Quantifying the feedback between Antarctic meltwater release and subsurface Southern Ocean warming

Lambert, Erwin; Le Bars, Dewi; van der Linden, Eveline; Jüling, André; Drijfhout, Sybren

doi:https://doi.org/10.5194/esd-16-1303-2025

Articles | Volume 16, issue 4

https://doi.org/10.5194/esd-16-1303-2025

© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/esd-16-1303-2025

© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 16, issue 4

Research article

|

11 Aug 2025

Research article |

| 11 Aug 2025

Quantifying the feedback between Antarctic meltwater release and subsurface Southern Ocean warming

Erwin Lambert, Dewi Le Bars, Eveline van der Linden, André Jüling, and Sybren Drijfhout

Download

Final revised paper (published on 11 Aug 2025)
Preprint (discussion started on 08 Oct 2024)

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2024-2257', Anonymous Referee #1, 12 Nov 2024

Please find attached a pdf with my review of “Quantifying the feedback between Antarctic meltwater release and subsurface Southern Ocean warming” by Lambert et al.

Citation: https://doi.org/10.5194/egusphere-2024-2257-RC1
- AC1: 'Reply on RC1', Erwin Lambert, 16 Jan 2025
  
  Dear reviewer, please find our reply in the attachment.
  
  Citation: https://doi.org/10.5194/egusphere-2024-2257-AC1
RC2:
'Comment on egusphere-2024-2257', Anonymous Referee #2, 15 Dec 2024

Review of “Quantifying the feedback between Antarctic meltwater release and subsurface Southern Ocean warming” by Lambert et al.

The manuscript by Lambert and colleagues uses output from Earth system models, ice sheet models, and observations to estimate the impact of an ocean-ice feedback on sea level projections. The majority of Earth system models do not yet include coupled ice sheets and ice shelves, and thus ice-ocean feedbacks are not currently included in sea level projections. Additional meltwater from Antarctica increases stratification, causing surface cooling and subsurface warming at the depth of ice shelf basal melting, which has the potential to accelerate melting. Making use of existing model output, the authors calculate linear response functions to estimate the impact of this feedback on sea level projections. They find at first pass, that this feedback could increase sea level projections for 2100 by 80%, which is a substantial effect. However, they also note that the basal melt parameter for this estimation has not been calibrated on observed historical melt rates. When accounting for calibration, they revise their estimate to suggest that the feedback could increase sea level projections for 2100 by 5%.

The findings of this manuscript will be useful to the scientific community working towards coupling dynamic ice sheets and shelves into Earth system models. The manuscript is well written, and the methodology is clear described. I have several minor comments specified below, which I expect will be straightforward to address. I also have one major comment, regarding bias / potential bias in the Earth system model ocean component that will influence the ocean response function and results – sensitivity to this should be investigated and presented – currently, it is not clear how dependent the results are too possible ocean-model bias in both the mean state and trends. I thus recommend major revisions for this manuscript.

Specific comments

Abstract line 10-14: The three sentences with reported percentages read a bit as unconnected listed items, and this has the effect of making the percentages seem unrelated. Rewording the calibration methodology and revised sea level enhanced projection sentences would help here.

Line 40: “SOFIAMIP project” would be better described as “Southern Ocean Freshwater Input from Antarctica (SOFIA) Imitative”.

Line 40-42: Can you add citations for these modelling studies in the last sentence – as it is currently written, this seems like it is referring to the SOFIA model output from Swart et al. (2023), which is not the case.

Line 50-51: Do any of models or the multimodel mean agree with the observations?

Line 89: Suggest “redistributed” here.

Line 99-102: It is implicit here that the Antarctic runoff generated in the CMIP6 version of the model is no longer generated in this version, but it would be helpful to describe this explicitly, so that it is clear you aren’t double counting this runoff plus your more realistic representation of meltwater release.

Line 142: Does “bias correct” mean bias correct the mean state, or to bias correct the trends, i.e. de-drift? I think it means the latter. How is this bias correction / de-drifting done? E.g. have you removed linear grid point trends from the piControl in the corresponding historical and SSP scenarios?

Further, the historical CMIP6 ocean trends around the Antarctic margins themselves are biased (Purich and England 2021) – have you corrected for this? For example, multimodel mean historical trends around the margins show more warming than observed in most regions, except the Amundsen Sea, over 1975-2012. The historical ocean trends will presumably exert a strong influence on your projections and gamma calibration. Sensitivity to model bias in both the mean-state and trends should be investigated and reported in the manuscript.

Purich, A., & England, M. H. (2021). Historical and future projected warming of Antarctic Shelf Bottom Water in CMIP6 models. Geophysical Research Letters, 48, e2021GL092752. https://doi.org/10.1029/2021GL092752

Line 169: Computation.

Line 204-207: It would be useful if you list the ice sheet models and ESMs used in this study in a table.

Line 220-222: It would be useful here to specify that you are using the observed historical ice mass loss, and the modelled ocean warming. What is the uncertainty in the historical ice mass loss? What is the uncertainty in the ocean warming?

Line 232-235: What does it mean that some model pairs require a negative gamma to reproduce the positive sea level contribution? Is the model representation of ocean temperature around the Antarctic margins suitable for this use?

Fig. 4 caption: Capitalisation should be “western Peninsula” and “eastern Peninsula”

Line 275: In this section it would be useful here to comment on this subsurface warming relative to other studies (particular in the Amundsen Sea), and comment on the resolution of this model / limitations to representation of high-resolution processes that may affect the response to meltwater.

Line 293-295: This saturation effect has been reported previously, e.g. in Schloesser et al. (2019).

Schloesser, F., T. Friedrich, A. Timmermann, R. M. DeConto, and D. Pollard, 2019: Antarctic iceberg impacts on future Southern Hemisphere climate. Nat. Climate Change, 9, 672– 677, https://doi.org/10.1038/s41558-019-0546-1.

Line 296-298: The upper limit seems reasonable. Did you experiment with other (non-linear) fits?

Line 319: Specify “subsurface” warming.

Line 353-354: Specify “the last decades of our projections”, so it is clear that you are not referring to the last decades of the observations. Also, here and elsewhere, the date that you are presenting projections for (i.e. 2100) should be specified.

Line 363: Specify “surface” cooling.

Line 364-365: Can you elaborate on why it could lead to an overestimation of sea-level projections?

Line 375-377: Have you shown the warming around Antarctica in EC-Earth3, compared to other models? Can you support this statement with references?

Line 380: “Contrast”

Conclusions: Some of the conclusions text seems like it is written in brief draft format. I recommend revising the text.

Citation: https://doi.org/10.5194/egusphere-2024-2257-RC2
- AC2: 'Reply on RC2', Erwin Lambert, 16 Jan 2025
  
  Dear reviewer, please find our reply in the attachment
  
  Citation: https://doi.org/10.5194/egusphere-2024-2257-AC2

Peer review completion

AR: Author's response | RR: Referee report | ED: Editor decision | EF: Editorial file upload

ED: Reconsider after major revisions (22 Jan 2025) by Ira Didenkulova

AR by Erwin Lambert on behalf of the Authors (27 Mar 2025) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (28 Mar 2025) by Ira Didenkulova

RR by Anonymous Referee #1 (03 Apr 2025)

Suggestions for revision or reasons for rejection

Dear Editor and Authors,

Thank you for sending the revised manuscript and the point-by-point responses to my comments. I enjoyed reading the manuscript again, and it is still a very nice paper that the community will value tremendously.

I have reviewed the revisions and am satisfied with the changes made. The authors have addressed all of my concerns appropriately, and I believe the manuscript has improved. I am happy to recommend the manuscript for publication pending a few minor revisions (see below).

Line 437 (in the tracked-changes document): I am happy with the added discussion here. But: “An inter-model comparison within the SOFIA initiative (Swart et al., 2023) showed a qualitative agreement on subsurface warming along the Antarctic coast (Chen et al., 2023).” This is not correct and needs to be changed/removed. Chen et al. show a qualitative agreement on subsurface warming in the zonal mean. This does not mean that the models show a warming along the coast. Current work in prep in the SOFIA community shows that this “zonal mean” warming is dominated by certain regions and that there is cooling along the coast in some regions in some models. This work is, sadly, not yet citable. I would either remove this sentence completely or state that a recent intermodel comparison shows that, in the zonal average, models agree on a subsurface warming but that regional differences are the focus of current investigations, such as those within the SOFIA initiative.

Regarding the constant freezing temperature, -1.7. I understand your wish for simplicity and transparency, and I see that rerunning the experiments with different values would be a lot of work for likely a very small change. The added argumentation in the methods section is an improvement, but I am not yet fully satisfied with it. Firstly, you state that you have chosen the surface freezing temperature, correct, but I think you need to add that this an unrealistic value and that the freezing temperature at depth is closer to –2.0 - -2.3 (for your depth ranges). Secondly, I think you need to be more quantitative and precise in the following statement: “We consider the uncertainty stemming from this idealization to be minimal in comparison to uncertainties from other methodological assumptions.” What is minimal? What difference approximately can we expect if we use -1.7 instead of -2.3? Can you provide an order of magnitude? And can you be more precise on which other assumptions you refer to her?

Gt/yr vs. Sv: I see you’ve noted that 1 Gt/yr = 0.0317 mSv, which is helpful. However, to improve clarity for the reader, I recommend consistently including the equivalent value in Sverdrups whenever you present a value in Gt/yr. This will save readers from having to do the conversion themselves each time. For instance, "an increase in ice mass loss of 400 Gt/yr (≈0.012 Sv)".

Hide

RR by Anonymous Referee #2 (13 May 2025)

Suggestions for revision or reasons for rejection

Second review of “Quantifying the feedback between Antarctic meltwater release and subsurface Southern Ocean warming” by Lambert et al.

In their revision, the authors have suitable addressed my comments from the first review. The revised manuscript makes it clear where results are based on a single coupled model and includes an expanded discussion on biases in EC-Earth. Where practicable for this study, a comparison of results across coupled models are presented. The findings of this manuscript will be useful to the scientific community working towards coupling dynamic ice sheets and shelves into Earth system models, and I consider this manuscript in its revised form suitable for publication in ESD.

I have two very minor comments for consideration:

Line 159-160 (of tracked changes version): I appreciate the clarification regarding dedrifting. It would be useful to briefly include mention of the period over which the piControl linear trend was calculated, e.g.100 years? The full available piControl length?

Line 257-260 (of tracked changes version): The revised text makes your methodology and reasoning clearer, but I wonder, for all regions, is regional net cooling over the historical period actually unrealistic? Over shorter periods (1975-2012) there is cooling in the Weddell and Ross (Schmidtko et al. 2014), which may be due to natural variability, a forced change causing the cooling, or sparse observations not fully sampling the real change. I don’t expect you to change your methodology here, but I think your phrasing of “unrealistic” needs refining, i.e. “which we consider to be unrealistic over a 150-year period with increasing global temperatures” or something like that – then it is explicit why you assume it to be unrealistic.

Hide

ED: Publish subject to minor revisions (review by editor) (13 May 2025) by Ira Didenkulova

AR by Erwin Lambert on behalf of the Authors (14 May 2025) Author's response Author's tracked changes Manuscript

ED: Publish as is (28 May 2025) by Ira Didenkulova

AR by Erwin Lambert on behalf of the Authors (28 May 2025)

Short summary

Ocean warming around Antarctica leads to ice melting and sea-level rise. The meltwater that flows into the surrounding ocean can lead to enhanced warming of the seawater, thereby again increasing melting and sea-level rise. This process, however, is not currently included in climate models. Through a simple mathematical approach, we find that this process can lead to more melting and greater sea-level rise, possibly increasing the Antarctic contribution to 21st century sea-level rise by 80 %.