Marine ecosystem role in setting up preindustrial and future climate

Jerry F Tjiputra; Damien Couespel; Richard Sanders

doi:10.1038/s41467-025-57371-y

. 2025 Mar 5;16:2206. doi: 10.1038/s41467-025-57371-y

Marine ecosystem role in setting up preindustrial and future climate

Jerry F Tjiputra ^1,^✉, Damien Couespel ¹, Richard Sanders ¹

PMCID: PMC11883021 PMID: 40044713

Abstract

The ocean ecosystem is a vital component of the global carbon cycle, storing enough carbon to keep atmospheric CO₂ considerably lower than it would otherwise be. However, this conception is based on simple models, neglecting the coupled land-ocean feedback. Using an interactive Earth system model, we show that the role ocean biology plays in controlling atmospheric CO₂ is more complex than previously thought. Atmospheric CO₂ in a new equilibrium state after the biological pump is shut down increases by more than 50% (163 ppm), lower than expected as approximately half the carbon lost from the ocean is adsorbed by the land. The abiotic ocean is less capable of taking up anthropogenic carbon due to the warmer climate, an absent biological surface pCO₂ deficit and a higher Revelle factor. Prioritizing research on and preserving marine ecosystem functioning would be crucial to mitigate climate change and the risks associated with it.

Subject terms: Carbon cycle, Climate and Earth system modelling, Marine chemistry, Marine biology

A new study reveals that the ocean biological pump plays a vital role in maintaining low atmospheric CO2 levels and facilitating efficient future anthropogenic carbon sinks. Without it, climate change would occur at an amplified and accelerated rate.

Introduction

As a key climate regulator, the ocean has slowed anthropogenic climate change by absorbing 91% of the heat trapped in the Earth system between 1971 and 2018¹ and 26% of anthropogenic carbon emitted since the preindustrial period². Subsequently, it will play a crucial role in determining future climate. Air-sea fluxes of carbon dioxide primarily arise from the air-sea disequilibrium of the partial pressure of CO₂, driven by two main mechanisms: (i) the solubility pump, associated with CO₂ gas solubility and circulation, and (ii) the biological pump, the photosynthetic conversion of near-surface dissolved inorganic carbon (DIC) into organic matter that is exported into the ocean interior via gravitational sinking and circulation³. Through mechanism (ii), the biological carbon pump (BCP) effectively reduces surface-preformed DIC while increasing remineralized DIC at depth. This vertical redistribution and variations in preformed to remineralized DIC determines the partitioning of ocean-atmosphere carbon stocks⁴. Using an atmosphere-ocean equilibrium relationship, it is estimated that when the remineralized carbon stock is removed, atmospheric CO₂ levels would be approximately 150-240 ppm higher in the new equilibrium state⁵.

Currently the BCP is often considered to operate at steady state, with the physically-driven solubility pump considered to have dominated the projected increase in current and future carbon sink and storage rates^6,7, due to the limited observational evidence and knowledge on the non-linear interactions and feedbacks between BCP and other Earth system components⁸. As a result, there is currently no consensus on the projected changes in BCP efficiency.

Repeated attempts to quantify the importance of BCP for future carbon storage have been conducted using ocean models of varying complexities. These include altering the chemical stoichiometry of organic matter production⁹, changing the remineralization depth¹⁰, enhancing micronutrient fertilization¹¹, and simulating future climate change with an abiotic ocean (i.e., by complete removal of all marine organisms)¹², among others. All these studies were performed with stand-alone ocean models, neglecting the carbon cycle feedback processes that occur between the ocean, land, and atmosphere. Such approaches can either underestimate or overestimate the impact of changing BCP on the long-term ocean carbon uptake⁸. In the face of rapidly intensifying climate change and anthropogenic pressures on marine ecosystems, better quantification of BCP’s role in mitigating future climate change is of fundamental importance for climate science and policy developments.

Here, we assess the role of the BCP in setting up the preindustrial climate and quantify the impacts of its removal on future carbon sequestration and climate change. We apply the state-of-the-art Norwegian Earth System Model (NorESM2)^13,14 in its fully interactive configuration to simulate quasi-equilibrium preindustrial climate and subsequently project historical and future climate scenarios with and without marine organisms (see “Methods”). The impacts across three future scenarios, ranging from high CO₂ emissions SSP5-8.5 (Shared Socioeconomic Pathways) to a strong mitigation SSP1-2.6¹⁵, are investigated. Our simulations allow for prognostic atmospheric CO₂ concentration, taking into account carbon fluxes from land and ocean and the associated feedback to the climate system. The preindustrial reorganization of the carbon pools across the Earth system in the absence of BCP leads to more than 50% (163 ppm) higher atmospheric CO₂ and a 1.6 °C warmer surface temperature. Consequently, the Earth system’s capacity to buffer anthropogenically-induced climate change is hampered. Despite having lower DIC content, an abiotic ocean counterintuitively has a higher surface DIC and pCO₂ concentration, leading to a higher Revelle factor and a reduced anthropogenic carbon uptake. Future land and ocean carbon sinks are projected to reduce considerably, accelerating and amplifying anthropogenic climate change.

Results

Preindustrial states

Two distinct quasi-equilibrium preindustrial climate states are simulated, with and without ocean biology (hereafter referred to as REF and Abiotic, respectively; see Methods and Supplementary Table 1). Key climate metrics and carbon budget estimates for these two preindustrial states are summarized in Table 1. The Abiotic ocean releases roughly 730 Pg C to the atmosphere during the 2000-year spin-up period (Supplementary Fig. 1). Of this, 345 Pg C is adsorbed by land, primarily through CO₂ fertilization-induced vegetation growth¹⁶. The remainder stays in the atmosphere, leading to a 163 ppm higher CO₂ concentration (445 ppm) compared to REF (282 ppm). This increases global mean surface temperature by 1.64 °C, while sea surface temperature (SST) increases by 1.15 °C. This warming leads to considerable impacts on the Earth system, such as reducing Atlantic Meridional Overturning Circulation (AMOC) strength (9%) and sea-ice area in the Arctic and Antarctic (23 and 24%, respectively; Table 1). Changes in sea-ice extent are seasonally non-uniform, with September Arctic sea-ice reducing the most, by more than half. Without ocean biology, global ocean temperature increases from 3.56 to 4.66 °C, and as the surface ocean equilibrates with the higher atmospheric CO₂ concentration, its Revelle factor increases by 16%.

Table 1.

Global mean and projected change in key climate metrics and carbon budgets in REF and Abiotic simulations

	Preindustrial		Δ (SSP5-8.5)
Variables	Reference	Abiotic	Reference	Abiotic	Units
Surface air temperature	14.52	16.16	3.77	4.88	°C
Atmospheric CO₂	282	445	727	925	ppm
Sea surface temperature	18.78	19.93	2.48	3.48	°C
Ocean temperature	3.56	4.66	0.36	0.62	°C
Maximum AMOC	20.40	18.62	− 10.45	− 10.53	Sv
Dissolved inorganic carbon	37694	37367	497	397	Pg C
Marine carbon sediment	1853	1462	203	− 29	Pg C
Sea-to-air CO₂ flux	− 0.11	0.07	− 5.45	− 3.59	Pg C yr⁻¹
Surface ocean Revelle factor	9.83	11.40	3.21	2.97	–
Land vegetation carbon	537	812	202	96	Pg C
Land soil and litter carbon	2578	2646	71	− 41	Pg C
Land-to-air CO₂ flux	0.25	− 0.02	− 5.84	− 3.26	Pg C yr⁻¹
Arctic sea-ice area	11.15	8.57	− 6.22	− 8.12	10⁶ km²
Antarctic sea-ice area	6.85	5.19	− 1.32	− 2.76	10⁶ km²

Open in a new tab

Mean values are averaged over the first 10 years of the preindustrial control simulations, and projected change (Δ) are the difference between the 2091–2100 and 1851–1890 periods from the SSP5-8.5 future scenario and historical simulations.

The spatial patterns of surface temperature in the two preindustrial climate states are similar but warmer everywhere in the Abiotic (Supplementary Fig. 2a, b). The Abiotic ocean exhibits stronger CO₂ outgassing in the tropical oceans and ingassing in the mid-to-high latitudes (Supplementary Fig. 2d, e). The zonal outgassing band along the Southern Ocean circumpolar fronts seen in REF, associated with the upwelling of remineralized carbon-rich watermasses¹⁷, switches into a net sink in Abiotic. In the subtropical oligotrophic, the rates of CO₂ flux are relatively similar in REF and Abiotic. The altered land carbon budget includes an increased vegetation carbon pool, with the largest increase simulated in the tropics region, followed by high-latitude ecosystems in the northern hemisphere (Supplementary Fig. 2g, h).

The stronger and expanded tropical outgassing in the Abiotic is consistent with (i) an absence of biological consumption of surface DIC and (ii) warming-induced lower CO₂ solubility, both of which increase surface pCO₂. In the high latitudes, stronger uptakes can be explained by the amplification of the seasonal cycle. Supplementary Fig. 3 illustrates that the removal of ocean biology fundamentally alters the seasonal cycle of the surface carbonate system. Firstly, the thermal-driven pCO₂ variability is amplified due to higher background pCO₂¹⁸ and higher SST seasonal variation. Secondly, the lack of summer productivity and winter upwelling of remineralized carbon-rich deep water amplify the thermally driven pCO₂ variability. Thirdly, the indirect effect of land biosphere changes (stronger summer productivity and winter respiration) leads to an amplification of the atmospheric CO₂ seasonal cycle (Supplementary Fig. 3a,d). These three effects combine to increase the air-sea pCO₂ gradient, with a stronger effect during winter, and manifest in a stronger high-latitude carbon sink in the Abiotic ocean (with up to a three-fold increase; Supplementary Fig. 3c, f). We note that in our quasi-equilibrium preindustrial Abiotic ocean, the sedimentary carbon content is reduced by 391 Pg C (Table 1), due to the absence of organic matter accumulations and dissolution of organic materials. This additional DIC to the water column could contribute to the overall oceanic carbon release. We also note that our experiment represents an extreme hypothetical case with a complete absence of marine productivity. When we consider a 10% reduction in primary production, a range projected by models¹⁹, Supplementary Fig. 4 shows that the preindustrial atmospheric CO₂ would approximately be 10 ppm higher, while the land carbon budget increases by 9 Pg C, implying ocean outgassing of approximately 30 Pg C.

Historical and future projections

Projections of historical and future climate (1850–2100) show accelerated climate change in Abiotic relative to REF (Fig. 1 and Supplementary Fig. 5). Atmospheric CO₂ in REF grows from 281 to 415 (SSP1-2.6), 569 (SSP2-4.5), and 1061 (SSP5-8.5) ppm by 2100 (Fig. 1a and Supplementary Fig. 5a), consistent with the CO₂ pathways defined in the CMIP6 (Coupled Model Intercomparison Project phase 6) protocol¹⁵. In the Abiotic, CO₂ increases from 445 to 685 (SSP1-2.6), 883 (SSP2-4.5), and 1433 (SSP5-8.5) ppm. The stronger CO₂ increase in the Abiotic (by 106, 150, and 208 ppm in SSP1-2.6, SSP2-4.5, and SSP5-8.5, respectively) is caused by the land and ocean carbon sinks diminishing by as much as 91% (i.e., land sink in SSP2-4.5). The weaker ocean carbon sink, despite higher atmospheric CO₂, is attributed to the higher surface ocean pCO₂ and lower air-sea pCO₂ disequilibrium (Fig. 1f). Figure 2 summarizes that without ocean biology, 68 to 83% of the fossil fuel emissions would remain in the atmosphere by 2100, as compared to only 37 to 65% in simulations that include ocean biology.

Fig. 2 — Shown are the initial CO₂ concentration at the start of the historical period (dark blue), cumulative release of fossil fuel emissions (yellow), net land sink (including land-use changes, green), net ocean sink (light blue), and the final CO₂ concentration at the end of the 21st century (red) in [ppm] units. The pie charts depict fractions [in percentage] of fossil fuel emissions taken up by the land, ocean, and atmosphere. Values are from *REF* and *Abiotic* simulations for (a, b) historical + SSP1-2.6, (c, d) historical + SSP2-4.5, and (e, f) historical + SSP5-8.5 scenarios.

These larger anthropogenic CO₂ residuals being stored in the atmosphere translate to a stronger climate sensitivity by accelerating climate change in Abiotic, with broad implications for various components of the Earth system (Fig. 1 and Supplementary Fig. 5). Under SSP5-8.5, the global mean surface temperature in Abiotic increases by nearly 5 °C by 2100, a 30% stronger warming rate than in REF (Table 1). Regionally, the largest warming occurs at high latitudes, with approximately 2 °C additional warming in the Arctic due to polar amplification²⁰ (Supplementary Fig. 6). Similarly, the AMOC strength, the sea-ice area, and the ocean mixed layer depth, are projected to decline faster in Abiotic (Fig. 1c, g–i). A complete removal of ocean biology leads to the system crossing the 2 °C warming threshold more than ten years earlier, with the disappearance of summer Arctic sea ice occurring more than 50 years earlier (Fig. 1b, g). This accelerated climate change is also evident in the lower emissions scenarios of SSP1-2.6 and SSP2-4.5 (Supplementary Table 2). The transient climate response to CO₂ emissions (TCRE) for the REF and Abiotic are 1.69 and 2.13 °C Eg C⁻¹, respectively (Supplementary Fig. 7), implying a 20% reduction of the allowable CO₂ emissions for a specified global warming target when ocean biology is removed.

The reduced terrestrial carbon sinks in the Abiotic reflects the saturation of land carbon uptake in a warmer, higher CO₂ world due to the higher preindustrial vegetation carbon pool^21,22 (Fig. 1e and Table 1), hence weakening the negative terrestrial carbon cycle feedback associated with CO₂ fertilization. In the ocean, warmer and higher surface pCO₂ reduces seawater CO₂ solubility and buffering capacity²³, leading to a lower air-sea pCO₂ disequilibrium and less CO₂ uptake for a given increase in atmospheric CO₂ concentration (Fig. 1d, f and Table 1). The lack of ocean production-induced pCO₂ deficit, which has significant impacts at higher atmospheric CO₂ levels, also plays an important role²⁴.

Next, we analyze how the spatio-temporal dynamics of the anthropogenic carbon (Cant) sink and storage in the ocean are altered by the absence of ocean biology. The cumulative ocean uptake of Cant (i.e., the difference between the transient historical + future and the preindustrial control simulations) over the 1850–2100 period is spatially heterogeneous, with deep water formation regions of the North Atlantic and the Southern Ocean emerging as the most intense sink regions (Fig. 3). These key gateways for Cant sinks allow for efficient transport of Cant-rich surface water to the interior for long-term storage^25,26. Stronger uptake is simulated in the high-emission SSP5-8.5 than SSP2-4.5 and SSP1-2.6, with regions of net Cant outgassing (more prominently in Abiotic) are simulated in the subtropics, where surface Cant converges²⁷. The Abiotic Ocean absorbs considerably less Cant than REF in the subpolar North Atlantic and parts of the Southern Ocean (Fig. 3c, f, i).

We investigate the driving mechanisms of the lower Abiotic Cant uptake, focusing in the subpolar North Atlantic (dashed green outlines in Fig. 3c, f, i). Since the seasonal cycle of CO₂ fluxes is fundamentally altered in Abiotic, we assess the projection of different drivers in each season (Supplementary Fig. 8). The air-sea gradient of pCO₂ determines to first order the direction of CO₂ flux. In REF, the oceanic pCO₂ stays below the atmospheric values, and the subpolar North Atlantic is a net CO₂ sink all year, consistent with observations and models²⁸. This is also seen in the Abiotic at the beginning of the historical period but is progressively altered by the faster oceanic pCO₂ growth rate, resulting in a remarkable decline in the CO₂ sink. More importantly, under SSP5-8.5 (and other scenarios) oceanic pCO₂ grows faster than atmospheric pCO₂ during spring and summer, therefore switching from uptake to outgassing. This is primarily driven by the combination of stronger warming and the higher initial surface Revelle factor, which reduce the Abiotic ocean capacity to offset the growing atmospheric pCO₂ (Supplementary Fig. 9), as depicted by the weaker surface DIC growth rate (Supplementary Fig. 8).

Discussion

Accurate assessment of the role of the marine ecosystem in regulating the Earth’s climate is challenging due to the non-linear interactions and feedback that link different components of the Earth system. Applying a fully interactive Earth system model, we show that a complete removal of marine ecosystems increases preindustrial atmospheric CO₂ by 163 ppm, at the lower end of the range estimate using atmosphere-ocean equilibrium relationship (150–240 ppm)⁵. A previous study using a coupled physical-biogeochemical ocean model, and thus excluding the land feedback, gave a greater increase of atmospheric CO₂ (> 200 ppm) after only 250 years²⁹. In contrast, a theoretical study linking atmospheric CO₂ and the efficiency of BCP predicts only a 100 ppm increase when BCP is shut down⁴. In our simulations, the positive feedback due to climate warming as atmospheric CO₂ increases is overshadowed by negative feedback from the land biosphere, which absorbs nearly half of the released oceanic CO₂. Nevertheless, there are other feedbacks that are not considered, such as (i) reduction in the biogenic marine dimethyl sulfide emissions, which act as aerosol cooling agents³⁰, and (ii) warming-induced methane release from the permafrost³¹. These feedbacks would enhance warming-induced climate feedback, further increasing the preindustrial atmospheric CO₂ level. Therefore, our estimated climate impacts likely represent the lower end of the expected range.

The reorganization of carbon pools within the Earth system alters the trajectories of anthropogenic climate change. The newly equilibrated abiotic ocean is warmer and contains less DIC. The absence of a biologically mediated vertical DIC gradient (Supplementary Fig. 10) leads to a higher surface pCO₂ and Revelle factor. This higher surface Revelle factor impedes the oceanic sink of anthropogenic CO₂²³, implying that future changes in primary production may negatively affect anthropogenic CO₂ uptake. Similarly, the reallocation of carbon from the ocean to the land biosphere saturates the vegetation growth response to higher atmospheric CO₂²¹. Depending on the emissions scenario, the ocean and land CO₂ uptake in Abiotic could be reduced by up to 34% and 91%, respectively, increasing the cumulative airborne fraction of CO₂ emissions from 37%, 50%, 65% to 68%, 76%, 83% for the SSP1-2.6, SSP2-4.5, SSP5-8.5 scenarios (Fig. 2). The projected climate change rate is therefore enhanced, particularly in the polar regions. Considering the lack of the above-mentioned feedback processes, these climate change figures are likely underestimated. This accelerated climate change also illustrates the potential consequence if the longevity of natural carbon sink is reduced faster or earlier than initially expected³².

The simulated changes shown here represent extreme end-member examples of the role that ocean biology plays in controlling atmospheric CO₂, both in the steady state and under greenhouse forcing. However, they are valuable and clearly show that estimates of the effects of biologically-mediated ocean CO₂ storage in the preindustrial has on atmospheric CO₂ based on simple metrics, i.e., integrated remineralised CO₂, need to be treated with some cautions as any reduction in this term would likely lead to an enhanced terrestrial carbon sink. In addition, our simulations indicate that ocean biological processes play a crucial role in driving the uptake of Cant in the modern ocean by adjusting the spatial and temporal patterns in the surface pCO₂ deficit.

Our assessment of Cant sequestration demonstrates that the largest impact of an abiotic ocean occurs in the ventilation regions of the North Atlantic and Southern Ocean, where long-term surface-to-interior Cant export is significantly reduced, particularly during summer, when the absence of biological export production and enhanced surface warming keep surface pCO₂ high (see Supplementary Fig. 9) and impede the ocean carbon sinks in the more stratified future ocean. Despite community consensus on the importance of ocean biology in the Earth system, its representations in state-of-the-art models are far from perfect. This leads to large uncertainty when simulating the observed CO₂ flux variability³³ and projecting future changes in ocean biology^7,19,34. In Earth system models, the BCP is generally positively associated with surface primary productivity, nevertheless there is a nuanced relationship between primary production, carbon export, and carbon sequestration. Spatial and temporal variations in surface productivity and ecosystem structure may have non-linear impacts on BCP efficiency, complicating the direct links between reductions in productivity and CO₂ sequestration³⁵. Addressing this complexity would be valuable to guide future climate modeling efforts, carbon cycle research, and policy development. Interdisciplinary research that integrates marine biological observations and models using various approaches to constrain the impacts of future global change should also be embraced³⁶. Our finding also contradicts the current notion that the biological carbon pump plays little to no role in adsorbing excess Cant from the atmosphere. Instead, it clearly plays an important role by setting up the disequilibria and Revelle factor conditions that determine the magnitude of this uptake.

In order to optimize the service the ocean offers us by mitigating anthropogenic climate change, our results emphasize the importance of sustaining a healthy and well-functioning marine ecosystem. Our simulations, though extreme, suggest that the dissolution of sedimentary organic material can alter the water column DIC budget and, subsequently, the air-sea CO₂ fluxes. This has obvious parallels to human-induced disturbances of the seafloor by trawling or dredging; however, these will be much smaller in size, and hence future studies applying a more realistic scenario would be necessary to investigate the impact of human-induced disruption of marine sediments.

In addition to improvements in various dynamical processes in the Earth system (cloud feedback, land carbon cycle, ocean circulation, albedo feedback, etc.), better representations of marine primary production and biological carbon pump should not be overlooked to constrain estimates of climate sensitivity and future climate projections. Future developments in ocean biogeochemistry models should emphasize improving the spatio-temporal biological processes in key ocean carbon sink regions (i.e., the North Atlantic and Southern Ocean). Sustained monitoring of marine biological carbon pumps would be necessary to improve our estimates of future ocean carbon sinks and, subsequently, the development of robust strategies for climate mitigation.

Methods

Model description and experimental design

We used the second generation of the Norwegian Earth System Model (NorESM2-LM), which contributed to the Coupled Model Intercomparison Project phase 6 (CMIP6)³⁷ and the sixth assessment report of the Intergovernmental Panel for Climate Change³⁸. It couples atmosphere, ocean, sea-ice, and land modules and simulates the physical and biogeochemical interactions between them. Here the NorESM2-LM was configured in a fully interactive mode, allowing for prognostic atmospheric CO₂ concentrations by accounting for online air-sea and air-land CO₂ fluxes. A full description of the model components and their performance has been extensively validated and documented^13,14. The NorESM2-LM simulates well the observed large scale pattern of surface primary productivity, and its annual rate is well within the range of observational estimates and other CMIP6 models³⁹. Two sets of preindustrial climate state and future projections were performed: (i) reference (REF), where the ocean biology was not modified, and (ii) Abiotic, where the ocean primary productivity was switched off, and the ocean carbon sources and sinks were driven only by physical processes. The riverine fluxes of biogeochemical substances were deactivated in Abiotic to balance the diminished sinks of materials into the sediments. We also applied climatological marine dimethyl sulfide (DMS) fluxes based on observations to compensate for the radiative imbalance at the top of the atmosphere. This last step was necessary since the impact of diminishing phytoplankton-produced DMS on the climate system is currently not well understood.

Prior to transient historical and future simulations, the NorESM2-LM was spun up until a quasi-equilibrium preindustrial climate and carbon cycle states were achieved. For REF, the model was initialized from observations and spun up for 1600 years with a prescribed constant atmospheric CO₂ concentration of 284 ppm, followed by an additional spin-up with prognostic CO₂ for 250 years. For Abiotic, the spin-up was started at the end of REF’s spin-up, with the marine productivity module deactivated and integrated for 2000 years. Following these spin-ups, both REF and Abiotic reach a stable preindustrial climate state with sufficiently low drifts for the purpose of this study (Supplementary Table 1).

Here, we summarize the transition state occurring in the Abiotic spin-up. Following a complete cessation of marine productivity, the ocean starts to outgass carbon to the atmosphere, leading to a rapid decrease in DIC storage and an increase in atmospheric CO₂ in the first few hundred years (Supplementary Fig. 1). The lower DIC content is reflected by declining remineralized DIC and is compensated by the gradual increase in preformed DIC and the dissolution of sedimentary carbon. The resulting higher atmospheric CO₂ induces higher terrestrial primary production and vegetation growth through the CO₂ fertilization effect¹⁶, removing a substantial fraction of the ocean carbon release. The soil and litter carbon pools also increase, though only slightly.

Starting from the two quasi-equilibrium preindustrial states, we performed transient historical (1850–2014) and three future scenarios (2015–2100) simulations following the CMIP6 protocol³⁷. The three future scenarios considered were: (i) low CO₂ emissions SSP1-2.6 (Shared Socioeconomic Pathways), which is consistent with a 2 °C warming by 2100 relative to the preindustrial, (ii) moderate CO₂ emissions SSP2-4.5, and (iii) high CO₂ emissions SSP5-8.5. Corresponding preindustrial control simulations (1850-2100) for each REF and Abiotic were also performed. These preindustrial control simulations were applied to correct model drifts and estimate the anthropogenic carbon (Cant) content in the ocean. At the start of the transient historical simulation (1850), the total carbon inventories for the atmosphere, land, and ocean (water column+sediment) reservoirs for REF (Abiotic) are 599 (945), 3115 (3458), and 39547 (38829) Pg C, respectively. We note that the noticeable decline in net land CO₂ uptake in the early 21st century under the SSP5-8.5 scenario (Fig. 1e) is associated with the loss of carbon due to the prescribed land use change and fires.

Uncertainty analysis

The results of our experiments, specifically on how the atmospheric CO₂ evolves with the release of oceanic carbon, depend on the land carbon cycle response or feedback to higher atmospheric CO₂. In our experiment the atmospheric CO₂ slowly increases from preindustrial level to 445 ppm in 2000 years. The most comparable experiment under the CMIP6 framework is the extended historical and SSP1-2.6 scenario (1850-2300), with the increasing atmospheric CO₂ stabilizes toward 396 ppm after 450 years of integration⁴⁰. Five ESMs have provided their outputs: ACCESS-ESM1.5, CanESM5, IPSL-CM6-LR, MIROC-ES2L, and UKESM1-0-LL. Supplementary Fig. 11 shows that the multi-model mean of cumulative land carbon sink plus total land-use-related carbon budget (from year 1850–2150 is 215 Pg C)⁴¹ amounts to 337 ± 26.73 Pg C, which is in good agreement with our total land carbon tbudget increase of 343 Pg C (Table 1) in the preindustrial Abiotic. We also assess the robustness of our projected preindustrial key climate response. Our simulated temperature increase, AMOC decline, and sea-ice cover decline are all within the range of CMIP6 ESMs (Supplementary Fig. 11). We note that under historical and SSP1-2.6 scenarios, in addition to atmospheric CO₂ increase, there are also aerosols and other prescribed forcings, such as volcanic eruption and other greenhouse gases, that could affect the climate.

Supplementary information

Supplementary Information^{(1.7MB, pdf)}

Peer Review file^{(245.7KB, pdf)}

Acknowledgements

This work was funded by the European Union under grant agreement no. 101083922 (OceanICU, J.T., D.S., and R.S.) and UK Research and Innovation (UKRI) under the UK government’s Horizon Europe funding guarantee [grant number 10054454, 10063673, 10064020, 10059241, 10079684, 10059012, 10048179]. The views, opinions, and practices used to produce this dataset/software are, however, those of the author(s) only and do not necessarily reflect those of the European Union or European Research Executive Agency. Neither the European Union nor the granting authority can be held responsible for them. This study is a contribution to the Research Council of Norway projects INES2 (no. 350390) and Navigate (no. 352142). We acknowledge Timothée Bourgeois for preparing the SSP2-4.5 CO₂ emissions file. High-performance computing and storage resources were provided by the Norwegian Research Infrastructure Services (projects nn1002k, nn2345k, and ns1002k).

Author contributions

Funding acquisition: R.S. and J.F.T.; Conceptualization: J.F.T.; Analysis of the results: J.F.T., D.C., and R.S.; Writing: J.F.T., D.C., and R.S.; Editing: J.F.T., D.C., and R.S.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.

Funding

Open access funding provided by NORCE Norwegian Research Centre AS.

Data availability

The NorESM2-LM model outputs (Reference and Abiotic) for the historical and future scenarios simulations can be obtained at the Earth System Grid Federation portal https://esgf-node.ipsl.upmc.fr and 10.11582/2024.00083.

Code availability

Model codes for the NorESMs model are publicly available in https://noresm-docs.readthedocs.io/en/noresm2/access/access.html.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-025-57371-y.

References

1.Pörtner, H.-O. et al. The Ocean and Cryosphere in a Changing Climate, pp. 39–69. Cambridge University Press, Cambridge, UK and New York, NY, USA (2019).
2.Friedlingstein, P. et al. Global carbon budget 2023. Earth Syst. Sci. Data15, 5301–5369 (2023). [Google Scholar]
3.Volk, T., Hoffert, M. in Ocean Carbon Pumps: Analysis of Relative Strengths and Efficiencies in Ocean-Driven Atmospheric CO2 Changes. 99–110, American Geophysical Union; Geophysical Monograph 32 (1985).
4.Ito, T. & Follows, M. J. Preformed phosphate, soft tissue pump and atmospheric CO₂. J. Mar. Res.63, 813–839 (2005). [Google Scholar]
5.Goodwin, P., Follows, M.J. & Williams, R.G. Analytical relationships between atmospheric carbon dioxide, carbon emissions, and ocean processes. Global Biogeochem. Cycles22, 10.1029/2008GB003184 (2008).
6.Aricò, S. et al. Integrated Ocean Carbon Research: A Summary of Ocean Carbon Research, and Vision of Coordinated Ocean Carbon Research and Observations for the Next Decade. (2021).
7.Wilson, J. D. et al. The biological carbon pump in CMIP6 models: 21st century trends and uncertainties. Proc. Natl. Acad. Sci. USA119, 2204369119 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Oschlies, A. Impact of atmospheric and terrestrial CO₂ feedbacks on fertilization-induced marine carbon uptake. Biogeosciences6, 1603–1613 (2009). [Google Scholar]
9.Schneider, B., Engel, A. & Schlitzer, R. Effects of depth- and CO₂-dependent C:N ratios of particulate organic matter (POM) on the marine carbon cycle. Global Biogeochem. Cycles18, 10.1029/2003GB002184 (2004).
10.Kwon, E. Y., Primeau, F. & Sarmiento, J. L. The impact of remineralization depth on the air-sea carbon balance. Nat. Geosci.2, 630–635 (2009). [Google Scholar]
11.Moore, J.K., Doney, S.C., Lindsay, K., Mahowald, N. & Michaels, A.F. Nitrogen fixation amplifies the ocean biogeochemical response to decadal timescale variations in mineral dust deposition. Tellus B Chem. Phys. Meteorol.58, 560–572 (2006).
12.Sarmiento, J. L. & Quéré, C. L. Oceanic carbon dioxide uptake in a model of century-scale global warming. Science274, 1346–1350 (1996). [DOI] [PubMed] [Google Scholar]
13.Seland, Ø. et al. Overview of the Norwegian Earth System Model (NorESM2) and key climate response of CMIP6 DECK, historical, and scenario simulations. Geosci. Model Dev.13, 6165–6200 (2020). [Google Scholar]
14.Tjiputra, J. F. et al. Ocean biogeochemistry in the Norwegian Earth System Model version 2 (NorESM2). Geosci. Model Dev.13, 2393–2431 (2020). [Google Scholar]
15.Meinshausen, M. et al. The shared socio-economic pathway (SSP) greenhouse gas concentrations and their extensions to 2500. Geosci. Model Dev.13, 3571–3605 (2020). [Google Scholar]
16.Hickler, T. et al. CO₂ fertilization in temperate face experiments not representative of boreal and tropical forests. Glob. Change Biol.14, 1531–1542 (2008). [Google Scholar]
17.Mikaloff Fletcher, S.E. et al. Inverse estimates of the oceanic sources and sinks of natural CO₂ and the implied oceanic carbon transport. Global Biogeochem. Cycles21, 10.1029/2006GB002751 (2007).
18.Takahashi, T. et al. Global sea-air CO₂ flux based on climatological surface ocean pCO₂, and seasonal biological and temperature effects. Deep Sea Res. Part II Top. Stud. Oceanogr.49, 1601–1622 (2002). [Google Scholar]
19.Tagliabue, A. et al. Persistent uncertainties in ocean net primary production climate change projections at regional scales raise challenges for assessing impacts on ecosystem services. Front. Clim.3, 10.3389/fclim.2021.738224 (2021).
20.Holland, M. M. & Bitz, C. M. Polar amplification of climate change in coupled models. Clim. Dyn.21, 221–232 (2003). [Google Scholar]
21.Arora, V. K. et al. Carbon–concentration and carbon–climate feedbacks in CMIP6 models and their comparison to CMIP5 models. Biogeosciences17, 4173–4222 (2020). [Google Scholar]
22.Asaadi, A. et al. Carbon cycle feedbacks in an idealized simulation and a scenario simulation of negative emissions in CMIP6 earth system models. Biogeosciences21, 411–435 (2024). [Google Scholar]
23.Egleston, E.S., Sabine, C.L. & Morel, F.M.M. Revelle revisited: Buffer factors that quantify the response of ocean chemistry to changes in DIC and alkalinity. Global Biogeochem. Cycles24, 10.1029/2008GB003407 (2010).
24.Hauck, J. & Völker, C. Rising atmospheric CO₂ leads to large impact of biology on Southern Ocean CO₂ uptake via changes of the Revelle factor. Geophys. Res. Lett.42, 1459–1464 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Tjiputra, J. F., Assmann, K. & Heinze, C. Anthropogenic carbon dynamics in the changing ocean. Ocean Sci.6, 605–614 (2010). [Google Scholar]
26.Williams, R. G., Ceppi, P., Roussenov, V., Katavouta, A. & Meijers, A. J. S. The role of the Southern Ocean in the global climate response to carbon emissions. Philos. Trans. R. Soc. A Math. Phys. Eng. Sci.381, 20220062 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Couespel, D. & Tjiputra, J. What goes in must come out: the oceanic outgassing of anthropogenic carbon. Env. Res. Lett.19, 014086 (2024). [Google Scholar]
28.Pérez, F. F. et al. An assessment of CO₂ storage and sea-air fluxes for the Atlantic Ocean and Mediterranean Sea between 1985 and 2018. Glob. Biogeochem. Cycles38, 2023–007862 (2024). [Google Scholar]
29.Maier-Reimer, E., Mikolajewicz, U. & Winguth, A. Future ocean uptake of CO₂: interaction between ocean circulation and biology. Clim. Dyn.12, 711–722 (1996). [Google Scholar]
30.Schwinger, J. et al. Amplification of global warming through pH dependence of DMS production simulated with a fully coupled Earth system model. Biogeosciences14, 3633–3648 (2017). [Google Scholar]
31.Schuur, E. A. G. et al. Climate change and the permafrost carbon feedback. Nature520, 171–179 (2015). [DOI] [PubMed] [Google Scholar]
32.Canadell, J. G. et al. in Climate Change 2021: The Physical Science Basis. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA (2021).
33.Rodgers, K. B. et al. Seasonal variability of the surface ocean carbon cycle: A synthesis. Glob. Biogeochem. Cycles37, 2023–007798 (2023). [Google Scholar]
34.Kwiatkowski, L. et al. Twenty-first century ocean warming, acidification, deoxygenation, and upper-ocean nutrient and primary production decline from CMIP6 model projections. Biogeosciences17, 3439–3470 (2020). [Google Scholar]
35.Frenger, I. et al. Misconceptions of the marine biological carbon pump in a changing climate: Thinking outside the “export” box. Glob. Change Biol.30, 17124 (2024). [DOI] [PubMed] [Google Scholar]
36.Tagliabue, A. ‘Oceans are hugely complex’: modelling marine microbes is key to climate forecasts. Nature623, 250–252 (2023). [DOI] [PubMed] [Google Scholar]
37.Eyring, V. et al. Overview of the Coupled Model Intercomparison Project phase 6 (CMIP6) experimental design and organization. Geosci. Model Dev.9, 1937–1958 (2016). [Google Scholar]
38.Forster, P.et al. The Earth’s Energy Budget, Climate Feedbacks, and Climate Sensitivity. in Climate Change 2021: The Physical Science Basis. Cambridge University Press, Cambridge, UK and New York, NY, USA (2019).
39.Séférian, R. et al. Tracking improvement in simulated marine biogeochemistry between CMIP5 and CMIP6. Curr. Clim. Change Rep.6, 95–119 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
40.O’Neill, B. C. et al. The Scenario Model Intercomparison Project (ScenarioMIP) for CMIP6. Geosci. Model Dev.9, 3461–3482 (2016). [Google Scholar]
41.Silvy, Y. et al. AERA-MIP: Emission pathways, remaining budgets and carbon cycle dynamics compatible with 1.5 and 2 °C global warming stabilization. Earth Syst. Dyn.15, 1591–1628 (2024). [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information^{(1.7MB, pdf)}

Peer Review file^{(245.7KB, pdf)}

Data Availability Statement

Model codes for the NorESMs model are publicly available in https://noresm-docs.readthedocs.io/en/noresm2/access/access.html.

[CR1] 1.Pörtner, H.-O. et al. The Ocean and Cryosphere in a Changing Climate, pp. 39–69. Cambridge University Press, Cambridge, UK and New York, NY, USA (2019).

[CR2] 2.Friedlingstein, P. et al. Global carbon budget 2023. Earth Syst. Sci. Data15, 5301–5369 (2023). [Google Scholar]

[CR3] 3.Volk, T., Hoffert, M. in Ocean Carbon Pumps: Analysis of Relative Strengths and Efficiencies in Ocean-Driven Atmospheric CO2 Changes. 99–110, American Geophysical Union; Geophysical Monograph 32 (1985).

[CR4] 4.Ito, T. & Follows, M. J. Preformed phosphate, soft tissue pump and atmospheric CO₂. J. Mar. Res.63, 813–839 (2005). [Google Scholar]

[CR5] 5.Goodwin, P., Follows, M.J. & Williams, R.G. Analytical relationships between atmospheric carbon dioxide, carbon emissions, and ocean processes. Global Biogeochem. Cycles22, 10.1029/2008GB003184 (2008).

[CR6] 6.Aricò, S. et al. Integrated Ocean Carbon Research: A Summary of Ocean Carbon Research, and Vision of Coordinated Ocean Carbon Research and Observations for the Next Decade. (2021).

[CR7] 7.Wilson, J. D. et al. The biological carbon pump in CMIP6 models: 21st century trends and uncertainties. Proc. Natl. Acad. Sci. USA119, 2204369119 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR8] 8.Oschlies, A. Impact of atmospheric and terrestrial CO₂ feedbacks on fertilization-induced marine carbon uptake. Biogeosciences6, 1603–1613 (2009). [Google Scholar]

[CR9] 9.Schneider, B., Engel, A. & Schlitzer, R. Effects of depth- and CO₂-dependent C:N ratios of particulate organic matter (POM) on the marine carbon cycle. Global Biogeochem. Cycles18, 10.1029/2003GB002184 (2004).

[CR10] 10.Kwon, E. Y., Primeau, F. & Sarmiento, J. L. The impact of remineralization depth on the air-sea carbon balance. Nat. Geosci.2, 630–635 (2009). [Google Scholar]

[CR11] 11.Moore, J.K., Doney, S.C., Lindsay, K., Mahowald, N. & Michaels, A.F. Nitrogen fixation amplifies the ocean biogeochemical response to decadal timescale variations in mineral dust deposition. Tellus B Chem. Phys. Meteorol.58, 560–572 (2006).

[CR12] 12.Sarmiento, J. L. & Quéré, C. L. Oceanic carbon dioxide uptake in a model of century-scale global warming. Science274, 1346–1350 (1996). [DOI] [PubMed] [Google Scholar]

[CR13] 13.Seland, Ø. et al. Overview of the Norwegian Earth System Model (NorESM2) and key climate response of CMIP6 DECK, historical, and scenario simulations. Geosci. Model Dev.13, 6165–6200 (2020). [Google Scholar]

[CR14] 14.Tjiputra, J. F. et al. Ocean biogeochemistry in the Norwegian Earth System Model version 2 (NorESM2). Geosci. Model Dev.13, 2393–2431 (2020). [Google Scholar]

[CR15] 15.Meinshausen, M. et al. The shared socio-economic pathway (SSP) greenhouse gas concentrations and their extensions to 2500. Geosci. Model Dev.13, 3571–3605 (2020). [Google Scholar]

[CR16] 16.Hickler, T. et al. CO₂ fertilization in temperate face experiments not representative of boreal and tropical forests. Glob. Change Biol.14, 1531–1542 (2008). [Google Scholar]

[CR17] 17.Mikaloff Fletcher, S.E. et al. Inverse estimates of the oceanic sources and sinks of natural CO₂ and the implied oceanic carbon transport. Global Biogeochem. Cycles21, 10.1029/2006GB002751 (2007).

[CR18] 18.Takahashi, T. et al. Global sea-air CO₂ flux based on climatological surface ocean pCO₂, and seasonal biological and temperature effects. Deep Sea Res. Part II Top. Stud. Oceanogr.49, 1601–1622 (2002). [Google Scholar]

[CR19] 19.Tagliabue, A. et al. Persistent uncertainties in ocean net primary production climate change projections at regional scales raise challenges for assessing impacts on ecosystem services. Front. Clim.3, 10.3389/fclim.2021.738224 (2021).

[CR20] 20.Holland, M. M. & Bitz, C. M. Polar amplification of climate change in coupled models. Clim. Dyn.21, 221–232 (2003). [Google Scholar]

[CR21] 21.Arora, V. K. et al. Carbon–concentration and carbon–climate feedbacks in CMIP6 models and their comparison to CMIP5 models. Biogeosciences17, 4173–4222 (2020). [Google Scholar]

[CR22] 22.Asaadi, A. et al. Carbon cycle feedbacks in an idealized simulation and a scenario simulation of negative emissions in CMIP6 earth system models. Biogeosciences21, 411–435 (2024). [Google Scholar]

[CR23] 23.Egleston, E.S., Sabine, C.L. & Morel, F.M.M. Revelle revisited: Buffer factors that quantify the response of ocean chemistry to changes in DIC and alkalinity. Global Biogeochem. Cycles24, 10.1029/2008GB003407 (2010).

[CR24] 24.Hauck, J. & Völker, C. Rising atmospheric CO₂ leads to large impact of biology on Southern Ocean CO₂ uptake via changes of the Revelle factor. Geophys. Res. Lett.42, 1459–1464 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR25] 25.Tjiputra, J. F., Assmann, K. & Heinze, C. Anthropogenic carbon dynamics in the changing ocean. Ocean Sci.6, 605–614 (2010). [Google Scholar]

[CR26] 26.Williams, R. G., Ceppi, P., Roussenov, V., Katavouta, A. & Meijers, A. J. S. The role of the Southern Ocean in the global climate response to carbon emissions. Philos. Trans. R. Soc. A Math. Phys. Eng. Sci.381, 20220062 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR27] 27.Couespel, D. & Tjiputra, J. What goes in must come out: the oceanic outgassing of anthropogenic carbon. Env. Res. Lett.19, 014086 (2024). [Google Scholar]

[CR28] 28.Pérez, F. F. et al. An assessment of CO₂ storage and sea-air fluxes for the Atlantic Ocean and Mediterranean Sea between 1985 and 2018. Glob. Biogeochem. Cycles38, 2023–007862 (2024). [Google Scholar]

[CR29] 29.Maier-Reimer, E., Mikolajewicz, U. & Winguth, A. Future ocean uptake of CO₂: interaction between ocean circulation and biology. Clim. Dyn.12, 711–722 (1996). [Google Scholar]

[CR30] 30.Schwinger, J. et al. Amplification of global warming through pH dependence of DMS production simulated with a fully coupled Earth system model. Biogeosciences14, 3633–3648 (2017). [Google Scholar]

[CR31] 31.Schuur, E. A. G. et al. Climate change and the permafrost carbon feedback. Nature520, 171–179 (2015). [DOI] [PubMed] [Google Scholar]

[CR32] 32.Canadell, J. G. et al. in Climate Change 2021: The Physical Science Basis. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA (2021).

[CR33] 33.Rodgers, K. B. et al. Seasonal variability of the surface ocean carbon cycle: A synthesis. Glob. Biogeochem. Cycles37, 2023–007798 (2023). [Google Scholar]

[CR34] 34.Kwiatkowski, L. et al. Twenty-first century ocean warming, acidification, deoxygenation, and upper-ocean nutrient and primary production decline from CMIP6 model projections. Biogeosciences17, 3439–3470 (2020). [Google Scholar]

[CR35] 35.Frenger, I. et al. Misconceptions of the marine biological carbon pump in a changing climate: Thinking outside the “export” box. Glob. Change Biol.30, 17124 (2024). [DOI] [PubMed] [Google Scholar]

[CR36] 36.Tagliabue, A. ‘Oceans are hugely complex’: modelling marine microbes is key to climate forecasts. Nature623, 250–252 (2023). [DOI] [PubMed] [Google Scholar]

[CR37] 37.Eyring, V. et al. Overview of the Coupled Model Intercomparison Project phase 6 (CMIP6) experimental design and organization. Geosci. Model Dev.9, 1937–1958 (2016). [Google Scholar]

[CR38] 38.Forster, P.et al. The Earth’s Energy Budget, Climate Feedbacks, and Climate Sensitivity. in Climate Change 2021: The Physical Science Basis. Cambridge University Press, Cambridge, UK and New York, NY, USA (2019).

[CR39] 39.Séférian, R. et al. Tracking improvement in simulated marine biogeochemistry between CMIP5 and CMIP6. Curr. Clim. Change Rep.6, 95–119 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR40] 40.O’Neill, B. C. et al. The Scenario Model Intercomparison Project (ScenarioMIP) for CMIP6. Geosci. Model Dev.9, 3461–3482 (2016). [Google Scholar]

[CR41] 41.Silvy, Y. et al. AERA-MIP: Emission pathways, remaining budgets and carbon cycle dynamics compatible with 1.5 and 2 °C global warming stabilization. Earth Syst. Dyn.15, 1591–1628 (2024). [Google Scholar]

PERMALINK

Marine ecosystem role in setting up preindustrial and future climate

Jerry F Tjiputra

Damien Couespel

Richard Sanders

Abstract

Introduction