of 14
Article
https://doi.org/10.1038/s41467-023-36177-w
Pelagic calcium carbonate production
and shallow dissolution in the North
Paci
fi
cOcean
Patrizia Ziveri
1,2,3,13
, William Robert Gray
4,5,13
, Griselda Anglada-Ortiz
1,6
,
Clara Manno
7
,MichaelGrelaud
1
, Alessandro Incarbona
8
,
James William Buchanan Rae
5
,AdamV.Subhas
9
, Sven Pallacks
1
,
Angelicque White
10
,JessF.Adkins
11
& William Berelson
12
Planktonic calcifying organisms play a k
ey role in regulating ocean carbonate
chemistry and atmospheric CO
2
. Surprisingly, references to the absolute and
relative contribution of these organis
ms to calcium carbona
te production are
lacking. Here we report quanti
fi
cation of pelagic calcium carbonate produc-
tion in the North Paci
fi
c, providing new insights on the contribution of the
three main planktonic calcifying gro
ups. Our results show that coccolitho-
phores dominate the living calcium carbonate (CaCO
3
) standing stock, with
coccolithophore calcite co
mprising ~90% of total CaCO
3
production, and
pteropods and foraminifera playing a
secondary role. We show that pelagic
CaCO
3
production is higher than the sinking
fl
ux of CaCO
3
at 150 and 200 m at
ocean stations ALOHA and PAPA, imp
lying that a large portion of pelagic
calcium carbonate is remin
eralised within the photic zone; this extensive
shallow dissolution explains the app
arent discrepancy between previous
estimates of CaCO
3
production derived from satel
lite observations/biogeo-
chemical modeling versus estimates f
rom shallow sediment traps. We suggest
future changes in the CaCO
3
cycle and its impact on atmospheric CO
2
will
largely depend on how the poorly-un
derstood processes that determine
whether CaCO
3
is remineralised in the photic zone or exported to depth
respond to anthropogenic warming and acidi
fi
cation.
The marine calcium carbonate (CaCO
3
) cycle is a key component of the
global carbon cycle, and is intimately related to atmospheric CO
2
(ref.
1
). The formation of CaCO
3
in the ocean is a process largely con-
trolled by the biological calci
fi
cation of marine organisms
2
. Planktonic
calcifying organisms at the base of the food web (from primary pro-
ducers to zooplankton) have played a key role since the Mesozoic
3
,via
processes including regulation of surface water alkalinity, ballasting of
organic matter and alkalinity export, and establishment of a pelagic
carbonate buffer capable of in
fl
uencing major CO
2
change
4
6
.
Since the seminal work of Milliman
7
9
on the production and
accumulation of CaCO
3
in the ocean, several studies have aimed to
quantify total CaCO
3
pelagic production
10
12
and the contribution of
speci
fi
c calcifying plankton groups. However, the relative contribution
of the main calcifying taxa to total CaCO
3
pelagic production has not
yet been directly quanti
fi
ed.
There is large uncertainty in total pelagic CaCO
3
production,
with current estimates varying between 0.7
4.7 Pg C yr
1
(0.6
3.9×10
14
mol CaCO
3
yr
1
)
7
,
10
13
. In general, estimates based on
satellite observations or modeling of ecosystems/carbonate chem-
istry in the surface ocean suggest higher CaCO
3
production
10
,
12
,
13
whereas estimates based on export from the production layer typi-
cally report lower values
7
,
8
,
11
.
Received: 23 August 2020
Accepted: 18 January 2023
Check for updates
A full list of af
fi
liations appears at the end of the paper.
e-mail:
Patrizia.Ziveri@uab.cat
;
william.gray@lsce.ipsl.fr
Nature Communications
| (2023) 14:805
1
1234567890():,;
1234567890():,;
There is also uncertainty on the make-up of pelagic CaCO
3
pro-
duction by different groups. Mainly based on sediment trap export
fl
uxes and sediment data, there is a general understanding that coc-
colithophores (single-celled haptophytes inhabiting the photic zone,
performing photosynthesis and producing calcite) and planktonic
foraminifera (single-celled marine eukaryotes producing calcite) each
contribute ~50% to the global pelagic CaCO
3
production and
sedimentation
8
,
14
16
. However, more recent papers have highlighted
the potential of shelled pteropods (specialized free-swimming pelagic
sea snails producing aragonite) as an important component of pelagic
CaCO
3
production
13
,
17
,
18
. Other taxa such as heteropods (holoplank-
tonic gastropods with aragonite shells, Pterotracheoidea) may also
contribute to a lesser degree.
These planktonic calcifying taxa have speci
fi
c mechanisms of
biogenic calci
fi
cation, and associated differences in vulnerability to
ocean acidi
fi
cation
19
,
20
with their shell solubility depending on their
speci
fi
c polymorph mineralogy and Mg content
21
23
. In addition,
planktonic calci
fi
ers exhibit a large range of particulate inorganic
carbon to particulate organic carbon ratios (PIC/POC)
18
,
24
,
25
,which
in
fl
uences the integrated carbon export rain ratio, an important
term for carbon cycling in the oceans and atmospheric CO
2
1
,
26
.
Furthermore, and as we discuss later, the association of PIC and POC
together within calcifying organisms may play a critical role in
driving CaCO
3
dissolution above the saturation horizon
27
.However,
despite the importance of pelagic calci
fi
cation to the marine carbon
cycle, key questions remain about pelagic CaCO
3
production rates,
standing stocks, and export
fl
uxes. Most importantly, it is critical to
determine the contributions of different planktonic calcifying
groups to pelagic calci
fi
cation, the proportion of aragonite versus
calcite, and the magnitude of CaCO
3
production compared to
export.
The North Paci
fi
c Ocean is a key region for understanding the
role of pelagic calci
fi
ers in the global CaCO
3
budget, due to its large
volume and the wide range of biogeochemical conditions from
the subtropical to subpolar gyres. In addition, the waters of the
North Paci
fi
c are some of the most undersaturated in the global
ocean with respect to calcite and aragonite, and thus calcifying
organisms in the region are most at risk of future ocean acidi
fi
cation
driven by anthropogenic CO
2
emissions
28
,
29
. Although there are
studies of the relative distribution of pelagic calci
fi
ers in the North
Paci
fi
c
30
,
31
, estimates of their relative contribution to CaCO
3
stand-
ing stock and production rates are severely lacking. The discovery
of excess alkalinity above the saturation horizon in the North Paci
fi
c
has sparked debate about the role of different pelagic calci
fi
ers and
their contribution to the alkalinity budget above the thermo-
dynamic saturation horizon
23
,
32
,
33
.
We conducted a research cruise from subtropical to subpolar
North Paci
fi
c waters in which we assessed the pelagic living CaCO
3
standing stock. We targeted the main planktonic calci
fi
ers at
fi
ve
survey stations, from Honolulu, Hawaii, to Seward, Alaska, (Fig.
1
,
Tables S1 and S2). We deployed plankton nets to sample calcifying
zooplankton and rosettes of Niskin bottles to target calcifying
phytoplankton. In addition, four intermediate planktonic towing
stations were sampled and integrated into the overall data set
(Fig.
1
, Tables S1, S2). Coccolithophores, foraminifera, pteropods,
and heteropods were quanti
fi
ed and the CaCO
3
biomass was esti-
mated, providing the
fi
rst overall picture of the total CaCO
3
living
standing stock (i.e. inventory), and the relative contribution of
the main calcite and aragonite planktonic producers in the pro-
ductive upper ocean. Using estimates of turnover time for each
group we estimate annual production (Methods), and compare this
to aragonite and calcite biomineral export out of the surface ocean
to 100 and 200 m water depth estimated using
fl
oating sediment
traps deployed during the time of sampling
34
and historical time
series in the region.
Results and discussion
North Paci
fi
cCaCO
3
standing stocks
The total CaCO
3
standing stock is lower in the nutrient-poor and less
productive subtropical gyre (~560
900 mg m
-2
;notehereand
throughout we refer to mg of CaCO
3
,unlessspeci
fi
ed otherwise), and
strongly increases into the nutrient-rich and productive subpolar gyre
(~1700
4500 mg m
2
total) (Figs.
2
and
3
a, Table S3), re
fl
ecting the
major ecological shift across the North Paci
fi
c, from low-CaCO
3
pro-
duction in the oligotrophic subtropics, to high-CaCO
3
production in
the subpolar region
32
,
35
.We
fi
nd that coccolithophores dominate the
CaCO
3
producing standing stock at all stations, demonstrating a mean
contribution of ~79% (with a range of 62
96% across all sites) to the
total CaCO
3
standing stocks. Pteropods contribute ~14% (3
29%
range), followed by foraminifera (~6% mean, 0.1
22% range), and het-
eropods (~1%, 0
2%). Calcite from coccolithophores and foraminifera
is thus the most abundant mineral, constituting ~86% of the standing
stock (71
96%), with aragonite making up ~14% of the standing
stock (4
30%).
The coccosphere CaCO
3
standing stock depth pro
fi
les follow the
overall chlorophyll
fl
uorescence, albeit with the scaling between
fl
uorescence and coccosphere CaCO
3
varying between stations, (Fig.
2
,
Fig. S1), indicating a substantial contribution by haptophytes/cocco-
lithophores to the total standing stock of photosynthetic algae
production
36
. We observe a shallowing of the chlorophyll maximum
depth and coccolithophore CaCO
3
standing stock maximum from
subtropical to subpolar stations (Fig.
2
).
Our estimates of living coccolithophore CaCO
3
standing stocks
range from 0.13 mg m
-3
at 175 m (St. 1) in the subtropical gyre to
110 mg m
-3
at 30 m (St. 5) in the subpolar gyre, with depth-integrated
estimates from 753 mg m
-2
to 3048 mg m
-2
at the same stations,
respectively (Fig.
2
b, c, Tables S2 and S3). Our results support previous
work in the North Paci
fi
c based on coccolithophore cell concentration,
which showed that their biomass is highest at high latitudes,
decreasing in temperate and subtropical regions
30
(Fig. S8). The sub-
polar and transitional North Paci
fi
c Ocean is also known as a region of
sustained seasonal
E. huxleyi
(the most abundant and cosmopolitan
coccolithophore species) blooms
37
with an estimated maximum
satellite-derived PIC concentration of ~0.8 mmol m
3
(CaCO
3
of
80 mg m
-3
) in August/September (note, satellite PIC is limited to
retrievals over the
fi
rst optical depth of satellite data, ~10 m). There is a
remarkable agreement between our estimated values of coccolitho-
phore CaCO
3
from the shallowest sampled water depth (~5 m, i.e. the
surface CaCO
3
concentration) and satellite-derived PIC concentrations
1
10
100
mg m
3
–180
–160
–140
–120
10
20
30
40
50
60
St.1
St.2
St.3
St.4
St.5
ALOHA
PA PA
St.2.5
St.3.5
St.4.5
St.6.5
PIC August climatology
Lon (°E)
Lat (°N)
Fig. 1 | Satellite PIC and location map.
August satellite-derived Particular Inor-
ganic Carbon
97
(PIC; mg CaCO
3
m
-3
) climatology (2002-2017) and location of C-
DisK-IV stations (black crosses) and long-term sediment trap studies (orange/pink
crosses). Large black crosses show the location of Niskin bottle rosette, plankton
tow, and
fl
oating sediment trap sampling sites at C-DisK-IV stations. Small black
crosses show sites with additional plankton tow sampling only at C-DisK-IV stations.
Note the logarithmic scale.
Article
https://doi.org/10.1038/s41467-023-36177-w
Nature Communications
| (2023) 14:805
2
(Figs. S2 and S3), which are mostly tuned to capture coccolithophore
PIC. This supports the high correlation between satellite-derived PIC
and measured PIC surface water concentrations suggested by Balch
and others;
10
,
38
however as previously noted
38
and discussed later,
much of the coccolith PIC production can occur below the depth of the
optical retrieval, particularly in the subtropics where the production
layer deepens.
We also quantify the CaCO
3
contribution of loose coccoliths
(calcite plates extruded to the cell surface forming the coccosphere;
Fig.
2
b; Table S5). Coccoliths are shed into the surrounding waters
following death and breakup of the coccosphere, or produced con-
tinuously by some species
38
40
.WefoundthatloosecoccolithCaCO
3
can contribute signi
fi
cantly to the total CaCO
3
standing stock in the
productive photic layer, with maximum values of 44 and 64 mg m
-3
in
E. huxleyi
blooms at Stations 3 and 5 (Fig. S3). Our results show loose
coccolith distribution is tied to the distribution of intact coccospheres
andPIC,asobservedinpreviousstudies
10
,
38
.Wenotethatourcocco-
lithophore living standing stock estimates (used to calculate CaCO
3
production rate) only include whole coccosphere cells, and excludes
loose coccoliths.
Our pteropod standing stock concentrations in the subpolar gyre
range from 109
802 ind. m
-3
, broadly within the published range of the
pteropod standing stocks in the northwestern Paci
fi
c(e.g.;
41
,
42
,Fig.
3
;
Table S3; Fig. S5) and the Gulf of Alaska
43
,
44
, although pteropods show a
signi
fi
cant seasonal and inter-annual variability in the coastal habitats
of the Gulf of Alaska
43
,
44
. The recent study of Bednar
š
ek et al.
45
found
that the abundance of pteropods collected in May 2015 in the subpolar
gyre and Gulf of Alaska was (42
423 ind m
-2
), around two orders of
magnitude lower than our abundances observed in August 2017 in the
same region (Stations 4
6, Table S3) and previous studies
43
.Thisdif-
ference in abundance may relate to seasonal and inter-annual varia-
bility in this region, and/or the use of larger mesh size in their study
(200
335
μ
m)
45
, which could have resulted in an underrepresentation
of the small size pteropod (juvenile) fraction. Our estimates from the
subtropical gyre (22
391 ind. m
-3
)(TableS3)aresimilartoprevious
estimates for this region and are higher than values observed across
the Atlantic Ocean
46
.
Our pteropod CaCO
3
concentrations range from 0.2
8.6 mg m
-3
(Fig.
3
a). We
fi
nd good agreement between our estimates of pteropod
CaCO
3
standing and North Paci
fi
c sites in the MAREDAT database
17
(Methods; Fig.
3
), which show a typical concentration of 0.5 (0.2
1.1,
68% CI) mg m
-3
. Despite our pteropod CaCO
3
standing stock results
being similar to/higher than previous estimates from the North Paci
fi
c/
North Atlantic, our values are lower than the proposed 23.17 mg CaCO
3
m
-3
global-mean value of shelled pteropod CaCO
3
standing stock
concentration
18
. However, our analysis shows that this global dataset is
heavily skewed (Skewness = 13.3, where a value above one indicates a
skewed dataset), with the median pteropod biomass value reported
18
being around three orders of magnitude smaller than the reported
mean. As such the global-mean value is not a useful descriptor of this
data compilation
18
. Our analysis shows typical pteropod CaCO
3
bio-
mass globally is 0.3 (0.08
0.9, 68% CI) mg m
-3
in the upper 250 m and
0.2 (0.07
0.8, 68% CI) mg m
-3
in the upper 1000 m (Methods; Sup-
plemental Fig. S6), two orders of magnitude lower than the global-
mean value reported by ref.
18
, and in line with our results from the
North Paci
fi
c.
Our vertically integrated pteropod CaCO
3
standing stocks range
between ~64
111 mg m
-2
in subtropical gyre and between
~215
1306 mg CaCO
3
m
-2
in the subpolar gyre (Fig.
3
; Table S3; Fig. S5).
We
fi
nd good agreement between our estimates and the vertically
integrated pteropod CaCO
3
standing stock calculated from North
Paci
fi
c sites in the MAREDAT database
17
(Methods; Fig.
4
), which show
a typical value of 121 (50
270, 68%) mg m
-2
, with our estimates thus
being slightly higher.
Our heteropod standing stock concentrations range from 5
40
ind. m
-3
and 0.01
0.1 mg CaCO
3
m
-3
, and their presence is limited to
the subtropics and transition zone. Although previous estimates of
heteropod standing stocks are extremely scarce, comparison to
previous abundances from a latitudinal Atlantic Ocean transect
con
fi
rmed that heteropods almost exclusively inhabit warm waters
and the recorded maximum of 0.7 ind. m
-3
(ref.
46
) is lower than our
estimates in the North Paci
fi
c. Our vertically integrated heteropod
CaCO
3
standing stocks range from 3-35 mg CaCO
3
m
-2
; heteropods
thus contribute between 3
12% of the total aragonite standing stock
300
250
200
150
100
50
0
Depth (m)
0.0
1.0
2.0
3.0
Fluorescence
(
mg/m
3
)
St.1
St.2
St.3
St.4
St.5
a
0
5
10
15
CaCO
3
(
mg m
3
)
b
0.00
0.05
0.10
0.15
(
mmol m
3
)
0 20
60
100
CaCO
3
(
mg m
3
)
c
0.0
0.4
0.8
1.2
(
mmol m
3
)
0123456
Omega
d
40
80
coccospheres
coccoliths
calcite
aragonite
subtropical gyre
transition zone/
subpolar gyre
Fig. 2 | Coccolithophore standing stock vertical pro
fi
les. a
chlorophyll
fl
uorescence. Coccosphere, and coccolith CaCO
3
from C-DisK-IV stations (
b
) 1 and 2 and, (
c
)3,4,
and 5. Note the different x axis range on panels (
b
)and(
c
).
d
Omega calcite and aragonite at the
fi
ve stations.
Article
https://doi.org/10.1038/s41467-023-36177-w
Nature Communications
| (2023) 14:805
3
in the subtropics and transition zone, but are absent from the sub-
polar region.
Our estimates of integrated foraminiferal CaCO
3
standing stock
range from 9
37 mg m
-2
in the subtropical gyre to 182
404 mg m
-2
in
the subpolar gyre. Although previous estimates of foraminiferal
standing stock in the North Paci
fi
c are scarce, our estimates of the
integrated vertical standing stock of the number of foraminifera from
the subtropical gyre/transition zone sites are similar to, or slightly
higher, than previous estimates of the integrated vertical standing
stock from the subtropical gyre/transition zone in the western North
Paci
fi
c
31
,
47
(Figs. S4, 5). Our estimates of the vertically integrated
standing stock of the number of foraminifera in the subpolar North
Paci
fi
c(190,000
250,000 ind. m
-2
) are generally higher than the esti-
mates of Taylor et al.
31
, which ranged up to ~80,000 ind. m
-2
,although
such high values are not unprecedented, with previous estimates of
the vertically integrated standing stock from the North Atlantic ran-
ging up to ~390,000 ind. m
-2
ref.
48
.
Pelagic CaCO
3
production
We calculate CaCO
3
production rate by dividing our measurements of
the living CaCO
3
standing stock by estimates of the turnover time (i.e.
typical life span) of each group (Methods, Fig.
3
b). Our approach
assumes all of the organisms within the standing stock are living; this is
valid for foraminifera, pteropods, and heteropods as individuals sink
relatively quickly after death
47
,
49
, and the individuals sampled con-
tained cytoplasm/soft tissue (Methods). For coccolithophores this
assumption is valid as we only consider intact coccospheres, which
disaggregate quickly upon death into the component coccoliths
50
,and
is supported by the fact that the peaks of coccolithophore CaCO
3
matchthepeaksinchlorophyll
fl
uorescence (Fig.
2
). We include the
caveat that our approach assumes the living standing stock is in
approximate steady state.
Given that coccolithophores have a shorter turnover time
(1.5
10 days) than the other calcifying groups (Methods,
Table
1
,Fig.
3
b) and dominate the CaCO
3
standing stock, they account
for ~86% (67
97%) of total CaCO
3
annual production across the sites.
Pteropods contribute ~10% to total production (2
17%), heteropods
~0.3% (0
1%), and foraminifera contribute ~2% (0.02
9%). As such, 89%
of the CaCO
3
production is calcite (70
97%), with the remainder being
aragonite (Fig.
3
c).
Given the large seasonality of PIC production
51
,
52
(Fig. S7), we
estimate annual CaCO
3
production correcting for seasonal bias
(Methods, Table S6). Our seasonally corrected annual CaCO
3
produc-
tion estimates range from 0.2
0.4 mol m
-2
yr
-1
in the subtropical gyre, a
similar range or slightly lower than the estimate of the production rate
of 0.7 mol m
-2
yr
-1
in the subtropical/tropical Atlantic
53
, although in
good agreement with the global mean estimate of 0.4 mol m
-2
yr
-1
(ref.
10
,Fig.
3
d). Our estimates from the transition zone and the
12345
0
1000
2000
3000
4000
5000
station
CaCO
3
standing st
ock
(
mg
m
2
)
ptero
hetero
foram
cocco
total
a
0
10
20
30
40
50
(
mmol
m
2
)
turnover time (days)
0
5
10
15
20
25
30
ptero
hetero
foram
cocco
b
12345
0
500
1000
1500
2000
2500
3000
3500
station
d
aily C
a
CO
3
produ
ctio
n
(
mg
m
2
d
ay
)
August
2017
c
0
5
10
15
20
25
30
35
(
m
mol
m
2
day
)
12345
0
1
2
3
4
5
station
annual CaCO
3
production
(
mol
m
2
yr
)
seasonally
adjusted
d
STG
SPG
TZ
STG
SPG
TZ
STG
SPG
TZ
Fig. 3 | Standing stock and production by pelagic calci
fi
ers. a
living CaCO
3
standing stock
b
turnover time of calcifying taxa used to calculate production from
standing stock (the range represented by the bar length is applied with a
fl
at
probability distribution in our error propagation)
c
CaCO
3
production per day
(August 2017)
d
CaCO
3
annual production corrected for seasonal bias using
satellite-derived PIC/chlorophyll
97
,
104
and zooplankton seasonality estimates (all
data and metadata are publicly available at hahana.soest.hawaii.edu/hot/hot-dogs/
interface.html). The total CaCO
3
production is shown by the violin plots in panels
(
c
)and(
d
), where the probability density of the estimate is represented by the
thickness of the shaded area and the grey lines show the 68% and 95% con
fi
dence
interval (CI); note the non-normal distribution with the high-tail on the upper
estimate. Error bars for the standing stock (
a
) and production (
c
,
d
)byindividual
taxon represent the 95% CI (Methods). STG, TZ, and SPG represent subtropical
gyre, transition zone, and subpolar gyre, respectively. Purple bands on panels
a
,
c
,and
d
show 68% range of pteropod standing stock and daily/annual production
calculated using the MAREDAT database (Methods). The blue stars on panel
d
show
the estimates of total production calculated with in-situ pH and fCO
2
measure-
ments at Ocean Station PAPA (ref.
57
, light blue), and estimates of production in the
subpolar North Paci
fi
c calculated using the seasonal cycle of alkalinity and dis-
solved inorganic carbon (ref.
12
, dark blue). STG, TZ, and SPG represent subtropical
gyre, transition zone, and subpolar gyre, respectively.
Article
https://doi.org/10.1038/s41467-023-36177-w
Nature Communications
| (2023) 14:805
4
productive subpolar gyre are higher (0.9
1.0 mol m
-2
yr
-1
)than
this global average even at the 95% con
fi
dence interval (Fig.
3
d);
however, they agree well with the estimate of production calculated
with in-situ pH and fCO
2
measurements at Ocean Station PAPA
through (1.2 mol m
2
yr
-1
;ref.
54
), and estimates of production in the
subpolar North Paci
fi
c calculated using the seasonal cycle of alkalinity
and dissolved inorganic carbon (0.9 ± 0.1 mol m
-2
yr
-1
;ref.
12
).
To explore the implications of our estimates of CaCO
3
produc-
tion for global production, we use a global climatology of satellite-
derived PIC (Fig. S7). While there is a high correlation between
satellite PIC and our estimates of surface PIC concentration (Fig. S2a;
ref.
38
), our results indicate depth integrated CaCO
3
production is
only twice as high in the nutrient-rich subpolar gyre than the
nutrient-poor subtropical gyre, smaller than the range expected from
the satellite PIC (CaCO
3
) estimates, which suggest PIC concentrations
~6
7 times higher in the subpolar region compared to the subtropics.
This difference likely re
fl
ects the deeper (coccolithophore) CaCO
3
production in the subtropics, as well as the thickness of the cocco-
lithophore productive layer (Fig.
2
) from the upper ~175 m in the
subtropics to the upper ~75 m in subpolar region, which will bias the
satellite-derived PIC estimates to lower values
38
. For Stations
1
4 surface PIC is below 10 mg m
-3
(both at the time of sampling and
in the annual mean climatology), yet we observe a depth integrated
seasonally corrected production of 0.2
1 mol m
-2
yr
-1
at these sites
(Table S6). Note, this surface PIC value is very similar to the threshold
proposed by Balch et al.
38
(0.13 mmol m
-3
/13mgm
-3
) between
surface-dominated and depth-dominated CaCO
3
production
regimes. Similar low surface PIC regimes (with annual surface PIC of
<10 mg m
-3
) represent 87% of the surface of the ocean (Fig. S8); thus,
assuming similar rates of CaCO
3
production as the seasonally cor-
rected production rates at Stations 1
4 globally puts a minimum
estimate (assuming the remaining 13% of the ocean with higher sur-
face PIC will have higher depth integrated production) for total
global CaCO
3
production of ~2.2×10
14
mol yr
-1
(2.6 Pg C yr
-1
), that is
0.0
0.5
1.0
1.5
2.0
0.00
0.01
0.02
Pteropods
0 - 250 m
CaCO
3
biomass
CDisk-4
MAREDAT
a
012345
0.00
0.05
0.10
0.15
0.20
0.25
0.30
probability density (frac. tot.)
P
32
P
50
P
68
P
max
0.09
0.22
0.49
0.03
0246810
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0
0.2 0.4 0.6 0.8 0.1
probability density (frac. tot.)
P
32
P
50
P
68
P
max
0.19
0.47
1.05
0.19
0
200
600
1000
0.00
0.05
0.10
0.15
0.20
0.25
0.30
probability density (frac. tot.)
P
32
P
50
P
68
P
max
49
117
262
48
0
50 100
200
0.0
0.1
0.2
0.3
0.4
0.5
probability density (frac. tot.)
P
32
P
50
P
68
P
max
5
12
26
4
C biomass (mg/m
3
)
CaCO
3
biomass (mg/m
3
)
integrated CaCO
3
standing stock (mg/m
2
)
daily CaCO
3
production (mg/m
2
/day)
b
c
d
e
(mmol/m
3
)
02
6
10
012
(mmol/m
2
)
48
(mmol/m
2
/day)
(mg/m
3
)
(mmol/m
3
)
Fig. 4 | Pteropod CaCO
3
biomass estimates. a
Pteropod CaCO
3
biomass estimated
from the MAREDAT database
17
and measured in this study; note, maximum values
extend above 2 mg m
-3
. Probability density of pteropod (
b
) Carbon biomass (
c
)
CaCO
3
biomass (
d
) Integrated CaCO
3
standing stock (
e
)dailyCaCO
3
production
calculated using samples in the upper 250 m of the North Paci
fi
cfromthe
comprehensive MAREDAT database
17
,
18
(Methods). Red shading indicates 32
68%
con
fi
dence interval range. Red values show the 32nd, 50th, 68th percentiles;
orange value shows the value with the highest probability (all values given in mg).
Note the distributions are highly skewed.
Article
https://doi.org/10.1038/s41467-023-36177-w
Nature Communications
| (2023) 14:805
5
0.8×10
14
mol yr
-1
using the production rate at Station 1 and 3.6×10
14
mol yr
-1
using the production rate at Station 3.
To make a
fi
rst-order approximation of the impact of deepening
CaCO
3
production on global CaCO
3
production we use a simple lin-
ear regression of total CaCO
3
production at our sites against satellite
PIC (CaCO
3
) (Fig. S2), with the deepening of production primarily
manifesting as a non-zero intercept (note the high variability in the
intercept coming from the changing production regimes from Sta-
tions 1 to 4 despite the low surface PIC at all Stations). We then apply
this relationship to global satellite PIC climatology (Fig. S8). We
include the caveat that 1) this assumes the bias caused by the dee-
pening of CaCO
3
production in the subtropics scales with surface PIC
in a similar way globally
38
, and 2) the regression in Fig. S2 is driven by
one station with high surface PIC (Station 5). While crude, this
approach allows us to make a
fi
rst-order approximation of the impact
of deepening CaCO
3
production on global CaCO
3
production.
Applying the relationship between total production and satellite PIC
(Table S7) to the global mean surface satellite-derived PIC climatol-
ogy (Fig. S8), and integrating globally (weighting by area) results in a
total CaCO
3
production of 3.1×10
14
mol yr
-1
(3.7 Pg C yr
-1
) globally. This
estimate is similar to, although toward the upper end, of previous
estimates of total pelagic calci
fi
cation based on satellites, upper
water column measurements, seasonal alkalinity changes, and eco-
system modeling which range from 0.9
3.9×10
14
mol CaCO
3
yr
-1
(1.1
4.7 Pg C yr
-1
)
10
,
12
,
13
,
28
,
32
,
55
. However, as previously noted by
others
10
,
13
this estimate is considerably higher than estimates of glo-
bal mean CaCO
3
export
fl
ux from the upper ocean which ranges from
0.5
0.6×10
14
mol yr
-1
(0.6
0.7 Pg C yr
-1
; refs.
7
,
8
).
While our estimate of the total amount of CaCO
3
produced agrees
well with that of Buitenhuis et al.
13
, there is a large discrepancy between
our results and those of Buitenhuis et al.
13
in terms of the dominant
CaCO
3
polymorph produced. We
fi
nd CaCO
3
production is dominated
by calcitic coccolithophores, however, their results suggested pelagic
CaCO
3
production is mainly driven by aragonite pteropods with coc-
colithophores and foraminifera playing a minor role.
Given the discrepancy with the results of Buitenhuis et al.
13
and the
limited temporal interval of our sampling and the potential for
large temporal variability of pteropod abundances
43
, we also calculate
pteropod CaCO
3
biomass and production in the North Paci
fi
c using the
comprehensive MAREDAT database
18
, which has excellent spatial and
seasonal sampling distribution in the North Paci
fi
c (Methods, Fig.
4
).
This results in a typical pteropod CaCO
3
biomass in the upper 250 m of
the North Paci
fi
c of 0.5 mg m
-3
(0.2
1, 32
68% CI; note the dataset is
highly skewed, Fig. S6), a vertically integrated pteropod CaCO
3
bio-
mass of 122 mg m
-2
(50
269, 32
68% CI), and a pteropod CaCO
3
pro-
ductionrateof12mgm
-2
day
-1
(5
27, 32
68% CI). The results calculated
using the MAREDAT database are thus in good agreement with the
estimates calculated using the samples collected during our own
cruise (Fig.
3
). We propose the discrepancy with the results of Bui-
tenhuis et al.
13
instead comes from three other factors: Firstly, their
model parametrization uses a
fi
xed PIC/POC ratio of 0.1 ref.
56
for
coccolithophores; this value is substantially lower than the published
review by Gafar et al.
25
which ranged from 0.19 to 2.30 and much lower
than the value 0.52 they used for pteropods, which is itself about two
times higher than the estimates of Bednar
š
ek et al.
18
(Table
1
). Sec-
ondly, they assume a similar turnov
er time for (single-celled) cocco-
lithophores and (complex) pteropods, contrary to the available
estimates from the literature
57
60
(Table
1
). Finally, within their calcu-
lation they assume all CaCO
3
dissolving above the calcite saturation
horizon is aragonite, an assumption which is likely to exaggerate ara-
gonite production; as we discuss below, previous studies
52
,
61
,
27
and our
results indicate substantial dissolution of coccolithophore calcite
above the calcite saturation horizon (which we attribute to respiration-
driven dissolution and dissolution within the guts of grazers) such that
this assumption is likely to be invalid.
CaCO
3
sinking and export
fl
uxes versus production
Our estimate of 3.1×10
14
mol CaCO
3
yr
-1
global pelagic CaCO
3
produc-
tion is ~30
300 times larger than required to meet the 0.8 10
12
-
1.1×10
13
mol CaCO
3
buried in deep sea sediments each year
8
,
11
,
62
64
and
balance the riverine of input of alkalinity to maintain steady-state,
reaf
fi
rming previous
fi
ndings that most of the CaCO
3
produced in the
surface ocean is dissolved and recycled within the ocean interior (e.g.,
ref.
8
). More surprisingly, at several stations our estimates of CaCO
3
production are larger than the export
fl
uxes at 100 to 200 m water
depth in
fl
oating sediment trap deployed during the plankton
sampling
34
, and our production estimates at Stations 1 and 5 are higher
than the long running shallow sediment traps at Station ALOHA
65
and
Ocean Station PAPA
64
,
66
,
67
)(Fig.
5
). While the discrepancy observed
with the
fl
oating traps deployed during the sampling interval (in place
for ~72 hrs) may be explained by a decoupling of CaCO
3
production
and natural mortality/sinking of pteropods, and coccolith aggregation,
such that there could be a time lag between production at the surface
and export through the water column, this time lag cannot explain the
discrepancy observed with the long running sediment traps at Ocean
Station PAPA and Station ALOHA. Our results show an annual pro-
duction of 0.4 (0.2-2.1, 95% CI) and 0.9 (0.5-3.8) mol CaCO
3
m
-2
yr
-1
at
Stations 1 and 5, which is ~5 times higher than the annual export of 0.08
and 0.16 mol CaCO
3
m
-2
yr
-1
at 150 m depth at ALOHA and 200 m at
PAPA (Fig.
5
c, d). We reiterate that previous estimates of annual CaCO
3
production at PAPA station based on seasonal cycle of seawater car-
bonate chemistry support our production value
12
,
54
,
57
.
This disparity between the amount of CaCO
3
produced, and the
amount of CaCO
3
that is exported out of the photic zone, suggests that
a large portion (~80%) of the total CaCO
3
produced in the photic zone
is never exported, and is instead remineralised in situ; that is, only
~20% of the total CaCO
3
produced is exported from the photic zone.
Bishop & Wood
52
suggestedupto92%ofthetotalCaCO
3
produced
dissolved within the upper 500 m in the subpolar North Paci
fi
c. In situ
remineralisation of such a high fraction of the CaCO
3
that is produced
within the photic zone explains the previous discrepancy between
higher estimates of global CaCO
3
production based on satellites, upper
water column measurements, seasonal alkalinity changes, and eco-
system modeling (which all estimate the total amount of CaCO
3
Table 1 | Ratio of Particular Inorganic car
bon (PIC) to Particular Organic Carbon (PO
C) and turnover time (life span) for the
calcifying taxon
Group
PIC:POC ratio
PIC:POC references
Turnover time (days)
Turnover time references
Coccolithophores
0.19
2.08
a
ref.
98
and references therein
0.6
10 days (0.1
1.5 cell divisions per day)
ref.
59
and references therein
Pteropods
0.20
0.56
b
ref.
18
,
99
5
16
ref.
13
,
58
,
100
Heteropods
0.28
0.45
This study
5
16
ref.
13
,
58
,
100
Planktonic foraminifera
3
6
ref.
101
and references therein; ref.
102
,
103
14
28
c
ref.
78
and references therein
a
C. leptoporus, E. huxleyi, C. pelagicus subsp. braarudii, G. oceanica, S. apsteinii, H. carteri, S. pulchra, U. sibogae.
b
Limaciniidae and Cavoliniidae families.
c
excludes deeper dwelling species with longer turnover times, however, these comprise only a very minor component of the assemblages
30
,
31
.
Article
https://doi.org/10.1038/s41467-023-36177-w
Nature Communications
| (2023) 14:805
6