the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 08 Dec 2025
Simulating the impact of an AMOC weakening on the Antarctic Ice Sheet using a coupled climate and ice-sheet model
Abstract. Climate model studies show that a shutdown of the Atlantic Meridional Overturning Circulation (AMOC) reduces northward heat transport into the North Atlantic, which causes an accumulation of heat in the sub-tropical Southern Ocean. The Antarctic Ice Sheet meanwhile has been shown to be particularly susceptible to temperature changes in ocean water flowing into the cavities of its grounded ice shelves. How AMOC-induced modulation of inter-hemispheric heat transport could influence the present-day state of the Antarctic Ice Sheet via a southward propagation of warm anomalies is little studied. As both, the AMOC as well as the West Antarctic Ice Sheet, are classified as climate tipping points, which can trigger irreversible changes in the Earth System, it is highly relevant how both systems interact with each other.
In this study, we simulate for the first time a shutdown of the AMOC in a global climate model interactively coupled to an ice-sheet model for Antarctica. In line with previous studies, an AMOC shutdown causes increased sea-surface temperatures in the Southern Hemisphere along with a small shift in the mid-latitude westerlies. However, Southern Ocean subsurface temperatures, which drive basal melt in Antarctica, do not change in most regions along the Antarctic margin for the first eight centuries post AMOC shutdown. Therefore, we do not find a change in the total Antarctic Ice volume in this time span. At later times, this is followed by a shift towards stronger Ross Sea convection, causing negative subsurface temperature anomalies of −1.4 °C on average. This cooling decreases basal melt in Antarctica, however increased calving balances the ice mass change. Even though our coupling approach strongly simplifies eddy mass and heat fluxes in the Southern Ocean, and does not resolve flows within ice-shelf cavities, our approach is an important first step to systematically investigate Earth-system stability in coupled climate–ice-sheet models.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(5960 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(5960 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-5128', Anonymous Referee #1, 05 Jan 2026
-
AC1: 'Reply on RC1', Anna Höse, 16 Jan 2026
We thank the reviewer for the very constructive feedback and the overall positive evaluation of our manuscript. Here we provide a general response to the raised points of concern, before providing a revised manuscript after receiving the second review.
Concerning the major comment:
We agree with the reviewer that the artificial freshwater forcing is indeed an important caveat of our study. Furthermore, we acknowledge the reviewer’s perspective that the implications of the experiment setup should be more present throughout the manuscript.
In a revised version, we will therefore
- rephrase the last paragraph of the abstract to include the caveat of freshwater hosing
- rewrite the discussion to put more weight on the implications of the artificial freshwater input into the North Atlantic
Concerning the impact of sea level rise on the Antarctic Ice Sheet, we want to clarify that the experiments were conducted without using the sea level forcing option in the coupling code. The ice sheet model thereby does not see any of the sea level changes in the ocean. Due to the general setup of our experiment, this choice is reasonable, as we do not aim to investigate the impact of sea level rise on Antarctic Ice Sheet dynamics, but are rather interested in changes induced by temperature and salinity forcing only.
Concerning the comment about the 0.1 Sv experiment:
Overall, we see a similar response in Antarctic basal melting when comparing the 0.1 Sv and 0.3 Sv experiments. The salinity decline in basal melt input (PICO) is at a slower rate in the 0.1 Sv case. As the influence of this salinity change on the basal melt response is weak in both the 0.1 Sv and the 0.3 Sv experiment, we conclude that the basal melting response is primarily temperature driven.
Concerning the other specific comments:
We thank the reviewer for reading our manuscript carefully and for pointing out several ambiguities and mistakes in wording. We will correct these and provide a point-to-point response to all raised comments when providing a revised version of the manuscript.
Citation: https://doi.org/10.5194/egusphere-2025-5128-AC1 - AC2: 'Point-by-point response to RC1', Anna Höse, 17 Apr 2026
-
AC1: 'Reply on RC1', Anna Höse, 16 Jan 2026
-
RC2: 'Comment on egusphere-2025-5128', Anonymous Referee #2, 02 Feb 2026
This study analyzes the response of the Antarctic ice sheet to an AMOC collapse using the commonly used “water hosing” approach. To the authors’ (and my) knowledge, this is the first hosing experiment with a coupled Antarctic ice sheet, even though the coupling is for now only via oceanic feedbacks. Despite this caveat, this is certainly a step forward in understanding AMOC-AIS interactions. The analysis is careful and detailed, and the manuscript is logically organized and well-written. Therefore, I recommend publication subject to moderate revisions.
Major comments
- Absence of atmospheric forcing anomalies for the AIS. This seems like a significant limitation and should therefore be acknowledged more explicitly already in the Methods (around L145). It is good that this aspect is discussed in some detail in L428 and following, but the discussion could be improved. For example, it would be very useful to add a supplementary figure that shows precipitation changes (currently “not shown” in L434). Then, could you indicate how including these temperature and precipitation changes might qualitatively influence the ice sheet mass changes (negatively or positively?). Could they even explain the (as you say, somewhat unexpected) absence of a significant AIS mass change?
- Deep convection and AABW formation. A maybe even more important limitation that likely influences some of the results is the typical but unrealistic AABW formation via open-ocean deep convection instead of on the shelves. One possibility to add better context would be to analyze salinity and density changes on the shelves (compare e.g. Li et al. 2023), but – as you write – this might be difficult to disentangle from the effects of the open-ocean convection. While it is adequately addressed in the Discussion, this shortcoming should be emphasized more clearly in other parts of the manuscript, especially in the model description and in the conclusions, e.g. around L457.
Please also consider mentioning the caveats related to points (1) and (2) in the abstract; they might be more important than the currently mentioned caveats (parametrizing eddies and resolving ice-shelf cavity flows [although I am not an expert on the latter]) - Model evaluation. Since this is, from my understanding, a new model configuration, it would be good to add a short section (maybe in the Appendix) for model evaluation, especially in terms of key Southern Ocean characteristics such as T and S profiles in the relevant ice shelf forcing regions, sea ice extent and AABW strength. The initial AMOC strength, which appears a bit high, should also be compared to observations.
- No salinity compensation. According to L181, the experiment follows the protocol of NAHosMIP (Jackson et al. 2023), but this is not the case for salinity compensation. Jackson et al. (2023) use a volume compensation term which compensates for the hosing throughout the global ocean, whereas no compensation is applied here. This should be mentioned clearly (and maybe justified, if there was a specific reason to not apply any compensation). In addition, it should be mentioned more explicitly that, when applying a large hosing term without compensation, the ocean continuously (and artificially) freshens.
This makes me wonder if any of the results presented in the manuscript are actually a direct effect of this global freshening. For example, the freshening in Fig. A13b looks suspiciously linear. It might be worth performing a back-of-the-envelope calculation how much the global mean salinity changes due to the hosing and then compare salinity changes against this number. - Significance testing. Currently, no significance testing is shown in the figures, but this would be important especially for the first period (years 100-200). Which of these features are significant signals and which ones could as well have arisen due to (low-frequency) internal variability (when compared to the control simulation)?
Specific comments
L16: “first step”: I think that some work in this direction has been done for Greenland, so maybe “an important step” is enough (or add “both hemispheres” somewhere in this sentence).
L19-36 presents several contentious issues (has the AMOC already declined during the historical period, how important is GrIs meltwater, what is the risk of 21st-century AMOC tipping) almost as facts, but there is often “another side” of the debate, which should also be referenced.
L38 could also reference the review by Lynch-Stieglitz (2017)
L51: “open-ocean convection is hardly observed” should be referenced
L67: Is Berdahl et al. 2024 based on proxy or model data?
L159: “pre-industrial atmospheric conditions” – as a caveat that arises from this, it could be mentioned somewhere that impacts of AMOC weakening have been shown to be quite strongly state-dependent (e.g. Bellomo & Mehling 2024).
L166-180: I see that the spin-up is already documented in Kreuzer et al. (2025) – this could be pointed out more clearly. Could you comment here on how equilibrated the model is at the end of the spin-up? (e.g. TOA imbalance, global ocean heat content trend or similar)
L222: What is the role of salinity for basal melting? For temperature it is intuitive, but less so for salinity. It could be mentioned here (and maybe in other places, see above) that at least some of the salinity decrease appears to come from the (non-compensated) freshwater input.
Fig. 2b: Worth repeating the depth range over which the temperature was averaged in the caption
L239: A reference could be added for the South Atlantic salinification, e.g. Zhu & Liu (2020)
L260: Is this change in convective activity in the Weddell Sea forced or linked to internal variability? Maybe to a mode similar to the well-documented centennial variability in the GFDL model (e.g. Zhang et al. 2019)?
L264 and following suggests that atmospheric temperatures influence SSTs. So, is there an atmospheric mechanism for the surface air temperature cooling? More generally, could you comment in the Discussion about the role/importance of atmospheric mechanisms?
L271, L278, L288, L305 etc.: You often refer to properties of upwelled water. Could this be more clearly demonstrated by showing Atlantic Ocean/SO sector cross-sections (latitude vs. depth) with isopycnals overlaid?
Section 3.3: Could you remind the reader at the beginning of this section why this period was selected for analysis?
L369: Stocker et al. (2007) compared surface compensation vs. no compensation, but here the relevant difference would be volume compensation (as in the NAHosMIP protocol) vs. no compensation. This should be mentioned somewhere in this paragraph as it likely makes quite a difference.
L373: For context, it could be added here how much SLE the Greenland ice sheet holds.
L391: The recent study by Aguiar et al. (2025), who argued for a role of vertical resolution, could be mentioned here
L404-420: The relevance of this paragraph for the conclusions is not so clear, especially since there are few citations and it is concluded that you “do not expect high impacts due to this limitation”. I think that such minor limitations can be covered in the methods section and that this paragraph could be removed to make the discussion a bit less extensive.
L432: “most changes in Antarctica are due to interactions at the ice-ocean interface” could use a reference
L451: It could be mentioned that there are now some models that do couple the AIS in their future projections, e.g. UKESM (Smith et al. 2021).
Technical corrections
L335 and elsewhere: Wunderling 2023 should be 2024
L339: warms temperatures -> increases temperatures
L367: Jackson & Wood 2018 should be Jackson et al. 2023?
L376: strictly idealized -> highly idealized
L431: no comma
Aguiar, W., et al. Antarctic Dense Water Formation Sensitivity to Ocean Surface Cell Thickness. Journal of Advances in Modeling Earth Systems, 17, e2024MS004913 (2025). https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2024MS004913
Bellomo, K. & Mehling, O. Impacts and State-Dependence of AMOC Weakening in a Warming Climate. Geophysical Research Letters 51, e2023GL107624 (2024). https://onlinelibrary.wiley.com/doi/abs/10.1029/2023GL107624
Li, Q., et al. Abyssal ocean overturning slowdown and warming driven by Antarctic meltwater. Nature 615, 841–847 (2023). https://www.nature.com/articles/s41586-023-05762-w
Lynch-Stieglitz, J. The Atlantic Meridional Overturning Circulation and Abrupt Climate Change. Annual Review of Marine Science 9, 83–104 (2017). https://www.annualreviews.org/doi/10.1146/annurev-marine-010816-060415
Smith, R. S., et al. Coupling the U.K. Earth System Model to Dynamic Models of the Greenland and Antarctic Ice Sheets. Journal of Advances in Modeling Earth Systems, 13, e2021MS002520 (2021). https://doi.org/10.1029/2021MS002520
Zhang, L., et al. Natural variability of Southern Ocean convection as a driver of observed climate trends. Nature Climate Change, 9, 59–65 (2019). https://www.nature.com/articles/s41558-018-0350-3
Zhu, C. & Liu, Z. Weakening Atlantic overturning circulation causes South Atlantic salinity pile-up. Nature Climate Change, 10, 998–1003 (2020). https://doi.org/10.1038/s41558-020-0897-7
Citation: https://doi.org/10.5194/egusphere-2025-5128-RC2 - AC3: 'Reply on RC2', Anna Höse, 17 Apr 2026
Peer review completion
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-5128', Anonymous Referee #1, 05 Jan 2026
-
AC1: 'Reply on RC1', Anna Höse, 16 Jan 2026
We thank the reviewer for the very constructive feedback and the overall positive evaluation of our manuscript. Here we provide a general response to the raised points of concern, before providing a revised manuscript after receiving the second review.
Concerning the major comment:
We agree with the reviewer that the artificial freshwater forcing is indeed an important caveat of our study. Furthermore, we acknowledge the reviewer’s perspective that the implications of the experiment setup should be more present throughout the manuscript.
In a revised version, we will therefore
- rephrase the last paragraph of the abstract to include the caveat of freshwater hosing
- rewrite the discussion to put more weight on the implications of the artificial freshwater input into the North Atlantic
Concerning the impact of sea level rise on the Antarctic Ice Sheet, we want to clarify that the experiments were conducted without using the sea level forcing option in the coupling code. The ice sheet model thereby does not see any of the sea level changes in the ocean. Due to the general setup of our experiment, this choice is reasonable, as we do not aim to investigate the impact of sea level rise on Antarctic Ice Sheet dynamics, but are rather interested in changes induced by temperature and salinity forcing only.
Concerning the comment about the 0.1 Sv experiment:
Overall, we see a similar response in Antarctic basal melting when comparing the 0.1 Sv and 0.3 Sv experiments. The salinity decline in basal melt input (PICO) is at a slower rate in the 0.1 Sv case. As the influence of this salinity change on the basal melt response is weak in both the 0.1 Sv and the 0.3 Sv experiment, we conclude that the basal melting response is primarily temperature driven.
Concerning the other specific comments:
We thank the reviewer for reading our manuscript carefully and for pointing out several ambiguities and mistakes in wording. We will correct these and provide a point-to-point response to all raised comments when providing a revised version of the manuscript.
Citation: https://doi.org/10.5194/egusphere-2025-5128-AC1 - AC2: 'Point-by-point response to RC1', Anna Höse, 17 Apr 2026
-
AC1: 'Reply on RC1', Anna Höse, 16 Jan 2026
-
RC2: 'Comment on egusphere-2025-5128', Anonymous Referee #2, 02 Feb 2026
This study analyzes the response of the Antarctic ice sheet to an AMOC collapse using the commonly used “water hosing” approach. To the authors’ (and my) knowledge, this is the first hosing experiment with a coupled Antarctic ice sheet, even though the coupling is for now only via oceanic feedbacks. Despite this caveat, this is certainly a step forward in understanding AMOC-AIS interactions. The analysis is careful and detailed, and the manuscript is logically organized and well-written. Therefore, I recommend publication subject to moderate revisions.
Major comments
- Absence of atmospheric forcing anomalies for the AIS. This seems like a significant limitation and should therefore be acknowledged more explicitly already in the Methods (around L145). It is good that this aspect is discussed in some detail in L428 and following, but the discussion could be improved. For example, it would be very useful to add a supplementary figure that shows precipitation changes (currently “not shown” in L434). Then, could you indicate how including these temperature and precipitation changes might qualitatively influence the ice sheet mass changes (negatively or positively?). Could they even explain the (as you say, somewhat unexpected) absence of a significant AIS mass change?
- Deep convection and AABW formation. A maybe even more important limitation that likely influences some of the results is the typical but unrealistic AABW formation via open-ocean deep convection instead of on the shelves. One possibility to add better context would be to analyze salinity and density changes on the shelves (compare e.g. Li et al. 2023), but – as you write – this might be difficult to disentangle from the effects of the open-ocean convection. While it is adequately addressed in the Discussion, this shortcoming should be emphasized more clearly in other parts of the manuscript, especially in the model description and in the conclusions, e.g. around L457.
Please also consider mentioning the caveats related to points (1) and (2) in the abstract; they might be more important than the currently mentioned caveats (parametrizing eddies and resolving ice-shelf cavity flows [although I am not an expert on the latter]) - Model evaluation. Since this is, from my understanding, a new model configuration, it would be good to add a short section (maybe in the Appendix) for model evaluation, especially in terms of key Southern Ocean characteristics such as T and S profiles in the relevant ice shelf forcing regions, sea ice extent and AABW strength. The initial AMOC strength, which appears a bit high, should also be compared to observations.
- No salinity compensation. According to L181, the experiment follows the protocol of NAHosMIP (Jackson et al. 2023), but this is not the case for salinity compensation. Jackson et al. (2023) use a volume compensation term which compensates for the hosing throughout the global ocean, whereas no compensation is applied here. This should be mentioned clearly (and maybe justified, if there was a specific reason to not apply any compensation). In addition, it should be mentioned more explicitly that, when applying a large hosing term without compensation, the ocean continuously (and artificially) freshens.
This makes me wonder if any of the results presented in the manuscript are actually a direct effect of this global freshening. For example, the freshening in Fig. A13b looks suspiciously linear. It might be worth performing a back-of-the-envelope calculation how much the global mean salinity changes due to the hosing and then compare salinity changes against this number. - Significance testing. Currently, no significance testing is shown in the figures, but this would be important especially for the first period (years 100-200). Which of these features are significant signals and which ones could as well have arisen due to (low-frequency) internal variability (when compared to the control simulation)?
Specific comments
L16: “first step”: I think that some work in this direction has been done for Greenland, so maybe “an important step” is enough (or add “both hemispheres” somewhere in this sentence).
L19-36 presents several contentious issues (has the AMOC already declined during the historical period, how important is GrIs meltwater, what is the risk of 21st-century AMOC tipping) almost as facts, but there is often “another side” of the debate, which should also be referenced.
L38 could also reference the review by Lynch-Stieglitz (2017)
L51: “open-ocean convection is hardly observed” should be referenced
L67: Is Berdahl et al. 2024 based on proxy or model data?
L159: “pre-industrial atmospheric conditions” – as a caveat that arises from this, it could be mentioned somewhere that impacts of AMOC weakening have been shown to be quite strongly state-dependent (e.g. Bellomo & Mehling 2024).
L166-180: I see that the spin-up is already documented in Kreuzer et al. (2025) – this could be pointed out more clearly. Could you comment here on how equilibrated the model is at the end of the spin-up? (e.g. TOA imbalance, global ocean heat content trend or similar)
L222: What is the role of salinity for basal melting? For temperature it is intuitive, but less so for salinity. It could be mentioned here (and maybe in other places, see above) that at least some of the salinity decrease appears to come from the (non-compensated) freshwater input.
Fig. 2b: Worth repeating the depth range over which the temperature was averaged in the caption
L239: A reference could be added for the South Atlantic salinification, e.g. Zhu & Liu (2020)
L260: Is this change in convective activity in the Weddell Sea forced or linked to internal variability? Maybe to a mode similar to the well-documented centennial variability in the GFDL model (e.g. Zhang et al. 2019)?
L264 and following suggests that atmospheric temperatures influence SSTs. So, is there an atmospheric mechanism for the surface air temperature cooling? More generally, could you comment in the Discussion about the role/importance of atmospheric mechanisms?
L271, L278, L288, L305 etc.: You often refer to properties of upwelled water. Could this be more clearly demonstrated by showing Atlantic Ocean/SO sector cross-sections (latitude vs. depth) with isopycnals overlaid?
Section 3.3: Could you remind the reader at the beginning of this section why this period was selected for analysis?
L369: Stocker et al. (2007) compared surface compensation vs. no compensation, but here the relevant difference would be volume compensation (as in the NAHosMIP protocol) vs. no compensation. This should be mentioned somewhere in this paragraph as it likely makes quite a difference.
L373: For context, it could be added here how much SLE the Greenland ice sheet holds.
L391: The recent study by Aguiar et al. (2025), who argued for a role of vertical resolution, could be mentioned here
L404-420: The relevance of this paragraph for the conclusions is not so clear, especially since there are few citations and it is concluded that you “do not expect high impacts due to this limitation”. I think that such minor limitations can be covered in the methods section and that this paragraph could be removed to make the discussion a bit less extensive.
L432: “most changes in Antarctica are due to interactions at the ice-ocean interface” could use a reference
L451: It could be mentioned that there are now some models that do couple the AIS in their future projections, e.g. UKESM (Smith et al. 2021).
Technical corrections
L335 and elsewhere: Wunderling 2023 should be 2024
L339: warms temperatures -> increases temperatures
L367: Jackson & Wood 2018 should be Jackson et al. 2023?
L376: strictly idealized -> highly idealized
L431: no comma
Aguiar, W., et al. Antarctic Dense Water Formation Sensitivity to Ocean Surface Cell Thickness. Journal of Advances in Modeling Earth Systems, 17, e2024MS004913 (2025). https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2024MS004913
Bellomo, K. & Mehling, O. Impacts and State-Dependence of AMOC Weakening in a Warming Climate. Geophysical Research Letters 51, e2023GL107624 (2024). https://onlinelibrary.wiley.com/doi/abs/10.1029/2023GL107624
Li, Q., et al. Abyssal ocean overturning slowdown and warming driven by Antarctic meltwater. Nature 615, 841–847 (2023). https://www.nature.com/articles/s41586-023-05762-w
Lynch-Stieglitz, J. The Atlantic Meridional Overturning Circulation and Abrupt Climate Change. Annual Review of Marine Science 9, 83–104 (2017). https://www.annualreviews.org/doi/10.1146/annurev-marine-010816-060415
Smith, R. S., et al. Coupling the U.K. Earth System Model to Dynamic Models of the Greenland and Antarctic Ice Sheets. Journal of Advances in Modeling Earth Systems, 13, e2021MS002520 (2021). https://doi.org/10.1029/2021MS002520
Zhang, L., et al. Natural variability of Southern Ocean convection as a driver of observed climate trends. Nature Climate Change, 9, 59–65 (2019). https://www.nature.com/articles/s41558-018-0350-3
Zhu, C. & Liu, Z. Weakening Atlantic overturning circulation causes South Atlantic salinity pile-up. Nature Climate Change, 10, 998–1003 (2020). https://doi.org/10.1038/s41558-020-0897-7
Citation: https://doi.org/10.5194/egusphere-2025-5128-RC2 - AC3: 'Reply on RC2', Anna Höse, 17 Apr 2026
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 2,544 | 2,471 | 149 | 5,164 | 97 | 138 |
- HTML: 2,544
- PDF: 2,471
- XML: 149
- Total: 5,164
- BibTeX: 97
- EndNote: 138
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Moritz Kreuzer
Willem Huiskamp
Stefan Petri
Georg Feulner
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(5960 KB) - Metadata XML
Please see the attached PDF