Review of cp-2022-8

Glacial state of the global carbon cycle: time-slice simulations for the last glacial maximum with an Earth-system model

by T. Kurahashi-Nakamura, A. Paul, U. Merkel, and M. Schulz

This manuscript addresses changes in the inventory and distribution of biogeochemical tracers during glacial periods using an Earth system model. The authors present the impacts of glacial-interglacial sea-level changes and associated changes in oceanic inventory of biogeochemical tracers on ocean biogeochemical cycles. The paper is novel in its approach. Of particular significance is their use of a sedimentation model in combination with the ESM to demonstrate the mass accumulation rate (MAR) of carbonate sediments and their discussion of AMOC changes in comparison with MAR observations. On the other hand, although I am familiar with the topic of this paper, I found some of the authors' explanations difficult to follow.

General comments:

In the introduction, the authors emphasized the importance of transiency to explain the glacial-interglacial carbon cycle. I would like to see more discussion of the time scale of the response of the carbon cycle. How fast do carbonate sediments change in response to changes in sea level? How do the time scales of carbonate sediment expansion and contraction differ from other processes that consist of the ocean carbon cycle (solubility, biological pumps, and ocean circulation)?

Although the authors focus on shallow-water carbonate sediments, carbonate is thought to be equally buried in the deep ocean (Cartapanis et al., 2018). I would like to see a clearer separation of these two contributions in the discussion.

In the LGM experiments, whole ocean alkalinity is increased to adjust the atmospheric CO2 concentration to the ice core data. The magnitude of this increase does not seem to be explicitly stated (not shown in Table 2), but is it appropriate? In section 4.1, the authors explained the change in alkalinity is related to the changes in shallow-water coral reefs but is it necessary to consider the effect of carbonate compensation, including deep-sea carbonate sediments (e.g., Brovkin et al., 2012; Ganopolski and Brovkin, 2017; Kobayashi et al., 2021)?

It is reported that PMIP models tend to simulate lower ocean carbon sequestration and higher atmospheric CO2 if they had a lower ocean volume at the LGM (Lhardy et al., 2021).  Do the ocean bathymetry and volume change in this study?　It is expected that a greater change in alkalinity would be required if a lower ocean volume is adopted.

Specific comments:

P1/L6: “The increase…”　I am not convinced by this statement because of the lack of information on the inventory change in alkalinity.

P2/L15: How much carbon (PgC) does the change in DIC in the deep ocean (μmol/kg) between the LGM and the present day correspond to?

P2/L16: I would like to know more about the coral reef hypothesis and the subsequent study's discussion of the impact of changes in shallow-water carbonate sedimentation on global carbon cycle changes.

P2/L20: I would like to know the explicit statement about the impact of changes in DIC and alkalinity on atmospheric CO2 concentrations. In other words, the increase in DIC in the deep ocean during glacial periods and the increase in alkalinity throughout the ocean both contribute to a decrease in atmospheric CO2.

P3/L21: Would you explain a little more about the advantages of using SolveSAPHE?

P3/L30: Changes in carbonate sediment burial result in changes in whole ocean alkalinity (Kobayashi et al., 2021). Is it difficult to investigate the changes in carbonate burial by using the sediment model?

P3/L30: I would like to see information on carbonate burial fluxes and sediment distribution throughout the ocean as calculated by the sediment model in expPI. p7/L8 includes a brief description, but I do not think it is sufficient.

P4/L13: Would you cite a reference for the Ruddiman belt?

P4/L15: Why did you choose 0.25 Sv for the freshwater input to the Southern Ocean?

P4/L34: What is the total change in alkalinity adjusted for "second-guess"? Would you add the information to Table 1?

P5/L14: Would you consider showing the stream function as well?

P5/L19: From my understanding, LGMss was an experiment in which freshwater was removed from the LGM in the North Atlantic and fresh water was added in the Southern Ocean, resulting in a stronger north-south density gradient at the sea surface. However, why do we see a somewhat weaker AMOC relative to the LGM?

P5/26: How about comparing the modeled deep-sea salinity to paleo records (e.g., Adkins et al., 2002; Insua et al., 2014; Homola et al., 2021）?

P6/L9: Can we assume that this change is consistent with estimates of changes in terrestrial carbon storage (e.g., Peterson et al., 2014; Jeltsch-Thömmes et al., 2019)?

P6/L9: Is the total amount of carbon in the atmosphere, ocean, and terrestrial reservoir the same for all experiments? It is difficult for me to understand the experimental design.

P6/L14: Would you show us d13C paleo records (e.g., Peterson et al., 2014) in the figure? It would make comparisons with other studies easier.

P6/L22: The impact of changes in the distribution of export production on nutrients and AOUs should also be considered. How about conducting sensitivity experiments with fixed biological fluxes?

P6/L25: The decrease in ideal age in the Southern Ocean in Fig. 4 is caused by changes in AABW flow or changes in local convective mixing. Please describe.

P6/L26: Some studies have reconstructed carbonate ions from B/Ca (e.g., Rickaby et al., 2010; Yu et al., 2013, 2020). Would you compare your modeling results with them? As stated in the discussion, the increase in alkalinity seems to be overestimated in the current setting.

P6/L31: How about showing the reconstructed changes in export production (Kohfeld et al., 2005)? The characteristics of the changes appear to be well reproduced in the model. Also, I would like to see a discussion of the effects of sea ice distribution and iron fertilization on export production.

P7/L12: Would it make sense to compare this model-data comparison of CaCO3 MAR for other ocean regions (the Southern Ocean and the Pacific Ocean)? It makes the model validity and shortcomings clearer.

P8/L15: The pyrite oxidation showed here as negative feedback on atmospheric CO2, but was insufficiently studied to constrain its quantitative contribution to the glacial-scale carbon cycle. Is my understanding of this correct?

P8/L32: A recent modeling study of Kobayashi et al. (2021) also used d13C to constrain their ocean carbon cycle fields in the LGM.

P9/L32: I asked a similar question about the export production, but does the change in sea ice coverage affect the sinking flux of CaCO3 in this region?

P10/L7 Typo? The references are not listed correctly.

P10/L13: The maximum values of ideal age in the Pacific appear to be getting younger in expPI, expLGM, and expLGMss (expLGMws). Is this related to the increased AABW-related deep-water flow?

References:

Adkins, J. F., McIntyre, K., and Schrag, D. P., The Salinity, temperature, and δ18O of the glacial deep ocean. Science, 298(5599), 176–1773. https://doi.org/10.1126/science.1076252, 2002.

Brovkin, V., Ganopolski, A., Archer, D., and Munhoven, G., Climate of the Past Glacial CO2 cycle as a succession of key physical and biogeochemical processes, Climate of the Past, 8, 251–264, https://doi.org/10.5194/cp-8-251-2012, 2012.

Cartapanis, O., Galbraith, E. D., Bianchi, D., and Jaccard, S. L., Carbon burial in deep-sea sediment and implications for oceanic inventories of carbon and alkalinity over the last glacial cycle, Climate of the Past, 14, 1819–1850, https://doi.org/10.5194/cp-14-1819-2018, 2018.

Ganopolski, A. and Brovkin, V., Simulation of climate, ice sheets and CO2 evolution during the last four glacial cycles with an Earth system model of intermediate complexity, Climate of the Past, 13, 1695–1716, https://doi.org/10.5194/cp-13-1695-2017, 2017.

Homola, K., Spivack, A. J., Murray, R. W., Pockalny, R., D'Hondt, S., and Robinson, R., Deep North Atlantic Last Glacial Maximum salinity reconstruction. Paleoceanography and Paleoclimatology, 36, e2020PA004088. https://doi.org/10.1029/2020PA004088, 2021.

Insua, T. L., Spivack, A. J., Graham, D., D'Hondt, S., & Moran, K., Reconstruction of Pacific Ocean bottom water salinity during the last glacial maximum. Geophysical Research Letters, 41, 2914–2920. https://doi.org/10.1002/2014gl059575, 2014

Jeltsch-Thömmes, A., Battaglia, G., Cartapanis, O., Jaccard, S. L., Joos, F., Low terrestrial carbon storage at the Last Glacial Maximum: Constraints from multi-proxy data, Climate of the Past 15, 849–879, https://doi.org/10.5194/cp-15-849-2019, 2019.

Kobayashi, H., Oka, A., Yamamoto, A. & Abe-Ouchi, A., Glacial carbon cycle changes by Southern Ocean processes with sedimentary amplification, Science Advances, 7, eqbg7723, https://doi.org/10.1126/sciadv.abg7723, 2021.

Kohfeld, K. E., Le Quéré, C., Harrison, S. P., Anderson, R. F., Role of marine biology

in glacial-interglacial CO2 cycles, Science, 308, 74–78, https://doi.org/10.1126/science.1105375, 2005.

Lhardy, F., Bouttes, N., Roche, D. M., Abe-Ouchi, A., Chase, Z., Crichton, K. A., et al., A first intercomparison of the simulated LGM carbon results within PMIP-carbon: Role of the ocean boundary conditions. Paleoceanography and Paleoclimatology, 36, e2021PA004302, https://doi.org/10.1029/2021PA004302, 2021.

Peterson, C. D., Lisiecki, L. E., Stern, J. V., Deglacial whole-ocean d13C change estimated　from 480 benthic foraminiferal records. Paleoceanography 29, 549–563, https://doi.org/10.1002/2013PA002552, 2014.

Rickaby, R. E. M., Elderfield, H., Roberts, N., Hillenbrand, C. -D., and Mackensen, A., Evidence for elevated alkalinity in the glacial Southern Ocean, Paleoceanography, 25, PA1209, https://doi.org/10.1029/2009PA001762, 2010.

Yu J., Anderson, R. F., Jin, Z., Rae, J. W. B., Opdyke, B. N., Eggins, S. M., Responses of the deep ocean carbonate system to carbon reorganization during the Last Glacialeinterglacial cycle, Quaternary Science Reviews, 76, 39–52, http://dx.doi.org/10.1016/j.quascirev.2013.06.020, 2013.

Yu, J., Menviel, L., Jin, Z. D., Anderson, R. F., Jian, Z., Piotrowski, A. M. et al., Last glacial atmospheric CO2 decline due to widespread Pacific deep-water expansion, Nature Geoscience, 13, 628–633, https://doi.org/10.1038/s41561-020-0610-5, 2020.

Kurahashi-Nakamura et al. present three time slice simulations of the LGM performed with the carbon isotope-enabled CESM1.2 fully-coupled GCM, that are used to investigate the global carbon cycle. The authors further investigate the sedimentary carbonate chemistry with an offline sediment model. The three experiments have in common that the artificial DIC and alkalinity additions have been tuned to achieve pCO2 concentrations and global mean DIC concentrations in agreement with reconstructions. These changes in inventories are then discussed in the context of potential processes that could have provided/removed the required budgets during the deglaciation. Yet, the three LGM runs differ in their physical ocean state which has been achieved through additions of idealized freshwater/salt fluxes in the North Atlantic and Southern Ocean. The authors discuss the effects of the different AMOC states on the distributions of biogeochemical tracers and compare their results to previous studies. The manuscript is generally well-written and illustrated, but I feel that the analysis could go into more depth with additional figures better showing for instance the model-data comparisons. As such, I believe the manuscript would benefit from some revisions that I outline below in detail.

Major points

1. In the introduction the authors emphasize the importance of simulations covering the entire glacial cycle to also capture the effects of slow processes. While it is clear that this is currently not possible due to the prohibitive computational costs, it would be interesting to discuss what effects and impacts would be expected from these slow processes and whether they could bias the results of the present study.

2. While by design of the experiments the largest changes are expected to be in the Atlantic, there are surely also important differences in the physical ocean states and biogeochemistry of the rest of the ocean. However, this is neither discussed nor shown in any of the figures.

3. In the two shallow LGM runs (LGMsw and LGMss) large changes in phosphate and carbonate ion concentrations and AOU exist. The authors argue that this is either related to the more sluggish deep ocean ventilation or the biological carbon pump. Yet, the ideal age tracer distributions, that in fact indicate younger bottom water in the entire Atlantic, and the stronger stream function in the deep Atlantic strongly suggest that this was only caused by the more efficient biological carbon pump in the Southern Ocean. Due to the importance of this process and the far-reaching effects I would like to see a more in-depth discussion and analysis of this matter.

4. In section 4.1 the authors mention that the applied alkalinity changes are in good agreement with previous estimates for carbonate deposition during the deglaciation. However, in section 4.2 it is then mentioned that the [CaCO3] are systematically too high most likely due to the uniformly increased alkalinity. How can this be reconciled?

5. Section 4.2 encompasses many comparisons of the LGM experiments to reconstructions of various parameters. However, I feel that there is a missed opportunity by not visualizing these results to a greater extent (currently only the CaCO3 MAR model-data comparison is shown). One could for instance show the LGM d13C and nutrient data by Oppo et al. (2018) in Figure 3. Further, the [CaCO3] gradients discussed in the text could be plotted against the reconstructions. This would surely help to better demonstrate where the model performs well and where there are biases.

6. The authors try to assess the validity of the DIC – ventilation age relationship used by Sarnthein et al. (2013) and Skinner et al. (2015) and come to the conclusion that it does not hold for the LGM. However, one has to note that the previous studies by Sarnthein et al. (2013) and Skinner et al. (2015) used radiocarbon ventilation ages while in the present study the ideal age of the model was used for the assessment. In this context it is noteworthy that the ideal age and the radiocarbon ventilation age behave quite differently in the (model) ocean, mostly due to the additional effect of limited air-sea gas exchange under sea-ice for radiocarbon that the ideal age tracer does not see. This effect should be much stronger for the LGM simulations than the PI due to the colder temperatures and hence larger sea ice extent. It is therefore possible (or even rather likely) that the radiocarbon ventilation age is much older in the LGM simulations than in PI while the ideal age is younger in the global mean and the previously proposed relationship still holds. The DIC – age relationship should therefore be reassessed with respect to this issue.

Minor points

P1, L3-5: This sentence is slightly confusing, considering that you simulated time slices but here argue with the evolution of the reservoirs.

P2, L2: The penultimate interglacial was MIS 7. Do you instead mean the last interglacial (i.e., the Eemian)?

P2, L5: Orbital configuration not orbital elements.

P2, L24: As far as I’m aware there are no reconstructed concentrations of DIC.

P3, L20: Do you mean that POM was fully remineralized in the bottommost cells?

P4, L8: Can you give a percentage for the adjusted mean salinity and nutrient concentrations.

P4, L10: Was the 2.5 kyr spinup enough to reach equilibrium?

P4, L14: The Ruddiman Belt is defined by the deposition of ice rafted debris during Heinrich Events. The reference to this could therefore lead to confusion. Instead, better simply refer to the latitudinal band where the freshwater was applied.

P4, L14: Since the freshwater addition was compensated for, I would try to avoid the word “hosing”.

P5, L1: Here you mention that the atmospheric d13C signature was prescribed. Isn’t this in conflict with freely evolving atmospheric CO2 concentrations in terms of the 13C budget?

P5, L15: It’s Atlantic Meridional Overturning Circulation not ocean circulation.

P5, L17-18: The zero isoline of the stream function is not equivalent to the separation of AABW and NADW as can be seen from dye experiments (e.g., for CESM: Gu et al., 2020).

P5, L21: To me, it appears from Figure 2 that LGMws does not have a stronger bottom circulation than LGM or PI.

P5. L26: Directly inferring from roughly correct SST changes the correct atmosphere-ocean partitioning of CO2 is quite a stretch. This completely ignores the other carbon pumps that also play a role in the partitioning.

P5, L27: This is in conflict with Figure 3c. However, I suspect that something went wrong in Figure 3c.

P5, L31: “In expPI, we obtained 276 ppm”. Please rephrase and expand this sentence.

P6, L1: The model is surely tuned to this PI pCO2, I therefore think that this is not necessarily an indication of the models “excellent ability” to predict pCO2.

P6, L11: Typo “relfected”.

P6, L18: Please try to avoid the word “observed” when talking about model results, as it suggests that the finding is derived from observations.

P6, L25: But the very deep is younger than PI not older and the stream function (Fig. 2) indicates stronger or equal advection in the deep of all LGM runs compared to PI. How does this fit together? Was the longer-lasting storage of organic matte only at the mid-depth between 2 and 3 km? Why is the phosphate concentration also elevated below 3 km? Have you diagnosed the remineralized phosphate fraction from the model?

P9, L13-14: As mentioned before, from the ideal age it is clear that the bottom water was in fact not more stagnant for all LGM simulations.

Figure 3c: The distribution looks rather strange compared to the other experiments and other tracers. Please double-check.

Figure 6: Most of the map is white. Does that mean that in ~80% of the grid cells the 1°x1° bathymetry is outside the POP2 depth domain and no CaCO3 MAR can be calculated? If yes, can this be improved to show a continuous map?

References:

Gu, S., Liu, Z., Oppo, D.W., Lynch-stieglitz, J., Jahn, A., Zhang, J., Wu, L., 2020. Assessing the potential capability of reconstructing glacial Atlantic water masses and AMOC using multiple proxies in CESM. Earth Planet. Sci. Lett. 541, 116294.

Oppo, D.W., Gebbie, G., Huang, K.F., Curry, W.B., Marchitto, T.M., Pietro, K.R., 2018. Data Constraints on Glacial Atlantic Water Mass Geometry and Properties. Paleoceanogr. Paleoclimatology 33, 1013–1034.

Sarnthein, M., Schneider, B., Grootes, P.M., 2013. Peak glacial 14C ventilation ages suggest major draw-down of carbon into the abyssal ocean. Clim. Past 9, 2595–2614.

Skinner, L., McCave, I.N., Carter, L., Fallon, S., Scrivner, A.E., Primeau, F., 2015. Reduced ventilation and enhanced magnitude of the deep Pacific carbon pool during the last glacial period. Earth Planet. Sci. Lett. 411, 45–52.