, I,Zeebe, RE,Rost, B,Riebesell, U.2001.
Decreasing marine biogenic calcification: A negative
feedback on rising atmospheric pCO
2
.Global Biogeo-
chemical Cycles 15(2): 507–516. DOI: http://dx.doi.
org/10.1029/2000GB001321.
How to cite this article: Seifert, M, Nissen, C, Rost, B, Hauck, J. 2022. Cascading effects augment the direct impact of CO
2
on phytoplankton growth in a biogeochemical model.
Elementa: Science of the Anthropocene
10(1). DOI: https://doi.org/
10.1525/elementa.2021.00104
Domain Editor-in-Chief: Jody W. Deming, University of Washington, Seattle, WA, USA
Associate Editor: Jean-E
´ric Tremblay, Universite
´Laval, Que
´bec, Canada
Knowledge Domain: Ocean Science
Published: September 16, 2022 Accepted: July 29, 2022 Submitted: November 9, 2021
Copyright: ©2022 The Author(s). This is an open-access article distributed under the terms of the Creative Commons
Attribution 4.0 International License (CC-BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium,
provided the original author and source are credited. See http://creativecommons.org/licenses/by/4.0/.
Elem Sci Anth
is a peer-reviewed open access
journal published by University of California Press.
Art. 10(1) page 32 of 32 Seifert et al: CO
2
effects on phytoplankton growth