Response to reviewer #1

This manuscript explores the ocean biogeochemical consequence of extended emissions scenarios up to 2500 using the MRI Earth system model. The analysis presented is robust, the manuscript is well-written, well referenced and adds an interesting long-term perspective to past publications which have overwhelmingly concentrated on 21st century perturbations. My comments are minor, typically focusing on text which could be clarified and additional references which could enhance the discussion in certain areas. Subject to these minor recommendations, I wholeheartedly support publication.

We thank the reviewer for the positive assessment of our study and for the many constructive suggestions.

Several comments highlighted the need for clearer explanations of the mechanisms underlying long-term changes in alkalinity, buffering capacity, and carbon uptake. In response, we have revised the manuscript to clarify the respective roles of physical and biogeochemical processes, particularly regarding alkalinity decline, nutrient redistribution, and carbonate chemistry changes.

We have also expanded the discussion by incorporating the suggested literature (e.g., Koven et al., 2022; Wilson et al., 2022; Terhaar et al., 2022), and by explicitly discussing model limitations associated with the absence of sediment carbonate feedbacks and ice-sheet freshwater forcing.

In addition, we improved figure references, clarified terminology (e.g., surface ocean vs. ocean), and revised several passages where the causal interpretation was unclear.

We believe these revisions have strengthened both the mechanistic interpretation and the broader context of our findings.

L15 Would be nice to mention the driver(s) of alkalinity declines.

We agree that briefly identifying the drivers of the simulated surface-ocean alkalinity decline would improve the clarity of the Abstract. We will therefore revise the text to indicate that enhanced upper-ocean stratification and a slowdown of the deep overturning circulation reduce the resupply of alkalinity-rich subsurface waters, leading to the simulated decline in surface-ocean alkalinity. The revised text will read:

"In parallel, enhanced upper-ocean stratification and a slowdown of the deep overturning circulation reduce the resupply of alkalinity-rich subsurface waters, leading to a decline in surface-ocean alkalinity that further enhances the influence of alkalinity on surface pCO2 during this period."

L65 I would rephrase this to make it clear the stratification and not the reduced buffering capacity that reduces alkalinity.

We have revised the sentence to clarify that enhanced upper-ocean stratification, rather than reduced buffering capacity, causes the decline in surface-ocean alkalinity. The revised sentence will read:

"Subsequent work shows that enhanced upper-ocean stratification reduces the upward transport of alkalinity-rich waters, leading to declining surface-ocean alkalinity and a subsequent reduction in carbon-system buffering capacity, thereby slowing CO2 uptake in later centuries."

L68 Maybe explain why this is the case (computing costs) and that EMICs often do run this far but with potentially less biogeochemical realism.

We agree that providing additional context on the limited availability of simulations extending to 2500 would improve the Introduction. We will therefore revise the text to explain that such simulations remain relatively uncommon with comprehensive Earth system models because of their high computational cost, while Earth system models of intermediate complexity (EMICs) can often be integrated over these timescales but generally employ less realistic representations of marine biogeochemical processes. The revised text will read:

"Yet most studies do not go beyond 2100 because multi-centennial simulations with comprehensive Earth system models remain computationally expensive, and the mechanisms shaping centennial-to-millennial responses (2100–2500) therefore remain poorly constrained. Although Earth system models of intermediate complexity (EMICs) have been integrated over such timescales, they generally employ simplified representations of marine biogeochemical processes. “

L102-104 Perhaps complement this with the model TCRE if known.

We agree that reporting the transient climate response to cumulative carbon emissions (TCRE) would provide additional context for interpreting the climate response of MIROC-ES2L. We will therefore revise the text to include the model's TCRE together with a brief comparison with the likely range assessed by the IPCC. The revised text will read:

"The equilibrium climate sensitivity (ECS) of this version of the model is 2.66 K, diagnosed using Gregory's method (Gregory et al., 2004) and previously reported (Tsutsui et al., 2020; Ohgaito et al. 2021). The transient climate response to cumulative carbon emissions (TCRE) of MIROC-ES2L is 1.3 °C per 1000 PgC (MacDougall et al., 2020; Hajima et al., 2020), which lies within the likely range assessed by the IPCC. Relative to the ECS range of the CMIP6 models, the ECS of MIROC-ES2L is relatively modest."

L122-124 I think a citation or reference to a figure is needed to support this statement.

We agree that this general statement should be supported by appropriate references. We will therefore revise the text to cite previous review studies on climate tipping elements and irreversible climate-system responses. The revised text will read:

"Additionally, changes in GMST levels are associated with non-linear and potentially irreversible responses in the climate system, such as tipping points in sea-ice extent, abrupt changes in El Niño–Southern Oscillation (ENSO) behavior, cloud regimes, and polar climate dynamics, as well as irreversibility in the Atlantic Meridional Overturning Circulation (AMOC) (Lenton et al., 2008; Weijer et al. 2019); Armstrong McKay et al., 2022.)"

L140-149 I think this excessive detail given this is quite fundamental knowledge.

We agree that the original description was unnecessarily detailed. We will therefore shorten this section by removing the explanatory discussion of the Revelle factor and its typical values, while retaining only the definitions of the Revelle factor and the buffering-capacity metrics used in the subsequent analysis.

L179-180 The logic here is confusing. Wouldn’t declining nutrients in deep water formation regions mean greater upper ocean nutrient availability (less loss to depth).

We thank the reviewer for identifying this unclear explanation. We agree that the original wording could be interpreted as implying that reduced nutrient transport to the deep ocean would increase nutrient availability in the upper ocean. Our intention was instead to describe the reduced resupply of nutrients to the surface ocean associated with a weakening of the AMOC. We will revise the text to clarify this mechanism and add a figure illustrating the relationship between the AMOC and marine productivity (NPP and export production) to support this interpretation. The revised text will read:

"The temporal evolution of detritus export is influenced by a complex interplay of environmental factors, including temperature, light availability, and nutrient concentrations, while the basin-scale distributions of these variables are reshaped by changes in ocean circulation. In particular, a weakening of the AMOC (Fig. A5) reduces the upward resupply of nutrients to the North Atlantic surface ocean, leading to lower surface macronutrient concentrations (Fig. A4) and a suppression of detritus export (Fig. xxx)"

Figure R1. Scatter plots showing the relationships between atmospheric CO2 concentration and (a) global export production (Pg C yr⁻¹) for all emission scenarios analyzed in this study, and (b) Atlantic Meridional Overturning Circulation (AMOC) strength (Sv) under the SSP1–2.6, SSP2–4.5, and SSP5–8.5 scenarios.  

L202-203 Is this confirmed by simulated fluxes? Presumably so if changes in winds are minimal.

We confirmed that the simulated air-sea CO2 fluxes are consistent with the interpretation based on (ΔpCO2), showing an intensification of ocean CO2 uptake until approximately 2080 followed by a subsequent decline. We will revise the manuscript to explicitly state this and add a figure showing the simulated air–sea CO2 fluxes. We will also examine changes in 10 m wind speed to clarify the possible contribution of gas-exchange changes. The revised text will read:

"This interpretation is supported by the simulated air–sea CO2 fluxes, which show that ocean CO2 uptake intensifies until approximately 2080 before gradually weakening thereafter. Although regional changes in 10 m wind speed occur, particularly over the Southern Ocean, the temporal evolution of the area-mean air–sea CO2 flux is consistent with that inferred from ΔpCO2."

Figure R2. Sea-surface distributions of (a) CO2 flux and (b) ΔpCO2 for the 2040s, 2090s, 2240s, and 2490s under the SSP5-8.5 scenario, shown as differences relative to the 2010s under the SSP1-2.6 scenario. Panels (c) show the 10 m wind speed under the SSP5-8.5 scenario for the same four periods, and panels (d) show the corresponding differences in 10 m wind speed relative to the 2010s under the SSP1-2.6 scenario.

L210 Does the Southern Ocean therefore remain a sink? Interestingly under concentration-driven overshoot scenarios the Southern Ocean tends to switch form sink to source in multi-model assessments (Koven et al., 2022). This difference in model behavior should probably be discussed.

We thank the reviewer for drawing our attention to Koven et al. (2022). We compared our results with the multi-model assessment presented in that study and found that our simulated Southern Ocean response is broadly consistent with their results. Specifically, the transition from a carbon sink to a carbon source was primarily reported for concentration-driven overshoot scenarios, whereas the Southern Ocean generally remained a carbon sink under high-emission scenarios without declining atmospheric CO2. We will therefore revise the manuscript to clarify this distinction and discuss how differences in the temporal evolution of atmospheric CO2 concentrations between overshoot and continuously increasing forcing scenarios influence the long-term Southern Ocean carbon balance. The revised text will read:

"Despite these widespread increases in ΔpCO2 elsewhere, the Southern Ocean (south of 30°S) consistently maintains negative ΔpCO2 until the 25th century under SSP5-8.5. Accordingly, the Southern Ocean continues to act as a carbon sink throughout the simulation. This behavior is broadly consistent with the multi-model results of Koven et al. (2022), in which a transition from a carbon sink to a carbon source occurred primarily under concentration-driven overshoot scenarios, whereas the Southern Ocean generally remained a carbon sink under high-emission scenarios without declining atmospheric CO2. This suggests that the long-term Southern Ocean carbon balance depends not only on the magnitude of anthropogenic forcing but also on the temporal evolution of atmospheric CO2 concentrations."

L230 Should the p in pCO2 be capitalized here to represent potential pCO2?

We will revise the notation by changing "pCO2" to "PCO2" where appropriate.

L237 Rephrase “neutralizing increases in DIC”. Should DIC be “acidity”?

We agree that the original wording was imprecise. We will revise the text to clarify that seawater carbonate chemistry buffers the increase in acidity associated with elevated DIC, rather than neutralizing the increase in DIC itself. The revised text will read:

"The ocean plays a buffering role in the global carbon cycle by absorbing CO2 and, through seawater carbonate chemistry, buffering the increase in acidity associated with elevated DIC, thereby mitigating the accumulation of CO2 in the atmosphere."

L242 Although closely related, I don’t think this is exactly the inverse as it has units (umol kg-1/uatm) unlike 1/R.

We agree that the original wording was not strictly correct because the buffering coefficient retains physical units and is therefore not mathematically equivalent to the inverse of the Revelle factor. We will revise the text to describe the buffering coefficient as a closely related metric rather than the inverse of the Revelle factor. The revised text will read:

"To assess the buffering capacity of the surface ocean carbon system, we examine the long-term evolution of the Revelle factor (∂ln pCO2ocn / ∂ln DIC), a widely used indicator of the ocean's resistance to changes in CO2. We focus primarily, however, on the buffering coefficient (∂DIC / ∂pCO2ocn), which is closely related to the Revelle factor and provides a more direct measure of the ocean's capacity to absorb additional CO2."

L246 Here and elsewhere be explicit when you are referring to the surface ocean.

We agree that the distinction was not always sufficiently clear. We will revise the manuscript throughout to explicitly refer to the surface ocean where appropriate and improve the consistency of the terminology.

L246-255 One wonders what are the respective roles of DIC vs. alkalinity in the differences in buffering capacity recovery. Presumably this is nearly all DIC driven?

We agree that the respective roles of DIC and alkalinity in the recovery of buffering capacity should be clarified. We will revise the discussion to explicitly state that the recovery of buffering capacity is primarily associated with the evolution of surface-ocean DIC, while changes in surface-ocean alkalinity provide a secondary contribution. The revised text will also refer readers to the subsequent analysis, where the relative influences of DIC and alkalinity on carbonate-system buffering are examined in more detail.

“... The carbonate system therefore gradually re-equilibrates. As discussed in a later section, this recovery is primarily associated with changes in surface-ocean DIC, while changes in surface-ocean alkalinity further modulate the recovery.”

L255 Maybe mention that there is still some regional variability in buffering capacity in SSP126 but this disappears in SSP585.

We agree that the contrasting evolution of the spatial variability in buffering capacity between SSP1–2.6 and SSP5–8.5 is an important feature of the simulations. We will revise the manuscript to explicitly note that regional variability persists under SSP1–2.6, whereas under SSP5–8.5 the much larger accumulation of anthropogenic carbon causes buffering capacity to approach its effective minimum across most of the global surface ocean, thereby largely eliminating regional differences. The revised text will read:

”...This indicates a state in which carbonate-system buffering is strongly diminished as continued accumulation of anthropogenic carbon drives buffering capacity toward its effective minimum across most of the global surface ocean. As buffering capacity approaches this effective minimum, regional contrasts become progressively weaker, resulting in a nearly homogeneous spatial distribution by around 2200.”

L270 I think “in contrast” should be “similarly”.

We agree that "Similarly" is more appropriate than "In contrast" here.

L268-277 Does this surface ocean alkalinity decline occur despite reductions in PIC export? i.e. is physics dominating biotic effects? It is worth mentioning here or later in the discussion on model limitations whether sediment carbonate dissolution is possible (presumably not). On these timescales, one might expect a non-negligible benthic alkalinity flux associated with such dissolution.

We agree that the relative roles of physical and biogeochemical processes in driving the simulated surface-ocean alkalinity decline should be clarified. We will therefore revise the Results section to explicitly note that, although calcium carbonate export decreases in the simulations, surface-ocean alkalinity nevertheless declines because enhanced upper-ocean stratification and a weakening of the overturning circulation reduce the resupply of alkalinity-rich subsurface waters to the surface ocean, overwhelming the alkalinity-conserving effect of reduced carbonate export. The revised text will read:

"After 2200, however, the influence of reductions in surface alkalinity, together with thermal effects, becomes increasingly important in driving further increases in surface pCO2. This influence is associated with enhanced upper-ocean stratification and a weakening of the meridional overturning circulation thereafter. Although calcium carbonate export decreases during this period, surface-ocean alkalinity nevertheless declines, indicating that these physical changes dominate over the alkalinity-conserving effect associated with reduced carbonate export."

We will also revise the "Future challenges and uncertainty" section to explicitly discuss the absence of sediment carbonate dissolution and benthic alkalinity feedbacks in the model as a limitation. The added text will read:

"Another limitation of the present model framework is the omission of sediment carbonate dissolution and benthic alkalinity feedbacks. Although these processes are expected to become increasingly important on centennial-to-millennial timescales, they are not represented in the present model. Sediment carbonate dissolution supplies alkalinity to the ocean and can partially offset ocean acidification and the long-term decline in surface-ocean buffering capacity. Consequently, the simulated reduction in surface-ocean alkalinity and buffering capacity may be somewhat overestimated on multi-centennial timescales."

L290-291 Does export production decline though, as seen in most CMIP6 models and therefore is it remineralization associated with increasing water mass age which is dominating ? (e.g. Wilson et al., 2022).

We thank the reviewer for this insightful comment and for drawing our attention to Wilson et al. (2022). We agree that biological export generally decreases under the high-emission scenario, consistent with previous CMIP6 studies. To clarify the mechanisms responsible for the increase in interior DIC, we will revise the manuscript to distinguish the contributions of preformed and regenerated carbon. We will also add the temporal evolution of preformed and regenerated carbon to Fig. 2. The revised text will read:

"Although export production generally decreases under SSP5–8.5 (Fig. 5), regenerated carbon nevertheless increases because longer water-mass residence times enhance the accumulation of remineralized carbon in the ocean interior. The increase in interior DIC is, however, dominated by preformed carbon, with regenerated carbon providing an additional contribution (Fig. 2)."

Figure R3. Time evolution of changes in ocean carbon storage (PgC) under the SSP1–2.6, SSP2–4.5, and SSP5–8.5 scenarios for the (a) 0–200 m, (b) 200–1000 m, (c) 1000–6000 m, and (d) 0–6000 m depth ranges. Solid lines denote total carbon storage, dashed lines denote preformed carbon storage, and dotted lines denote regenerated carbon storage. Regenerated carbon was estimated from apparent oxygen utilization, and preformed carbon was obtained as the residual.

L322 Are undersaturation metrics with respect to the surface ocean only? Clarify this.

Aragonite and calcite undersaturation are calculated for the full ocean, whereas the discussion in this section focuses on changes in the surface ocean because they are most directly relevant to ecosystem stress. We will revise the manuscript to clarify this distinction. The revised text will read:

"This intensification is accompanied by pronounced surface-ocean acidification on a global scale."

L379 Maybe the relative pCO2atm decline with respect to increase should be mentioned here. This relative decline is much greater in SSP126 than SSP585

We agree that the contrasting evolution of atmospheric pCO2 between the two scenarios is important for interpreting the differences in long-term ocean CO2 uptake. We will revise the manuscript to explicitly note that atmospheric pCO2 stabilizes and declines much more rapidly under SSP1–2.6 than under SSP5–8.5, leading to a more rapid reduction in the ocean–atmosphere pCO2 gradient. The revised text will read:

"As pCO2atm stabilizes and declines much more rapidly under SSP1–2.6 than under SSP5–8.5, the negative ocean–atmosphere pCO2 difference progressively decreases, indicating a more rapid reduction in the chemical potential for further ocean CO2 uptake."

L380-381 Would be useful to compare with Koven et al. (2022) here.

We agree that comparison with Koven et al. (2022) provides useful context for interpreting the long-term evolution of the Southern Ocean carbon sink. We will revise the manuscript to compare our results with their multi-model assessment and clarify that the persistence of the Southern Ocean carbon sink in our SSP5–8.5 simulation is consistent with their results for non-overshoot high-emission scenarios. The revised text will read:

"The Southern Ocean continues to act as a major sink, consistent with the multi-model assessment of Koven et al. (2022), in which the Southern Ocean generally remained a carbon sink under high-emission scenarios without declining pCO2atm."

L383 This wording “as expressed as a decline...” doesn’t work in light of the previous sentence.

We agree that the original wording was awkward and did not connect properly with the preceding sentence. We will revise the text to improve the logical flow and explicitly relate the decline in buffering capacity to the increase in the Revelle factor. The revised text will read:

"As anthropogenic carbon accumulates, the buffering capacity of the carbonate system declines, particularly under SSP5–8.5. This decline corresponds to an increase in the Revelle factor, indicating heightened sensitivity of surface pCO2ocn to additional increases in DIC."

L389 and 391 Can you say “reductions” instead of “changes”?

We agree that “reductions” more accurately describes the behavior discussed in this section. The text has been revised accordingly.

L398-399 It might be worth reelecting on biogeochemical drivers of alkalinity anomalies here.

We agree that the discussion would benefit from considering the biogeochemical drivers of the simulated alkalinity anomalies in addition to the physical mechanisms. We will revise the manuscript to note that changes in calcium carbonate export may also influence surface-ocean alkalinity, but that our results indicate the dominant control is exerted by changes in ocean circulation and upper-ocean stratification. The revised text will read:

"Although changes in calcium carbonate export may also contribute to regional alkalinity anomalies, the present simulations suggest that the dominant driver of the long-term surface-ocean alkalinity decline is the reduced resupply of alkalinity-rich subsurface waters associated with enhanced upper-ocean stratification and changes in overturning circulation. However, the circulation pathways..."

Figure R4. Differences in alkalinity between the SSP5–8.5 scenario (2040s, 2090s, 2240s, and 2490s) and the SSP1–2.6 2010s. Panels show Atlantic meridional sections, Southern Ocean zonal sections, and Pacific meridional sections.

L415 Not sure it makes sense to call the carbonate system CO2 dominated as CO2 will still be <<10% of DIC with nearly everything presumably bicarbonate.

We agree that the original wording was imprecise and could be interpreted as implying that dissolved CO2 becomes the dominant DIC species. We will revise the manuscript to avoid this implication and instead describe the transition in terms of the declining buffering capacity of the carbonate system. The revised text will read:

"This convergence marks a transition toward a weakly buffered carbonate-system regime, in which further increases in DIC generate disproportionately large increases in surface pCO2ocn. As buffering weakens, ..."

L424 Could also mention the absence of sediment feedbacks here.

We agree that the absence of sediment carbonate dissolution and associated benthic alkalinity feedbacks should also be noted in this context. We will revise the manuscript to explicitly identify this process as an additional limitation of the model and discuss its potential implications for long-term projections. The revised text will read:

"... changes in CaCO3 production and dissolution. In addition, the model does not include sediment carbonate dissolution or associated benthic alkalinity feedbacks, which could partially offset surface-ocean alkalinity declines and modify the long-term evolution of ocean buffering capacity and carbon uptake on multi-centennial timescales. Recent work suggests that ..."

L444-445 This is a little confusing. Please clarify what is meant by this decoupling.

We agree that the term "decoupling" was not sufficiently clear in the original manuscript. We will revise the text to explicitly explain that, despite continued accumulation of carbon in the deep ocean, deep-ocean PCO2 remains well below atmospheric pCO2 because deep-ocean ventilation cannot keep pace with the rapid rise in atmospheric CO2. Consequently, the deep ocean retains substantial capacity for further carbon uptake. The revised text will read:

"This reflects the rapid rise in atmospheric pCO2 outpacing deep-ocean ventilation, such that the deep ocean remains far from chemical equilibrium with the atmosphere despite continued DIC accumulation. Consequently, the deep ocean retains substantial capacity for further CO2 uptake, highlighting that carbon storage and chemical equilibration proceed on different timescales under sustained high emissions."

L447 “Ice sheet derived” freshwater forcing? What follows this is also a bit confusing. Is this a consequence of enhanced remineralization in these older waters?

We agree that the original wording was not sufficiently clear. The freshwater forcing referred to here is specifically freshwater input from ice-sheet melt, which is not included in the present simulations. We will revise the manuscript to make this explicit and clarify that the expected enhancement of deep-ocean carbon storage arises primarily from reduced AABW formation and weakened deep-ocean ventilation, rather than from enhanced remineralization associated with older water masses. The revised text will read:

"Freshwater input from ice-sheet melt, which is not included in the present simulations, is expected to amplify the long-term impact of anthropogenic carbon storage on seawater chemistry by reducing AABW formation and deep-ocean ventilation. The resulting increase in deep-water residence time would prolong the isolation of carbon-rich waters from the atmosphere, reinforcing deep-ocean acidification (Lago and England, 2019; Chen et al., 2023)."

L452-455 I’m not sure I have seen robust increases in export even under high mitigation scenarios in multi-model comparisons (e.g. Wilson et al., 2022). The novelty in these projections should be highlighted.

We thank the reviewer for this insightful comment and for referring us to Wilson et al. (2022). We agree that the projected increase in export production under the low-emission scenarios differs from the general behavior reported in previous CMIP6 multi-model assessments, and that the novelty of this result should be highlighted more clearly. We will revise the discussion to emphasize that the long-term simulations reveal a transient response that is not apparent from end-of-century analyses alone. In particular, export production responds rapidly during the 21st century and closely follows the evolution of the AMOC. The transition from increasing to decreasing export occurs once atmospheric CO2 exceeds approximately three times the pre-industrial level, when the weakening of the AMOC becomes more pronounced under stronger forcing. We are also considering adding a figure showing the temporal evolution of export production together with AMOC strength to illustrate this relationship more clearly. The revised text will read:

"Simulation results for detritus export at the depth of 100 m vary substantially among emission scenarios. Detritus export exhibits a long-term increase under SSP1–2.6 and SSP2–4.5, in contrast to the decline projected under SSP3–7.0 and SSP5–8.5 (Fig.~\ref{fig:fzd100_PgC}). Most of this divergence develops during the 21st century and closely follows the rapid adjustment of the large-scale ocean circulation. Export production declines as the weakening of the AMOC becomes more pronounced under stronger forcing; this transition occurs when atmospheric CO2 exceeds approximately three times the pre-industrial level (Fig. xxx). Consequently, the long-term behavior is largely established by the end of the 21st century and remains comparatively stable thereafter, a feature that is difficult to identify from end-of-century analyses alone. Regionally, detritus export decreases across the high-latitude North Atlantic under all scenarios, with the largest decline occurring under SSP5–8.5, consistent with the strongest weakening of the AMOC."

L468 “they” typo

We have corrected "they" to "their" in the revised manuscript.

L544 Or an overestimation. Greenland hosing simulations have been shown to reduce ocean carbon uptake (e.g. Swingedouw et al., 2007).

We thank the reviewer for this important comment and for drawing our attention to Swingedouw et al. (2007). We agree that the sign of the bias introduced by the omission of ice-sheet freshwater forcing is not straightforward and may depend on the source of the freshwater perturbation and the associated circulation response. We will therefore revise the manuscript to avoid implying a one-sided bias and instead emphasize that the omission of dynamic ice-sheet freshwater forcing represents an important source of uncertainty in long-term projections. The revised text will read:

"... its influence on buoyancy-driven overturning (Rahmstorf et al., 2002). Recent studies have shown that Antarctic ice sheet melt can weaken AABW formation, alter Southern Ocean circulation, and substantially modify ocean carbon uptake and storage (Bronselaer et al., 2018; 2020), whereas freshwater perturbations from the Greenland Ice Sheet have been shown to reduce ocean carbon uptake through their influence on the AMOC (Swingedouw et al., 2007). The omission of dynamic freshwater forcing in the present study therefore represents an important source of uncertainty in projections of long-term ocean carbon storage and acidification. Future modeling efforts ..."

L563-565 This is a feature of carbonate chemistry projections at depth but not in the global surface ocean where there is generally high model agreement for acidification projections with the exception of regions dominated by sea ice dynamics or riverine fluxes. I therefore suspect the impact on ocean carbon uptake to be minimal. That being said, a general overestimation of Revelle factor has been shown to low bias CMIP6 ocean carbon uptake simulations (Terhaar et al., 2022).

We thank the reviewer for this helpful comment and for drawing our attention to Terhaar et al. (2022). We agree that projections of surface-ocean acidification are generally robust across models, except in regions strongly influenced by sea-ice processes or riverine inputs. We will revise the discussion to distinguish the relatively high confidence in surface-ocean acidification projections from the greater uncertainty associated with carbonate-system buffering capacity and its implications for long-term ocean carbon uptake. We will also cite Terhaar et al. (2022), who showed that an overestimation of the Revelle factor can lead to an underestimation of ocean carbon uptake in CMIP6 models. The revised text will read:

"... ocean buffering capacity. Although projections of surface-ocean acidification are generally robust across Earth system models, substantially greater uncertainty remains in the timing and magnitude of changes in carbonate-system buffering capacity and the resulting limits on long-term ocean carbon uptake (Orr et al., 2005; Egleston et al., 2010). In particular, biases in the simulated Revelle factor have been shown to contribute to biases in projected ocean carbon uptake in CMIP6 models (Terhaar et al., 2022). At the same time, ocean acidification poses significant risks to marine ecosystems, particularly calcifying organisms (Doney et al., 2009). "

Response to Reviewer #2

Overall Assessment and Recommendation

I appreciate the overall topic of this study, which addresses an important and timely question:  the multi-centennial response of ocean biogeochemistry and marine ecosystems to anthropogenic CO2 forcing. Long-term projections beyond 2100 remain relatively scarce, and the results presented here have the potential to provide valuable insights into both the future evolution of the ocean carbon sink and the persistence of ecosystem stressors. These are important issues that deserve attention and publication.

However, I find that the current manuscript suffers from significant organizational and conceptual weaknesses. First, the scientific objectives are not clearly defined or sufficiently justified. The Introduction combines several distinct themes—including ocean carbon uptake, low-latitude productivity, nutrient cycling, ocean acidification, and ecosystem stress—without developing a coherent argument that explains why the long-term response (2100–2500) should be investigated and which specific scientific questions are being addressed. As a result, the narrative lacks focus and the reader is left uncertain about the primary goals of the study.

The Results section reflects this lack of focus. It concentrates disproportionately on the evolution of DIC and NPP, while other potentially more impactful aspects of the simulations receive limited attention. More generally, the manuscript attempts to address two major questions—long-term carbon uptake and long-term ecosystem stressors—without clearly separating them or establishing strong links between them.

In its current form, I do not believe that the manuscript is ready for publication in Biogeosciences. I strongly recommend a major reorganization of the paper. The authors should either clearly separate the carbon uptake and ecosystem stress components into distinct parts of the manuscript, each with dedicated objectives and analyses, or alternatively choose to focus on only one of these themes. Such a restructuring would require a substantial rewriting of the Introduction, a clearer formulation of the scientific objectives, and a reconsideration of the results presented in the central sections of the paper. I believe that these changes would significantly improve the clarity, impact, and overall scientific value of the study.

We thank the reviewer for the thoughtful and constructive assessment of our manuscript. We appreciate the recognition that multi-centennial projections of ocean biogeochemistry remain scarce and that such simulations have the potential to provide important insights into the long-term evolution of the ocean carbon sink and marine ecosystems.

We agree that the original manuscript did not sufficiently articulate its central scientific objectives or clearly explain the common scientific question linking the different components of the study. Our objective is to understand how the long-term adjustment of the ocean to anthropogenic CO2 forcing differs among emission pathways, and how these differences shape the evolution of ocean carbon uptake, carbonate chemistry, biological productivity, and ecosystem stress over multi-centennial timescales. Rather than examining each of these processes in isolation, this study aims to understand how these components co-evolve and interact during the long-term adjustment of the ocean to anthropogenic CO2 forcing. We respectfully consider that ocean carbon uptake and ecosystem stress are not independent themes, but complementary aspects of the long-term response of the ocean to anthropogenic CO2 forcing.

In response to the reviewer's comments, we will substantially revise the manuscript. We will reorganize the Introduction to better motivate the scientific importance of simulations extending beyond 2100, explicitly state the scientific questions addressed in this study, and clarify the rationale for investigating long-term ocean biogeochemical responses across multiple emission scenarios. We will also strengthen the motivation for investigating ecosystem stress by introducing recent studies that use Earth system model projections to assess impacts on higher trophic levels (e.g., FISHMIP), thereby clarifying how long-term changes in ocean biogeochemistry provide an important foundation for understanding future marine ecosystem responses. In addition, we will expand the discussion of previous long-term Earth system simulations and more clearly position the present study within this broader context.

We will also reorganize parts of the Results section to better distinguish responses established during the 21st century from slower multi-centennial adjustment processes while emphasizing their common physical and biogeochemical drivers. To strengthen the mechanistic interpretation, we will expand the analyses of biological productivity and ocean carbon storage. In particular, we will distinguish the contributions of preformed and regenerated carbon to long-term carbon sequestration, further examine the transient evolution of net primary production and export production in relation to atmospheric CO2 and changes in the Atlantic Meridional Overturning Circulation, and clarify the regional characteristics of these responses. We will also add new figures, where appropriate, and promote selected supplementary figures to the main text to better illustrate these relationships.

These additional analyses will clarify that many key ocean biogeochemical responses are established primarily during the 21st century, whereas their adjustment continues over subsequent centuries. By comparing multiple emission scenarios, the revised manuscript will more clearly demonstrate how both the magnitude and persistence of these long-term responses depend on the trajectory of anthropogenic forcing, highlighting an aspect that cannot be inferred from conventional end-of-century analyses alone.

Finally, we will expand the Discussion to clarify the physical and biogeochemical mechanisms responsible for the simulated changes, compare our results more comprehensively with previous studies, and explicitly discuss important model limitations and uncertainties, including alkalinity feedbacks, sediment processes, freshwater forcing from ice-sheet melt, and uncertainties in carbonate-system buffering.

We believe that these planned revisions will substantially improve the clarity, organization, and scientific focus of the manuscript while more clearly communicating its central scientific objective of understanding how long-term ocean adjustment differs among emission pathways and the implications of these differences for ocean carbon uptake and marine ecosystems.

Specific Comments

(1) Introduction

The Introduction requires substantial reorganization. As it stands, it interweaves results and concepts related to the evolution of the ocean carbon sink, primary productivity, and ecosystem stressors (e.g., acidification), without clearly distinguishing these components or explaining why they should be considered together, nor how they potentially interact or feedback on one another. Some statements appear overly general and insufficiently supported by the literature; for instance, the claim that “low-latitude productivity is crucial for the global carbon cycle” is too simplistic and would benefit from a more explicit and better substantiated discussion.

In addition, the paragraph starting at line 59 (“Analyses using CMIP models have also advanced understanding of present-day oceanic uptake of anthropogenic CO2”) transitions abruptly from present-day understanding to long-term projections without a clear logical bridge. Given the relatively limited number of studies addressing multi-centennial ocean biogeochemistry, the motivation for focusing on long-term responses needs to be strengthened. In particular, the authors should better justify why centennial-to-millennial timescales are relevant, explicitly discussing ocean adjustment timescales, (ir)reversibility of carbon-cycle changes, and the relevance for net-zero or overshoot scenarios.

We thank the reviewer for this important and constructive comment. We agree that the original Introduction did not sufficiently articulate the overarching motivation of the study or clearly define the scientific questions addressed. We also agree that the connections among long-term ocean carbon uptake, carbonate chemistry, biological productivity, and ecosystem stress were not adequately explained, and that the motivation for extending the analysis beyond 2100 required a stronger justification.

To address these concerns, we will substantially revise the Introduction. We will reorganize the Introduction to more clearly motivate the need for simulations extending beyond 2100 and explicitly state the scientific questions addressed in this study. We will strengthen the motivation by discussing the long adjustment timescales of the ocean and their relevance to future mitigation pathways. We will also clarify why ocean carbon uptake, carbonate chemistry, biological productivity, and ecosystem stress are investigated together by emphasizing that they represent complementary components of the long-term ocean biogeochemical response within a common Earth system framework. Furthermore, we will strengthen the Introduction by incorporating recent studies that link Earth system model projections to higher trophic-level ecosystem impacts (e.g., FISHMIP), thereby providing a clearer motivation for examining long-term ecosystem stress. We will also revise several overly general statements, including the discussion of low-latitude productivity, to provide a more balanced description supported by the relevant literature.

The revised Introduction will conclude with the following paragraph:

"Despite growing understanding of 21st-century ocean change, the long-term evolution of ocean biogeochemistry after atmospheric CO2 stabilization remains poorly constrained. Because ocean circulation, carbon storage, nutrient redistribution, and ecosystem responses operate on timescales ranging from decades to centuries, substantial adjustments may continue long after atmospheric CO2 concentrations cease increasing. Understanding these delayed responses is particularly relevant for evaluating the long-term consequences of mitigation pathways. In this study, we use extended SSP simulations with MIROC-ES2L through the year 2500 to address three questions: (1) How does ocean carbon uptake evolve after atmospheric CO2 stabilization? (2) What processes control the long-term evolution of ocean carbonate chemistry and buffering capacity? (3) How do biological productivity and ecosystem stress evolve on multi-centennial timescales? By addressing these questions within a single Earth system model framework and across multiple emission scenarios, we identify how long-term ocean adjustment depends on the trajectory of anthropogenic forcing and distinguish responses established during the 21st century from slower adjustment processes that continue for centuries thereafter."

(2) Results structure and metrics

The first results section is relatively short and focuses primarily on dissolved inorganic carbon (DIC) and net primary production (NPP)/export production. This choice is not sufficiently justified. In particular, it is unclear why the evolution of ocean carbon uptake itself is not treated as a central diagnostic, given that it is arguably one of the key variables of the study.

We thank the reviewer for this helpful comment. We agree that the central role of ocean carbon uptake was not sufficiently emphasized in the original presentation of the Results. Our intention was to first describe the large-scale redistribution of dissolved inorganic carbon, which provides the physical and biogeochemical context for the subsequent analyses of ocean carbon uptake, carbonate chemistry, and deep-ocean carbon storage.

To make this progression clearer, we will revise the first Results section to more explicitly highlight the evolution of ocean carbon uptake. In particular, we will add the time evolution of the accumulated ocean carbon inventory for each scenario, thereby providing a more direct diagnostic of long-term ocean carbon uptake and complementing the redistribution of dissolved inorganic carbon. We will also revise the subsection title to better reflect that this section describes both the redistribution of dissolved inorganic carbon and the evolution of ocean carbon uptake.

Figure R1. Time evolution of changes in ocean carbon storage (PgC) under the SSP1–2.6, SSP2–4.5, and SSP5–8.5 scenarios for the (a) 0–200 m, (b) 200–1000 m, (c) 1000–6000 m, and (d) 0–6000 m depth ranges. Solid lines denote total carbon storage, dashed lines denote preformed carbon storage, and dotted lines denote regenerated carbon storage. Regenerated carbon was estimated from apparent oxygen utilization, and preformed carbon was obtained as the residual.

(3) NPP response under contrasting scenarios

The evolution of NPP differs markedly between scenarios, with a decrease under high-emission conditions and an increase under low-emission conditions. This is a striking and potentially highly significant result. However, it is not discussed or explained in any depth, which represents a missed opportunity given its implications.

We thank the reviewer for highlighting the importance of this result. We agree that the contrasting responses of net primary production under low- and high-emission scenarios are one of the key findings of this study and deserve further discussion.

In response to this comment, we will substantially expand the corresponding Results and Discussion sections. In particular, we will examine the temporal evolution of net primary production in relation to atmospheric CO2 and changes in the Atlantic Meridional Overturning Circulation, highlighting that much of the divergence among emission scenarios is established during the 21st century and subsequently evolves more gradually over multi-centennial timescales. We will also clarify the regional characteristics of the simulated responses and discuss their consistency with previous studies.

At the same time, we recognize that a complete attribution of the underlying mechanisms would require additional process-oriented diagnostics and dedicated sensitivity experiments to separate the respective roles of physical circulation, nutrient redistribution, and biological feedbacks. Such analyses would require dedicated sensitivity experiments and additional diagnostic calculations to isolate the relative contributions of physical circulation, nutrient redistribution, and biological feedbacks, and are therefore beyond the scope of the present study.

Figure R2. Scatter plots showing the relationships between atmospheric CO2 concentration and (a) global export production (Pg C yr⁻¹) for all emission scenarios analyzed in this study, and (b) Atlantic Meridional Overturning Circulation (AMOC) strength (Sv) under the SSP1–2.6, SSP2–4.5, and SSP5–8.5 scenarios.

(4) Carbon uptake drivers and biological pump

The section describing the drivers of carbon uptake is generally well developed and clearly presented. However, it would be particularly interesting to further explore the role of the biological pump across these long timescales, especially in comparison with the physical pump. In particular, the manuscript could attempt to assess whether biological processes become increasingly important in controlling carbon uptake over centennial to millennial scales.

We thank the reviewer for this valuable suggestion. We agree that the relative roles of the biological and physical pumps in regulating long-term ocean carbon uptake deserve further discussion.

In response to this comment, we will expand the Discussion by explicitly considering the complementary roles of physical circulation and biological processes over multi-centennial timescales. Building on the newly added analysis of preformed and regenerated carbon storage, we will discuss how these diagnostics provide additional insight into the mechanisms controlling long-term carbon sequestration. In particular, we will emphasize that regenerated carbon storage continues to increase even as export production declines under the high-emission scenarios, whereas preformed carbon remains the dominant contributor to the long-term increase in deep-ocean carbon storage. These contrasting responses suggest that long-term regenerated carbon storage is influenced not only by biological export but also by changes in ocean circulation and water-mass residence time.

Accordingly, we will add the following discussion to Section 4.1.3 “Deep-ocean carbon sequestration, circulation pathways, and potential CO2 outgassing”:

"Separating deep-ocean carbon storage into preformed and regenerated components further illustrates the complementary roles of the physical and biological pumps in long-term carbon sequestration. Although export production declines under the high-emission scenarios, regenerated carbon storage continues to increase. This contrasting behavior suggests that long-term regenerated carbon storage is influenced not only by biological export but also by changes in ocean circulation and water-mass residence time. Nevertheless, preformed carbon remains the dominant contributor to the long-term increase in deep-ocean carbon storage. These results suggest that changes in ocean circulation remain a primary control on long-term carbon sequestration, while biological processes continue to influence the redistribution and storage of carbon within the ocean interior."

A complete quantitative attribution of the respective contributions of the physical and biological pumps would require additional process-oriented diagnostics and dedicated sensitivity experiments, and is therefore beyond the scope of the present study.

(5) Ecosystem focus

The analysis of specific ecosystems (e.g., kelp forests, coral reefs) appears underdeveloped and only marginally integrated into the overall narrative. It is not entirely clear that this level of ecosystem-specific detail is essential for the scope of the manuscript, and reconsideration of this section may help improve focus and coherence.

We thank the reviewer for this helpful suggestion. We agree that the role of the ecosystem-specific examples was not sufficiently clear in the original manuscript.

Our intention is not to provide detailed assessments of individual ecosystems, but rather to illustrate how long-term changes in ocean biogeochemistry may translate into persistent ecosystem stress under different emission pathways. To clarify this objective, we will revise both the Introduction and the Results. In the Introduction, we will strengthen the motivation for including ecosystem stress by discussing recent studies that use Earth system model projections to assess impacts on higher trophic levels (e.g., FISHMIP), thereby placing the ecosystem analyses within a broader Earth system context. In the Results, we will reduce descriptive details associated with individual ecosystem types and instead emphasize that coral reefs, kelp forests, and other examples are intended to illustrate broader patterns of long-term ecosystem vulnerability rather than to provide ecosystem-specific assessments.

These revisions will strengthen the connection between the ecosystem analyses and the central objective of the manuscript, namely understanding how long-term ocean biogeochemical adjustment differs among emission pathways and how these differences influence future marine ecosystem stress.