Extier et al. study the transfer of organic terrestrial carbon to the ocean due to sea-level rise during the last deglaciation. These results are from a transient simulation of the last deglaciation performed with the MPI ESM, which includes an interactive adjustment of the land-sea interface as well as dynamic vegetation. This is a very novel and interesting study, allowing the quantification of the carbon cycle response to changes in deglacial sea-level from land flooding. It is however unclear how the simulated changes in sea-level (derived from the simulated meltwater input into the ocean) compare to paleo-estimates. A discussion on the simulated sea-level and its implication on the results is needed. In addition, the introduction needs to be revised.

1) Title: I suggest “Local oceanic CO2 outgassing….”

From the title one can think that the terrestrial organic carbon input to the ocean during the deglaciation leads to significant CO2 outgassing, whereas that’s not the case. “Local outgassing” seems to better reflect the conclusions.

2) Introduction:

The structure of the introduction needs to be improved and some aspects that are the focus of the manuscript are currently missing.

The introduction does not properly mention the current hypotheses to explain glacial-interglacial changes in pCO2. L52-53 is confusing, as there has been a lot of work done in understanding glacial-interglacial changes in atmospheric CO2, including some transient simulations. On the other hand, the introduction includes one paragraph on the impact of Heinrich events on the carbon cycle from a modelling perspective, but links with the current studies are not made.

The introduction should also include a paragraph on estimates of glacial-interglacial changes in terrestrial carbon to provide a perspective on the modelling outputs.

Finally, since MWP1A is mentioned throughout the manuscript, a brief paragraph on MWP1a should be added. This paragraph could describe estimates of the timing of MWP1A, its magnitude and the potential origin of this meltwater pulse (NH vs SH).

3) Deglacial sea-level rise

Since the results of the present study are dependent on the sea-level rise, a timeseries of the simulated deglacial sea-level rise along with paleo-estimates should be included in Figure 2. The implications of potential differences between simulated and estimated deglacial sea-level rise should be discussed.

4) Line by line comments

P1, L. 1: The first sentence of the abstract is odd. It needs to be rephrased, and most likely split in two sentences.

P1, L.2: I suggest “induces a sea-level rise”

P1, L. 10: I suggest “leads to 21.2 GtC transfer of terrestrial organic carbon to the ocean”

P1, L. 20: “including” instead of “triggered”

P2, L. 28: Consider adding references to Lambeck et al. 2014.

P7, L. 190: Weren’t the nutrient concentrations adjusted for the lower sea-level at the LGM?

P8, L. 218-220: I find this sentence confusing. What do you suggest the relationship between oceanic circulation and CO2 uptake is? A well ventilated Southern Ocean is usually associated with CO2 outgassing and not uptake.

In addition, the ocean to atm. CO2 flux shown in Fig. 2c is not really explained, and barely mentioned in the text. However, given the experimental setup, it might be simply responding to the forced changes in atmospheric CO2, and to changes in surface solubility.

A timeseries of atmospheric CO2 should be included in figure 2 to better understand the ocean-atm CO2 flux (2c), and maybe a timeseries of globally averaged SST.

P9, L. 230: The Pa/Th record from the Bermuda rise suggest an AMOC weakening during HS1, but not necessarily during MWP1a. I think that the most recent chronology suggest the end of HS1 and thus beginning of Bolling Allerod at 14.6 ka, contemporary with the beginning of MWP1a.

P12, L. 283 and throughout: I understand why you are referring to “terrestrial organic carbon input to the atmosphere”, however this is not correct and could be confusing. It might be better to simply refer to “terrestrial carbon input to the atmosphere”

P19, L. 369 and throughout manuscript: I am not sure that the use of North/central and south Indonesia is correct. Maybe it is more appropriate to refer to the Sunda and Sahul shelves.

Figures:

Figure 4: Comparing 15ka vegetation with reconstructions from 21 ka does not seem appropriate. Please show the JSBACH field at 21 ka compared to LGM proxies. I however understand that given that the type of vegetation at the area of flooding will impact the terrestrial organic transfer, it might also be necessary to show JSBACH at 15ka.

Figure 8: Add AMOC and/or meltwater timeseries?

Figure 11: Add a sentence in the caption stating that the timeseries start when the flooding event at that location occurs (if that’s the case).

Review of "Oceanic CO₂ outgassing triggered by terrestrial organic carbon fluxes during deglacial flooding" by Thomas Extier et al.

Thomas Extier and colleagues present a new implementation of terrestrial organic carbon fluxes between the land and ocean as land is flooded and becomes ocean or vice versa in the MPI-ESM as well as results from a deglacial simulation.

I was very pleased to see this process being implemented in an Earth system model (ESM) and support the publication of the well written article with a clear documentation of the implementation and am looking forward to see further applications.

Nevertheless, I would like to add a few comments.

1) Introduction:

Maybe expand the section on glacial-interglacial CO₂ variations (p.2, ll.39-49). In the context of the current study, recent estimates of total changes in land carbon storage between the last glacial maximum (LGM) and preindustrial (PI) might be of interest (e.g. Müller and Joos, 2020 BG; Jeltsch-Thömmes et al., 2019 CP).

Further, many studies have invoked processes other than physical changes in the ocean (see e.g. Menviel et al., 2012 QSR or Sigman et al., 2010 Nature for a review, and many others) to explain glacial-interglacial CO₂ variations.

2) Prescribed atmospheric CO₂ concentration:

At the end of section 2.1 the authors note that all atmospheric concentrations are prescribed in the simulations. 

One of the goals of glacial-interglacial simulations with ESMs is to simulate the change in atmospheric CO₂ concentration, i.e. the ~90 ppm increase since the LGM. While making sure to have the correct atmospheric inventory, prescribing atm. CO₂ comes with drawbacks. For example, without other changes this would lead to a smaller LGM DIC inventory as the atmosphere would act as a sink until equilibrium is reached with the ocean. The authors circumvent this by initializing the spinup simulations with higher alkalinity concentrations. Is there a specific reason for not letting atm. CO₂ evolve freely over the course of the simulation? I don't think, though, this would change the findings of the study, as the effect of terrestrial organic carbon fluxes is diagnosed from the difference of two runs, but would like to see at least a short discussion of this choice.

Are changes in tracer concentrations as a result of lower sea-level considered here as well?

3) Simulated terrestrial carbon inventory

In general, I was a bit surprised to read that the effect of terrestrial organic carbon fluxes as a result of flooding are rather small and am wondering whether this might link to the size of the simulated terrestrial carbon inventory and thus the amount of carbon available in flooded gridcells.

On page 10, l.232-233 the authors state that the terrestrial carbon inventory increased from 922.9 GtC to 1302.7 GtC between 21-15 kaBP and amounts to 1563.6 GtC in 12 kaBP. I am no expert on land modeling, but in a recent paper Müller and Joos (2020, BG) simulate total terrestrial carbon at the LGM at about 2000 GtC, which increases to about 2500 GtC in 12 kaBP. This is almost twice the amount shown in this study. Also, Ganopolski and Brovkin (2017, CP) simulate a larger terrestrial carbon inventory. Is the assumption correct that a higher terrestrial carbon pool would also increase the terrestrial organic carbon flux during flooding? If yes, this might be a point to be included in the discussion of uncertainties of the findings.

In the same paragraph (p.10, ll.235-237) include Müller and Joos, 2020 BG into the estimates of terrestrial carbon evolution.

Are peatlands included in the land component of the model?

Are there other uncertainties that would be good to be discussed (other than C:N:P ratios)?

Minor comments:

- make sure text in figures is readable (size), for example legends in Fig. 11 are very small

- p.5, ll.129-130: either 'presented a new development' or 'presented new developments'

- p.11, Fig. 4: why not compare 21 ka model with 21 ka reconstruction?

References:

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, Clim. Past, 13, 1695–1716, https://doi.org/10.5194/cp-13-1695-2017, 2017.

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

Menviel, L., Joos, F., and Ritz, S. P.: Simulating atmospheric CO₂ , δ¹³C and the marine carbon cycle during the Last Glacial-Interglacial cycle: Possible role for a deepening of the mean remineralization depth and an increase in the oceanic nutrient inventory, Quaternary Sci. Rev., 56, 46–68, https://doi.org/10.1016/j.quascirev.2012.09.012, 2012.

Müller,  J.  and  Joos,  F.:  Global  peatland  area  and  carbon  dynamics  from  the  Last  Glacial  Maximum  to  the  present - aprocess-based  model  investigation,  Biogeosciences,  17,  5285–5308, https://doi.org/10.5194/bg-17-5285-2020, 2020.

Sigman, D. M., Hain, M. P., and Haug, G. H.: The polar ocean and glacial cycles in atmospheric CO₂ concentration, Nature, 466, 47–55, https://doi.org/10.1038/nature09149, 2010.