the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 25 Sep 2025
Response of ice sheets, sea-ice and sea level in climate stabilisation and reversibility simulations using a state-of-the-art Earth System Model
Abstract. We have conducted an ensemble of idealised climate overshoot simulations in a state-of-the-art Earth system model in which global mean temperature is increased at a constant rate to various global warming levels (GWLs) by prescribing constant CO2 emissions, followed by a period of zero CO2 emissions started at different GWLs, and then a period of cooling in which CO2 is removed from the atmosphere. We give an overview of the ice sheet, sea-ice and sea level responses in these simulations highlighting long term responses at different GWLs and discuss the degree to which those responses can be simply reversed as global surface temperature cools. We show a broad divide between the two hemispheres, in which northern hemisphere polar processes have larger direct responses to warming which are more simply reversible than those in the southern hemisphere.
Cessation of CO2 emissions at most GWLs stabilises surface temperatures at high northern latitudes, although a slow warming trend continues at high southern latitudes. Northern hemisphere sea-ice extent and Greenland surface mass balance both stabilise under zero CO2 emissions and appear to return toward preindustrial levels along with surface temperatures when CO2 is sequestered, but southern hemisphere sea-ice continues to decline under zero emissions and does not simply increase again as global temperatures cool. Likewise, the Antarctic circumpolar westerlies exhibit strengthening and shift poleward under global warming, but do not simply return to their preindustrial state when the climate is cooled, with implications for ocean circulation and Antarctic surface mass balance. The thermosteric contributions to global mean sea level from ocean heat uptake and the Greenland ice sheet continue when CO2 emissions are ceased or reversed, at rates which slowly decline on centennial timescales. The net mass balance of Antarctica and its contribution to sea level do not simply scale with global temperature; they result from a complex interaction between the basal and surface mass balances, the ice-dynamic response to those forcings and significant trends inherent to the initialisation of our ice sheet model state.
- Preprint
(1306 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2025-4476', Anonymous Referee #1, 28 Oct 2025
-
RC2: 'Comment on egusphere-2025-4476', Anonymous Referee #2, 09 Dec 2025
Smith and co-authors present the first description of new idealized Earth system model simulations conducted with the UKESM, which is used here in its fully coupled configuration, including dynamic Greenland and Antarctic ice sheets. The simulations encompass the Tier 1 experiments of the Tipping Point Model Intercomparison Project (TIPMIP) but extend substantially beyond them by incorporating additional global warming levels and branched simulation pathways. Running a model of this complexity, particularly with fully dynamic ice sheets, is a significant achievement requiring expertise across multiple disciplines. Nevertheless, the model setup remains somewhat unsatisfactory, as the ice sheets are initialized directly from present-day conditions, resulting in noticeable model drift even in unforced scenarios. The authors acknowledge these limitations and point to ongoing research aimed at addressing them.
The manuscript is well written and provides a concise overview of the simulations, with a particular emphasis on high-latitude processes. However, it remains largely descriptive and does not dive deeply into underlying mechanisms or dynamics. This is also reflected in the figures, which primarily present simple time series and omit visualizations of the more complex dynamics governing these overshoot scenarios. Given this, I wonder whether the manuscript may be more appropriately suited to Earth System Science Data rather than Earth System Dynamics.
Â
Beyond these general comments, I have two major points that I would like the authors to address:
1. Definition of reversibility.
The manuscript refers repeatedly to reversibility and irreversibility of sea ice, ice sheets, and other components, based on the path-(in)dependence of variables during ramp-up versus ramp-down phases. However, these responses are highly transient and differ fundamentally from the traditional concept of reversibility associated with hysteresis, which concerns (quasi) steady-state behavior. I encourage the authors to provide a more thorough definition of reversibility, one that explicitly accounts for the transient nature of the experiments and the characteristic internal timescales of different components (e.g., sea-ice versus ice-sheets).
Â
2. Focus on high latitudes despite ad-hoc ice-sheet initialization.
The strong emphasis on high-latitude and ice-sheet responses is difficult to reconcile with the highly idealized ice-sheet initialization, which the authors themselves acknowledge as a major limitation. This is not a criticism of their general approach; coupling dynamic ice-sheets within fully complex ESMs remains extremely challenging due to their large inertia, long equilibration timescales, and the high computational cost of such simulations. However, given the substantial biases introduced, particularly for the Antarctic ice sheet, it is unclear whether the resulting ice-sheet behavior can be meaningfully interpreted. As there is no straightforward solution to the initialization problem, I recommend reducing the emphasis on ice-sheet responses and instead providing a more detailed analysis of other Earth system components, such as the AMOC, ocean circulation in general, or terrestrial changes.
Â
Â
Minor comments:
L2: Please mention the range of GWLs here.
L11: Mention after how many years of zero CO2 emissions, CO2 was removed again from the atmosphere.
L67: branched off instead of spawned?
L73: I would prefer to have Appendix A moved to the main text, to further improve readability and look up of simulation names.
L96: Can you briefly explain why the negative emission rate was set to half the positive one and not the same?
L115: Is JULES also running on a 1° grid?
L120: Please briefly mention the solver for the ice-sheet dynamics (SIA, SSA, hybrid …)
L142: Are the ice-sheets in thermal equilibrium at time of branching?
L295: Explicitly mention again that this refers to the first 550 yr of the simulation.
L339: Typo: THe PI
Â
Fig. 1: Can you add a panel (or a separate figure) depicting the AMOC evolution for these different model runs? In panel c, why is there an offset at the beginning of the simulations of PI and Up8?
Fig. 7: I find the term Mass Change slightly confusing here. Initially, I thought this refers to the net mass change, instead of the loss term (correct?).
Citation: https://doi.org/10.5194/egusphere-2025-4476-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 514 | 160 | 20 | 694 | 23 | 16 |
- HTML: 514
- PDF: 160
- XML: 20
- Total: 694
- BibTeX: 23
- EndNote: 16
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Review of ‘Response of ice-sheets, sea-ice and sea level in climate stabilisation and reversibility simulations using a state-of-the-art Earth System model’ by R.S. Smith et al., submitted to EarthSystemDynamics.
The paper describes an ensemble of simulations investigating the effect of stabilisation of CO2 levels at different global temperature anomalies using a coupled ESM/ice sheet model for Greenland and Antarctica including a scenario mimicking the effect of removal of the entire anthropogenic CO2. The experimental design follows the TIPMIP protocol and fills in additional levels at which stabilization takes place. The topic is highly actual and fits perfectly into ESD. In general, the paper is well written. The analysis could go more into depth, but the authors see this as a first paper, introducing the set of experiments and leaving the detailed analysis for later paper.
Â
It is sufficient material to be published, but in some places not enough analysis to make me really happy. One example: Following your argument in section 3.6, the Antarctic SMB closely follows the GSAT (306/7) and stabilizes as soon as emission stop (307). This is obviously differently then the behavior of Antarctic SAT, which continues to warm even after stopped emission (Fig.2b and discussion). Here a bit more careful discussion and analysis of the different behaviors after stabilization and the causes behind would be essential. Why is the local Antarctic SAT not relevant?
Â
In general the paper can be published after a bunch of minor corrections.
Â
Some general comments to the figs.:
The yellow line is almost invisible in my printed version, on the screen it looks fine. Changing this into orange could be a compromise.
In the (mostly temp.) anomaly time series (1cd, 2ab, 4a, 7ab) a zero line should be plotted. That makes it considerably easier to assess potential drift.
In some places PI is used as reference, in others ZE-0. This is rather inconsistent. I recommend to plot them both. This would also allow the reader to estimate, whether the drift in the ice sheet has an effect on the southern ocean climate (sea ice, temp) or not.
Â
Detailed comments
37 crucial ?that?
62 specified emission of CO2 be more specific and give the number
96-98 specify the length of the runs
100-104 does not make sense to describe experiments specifically that are not used in the paper. Here a vague hint to more experiments should be sufficient
187 southern hemisphere SAT is inaccurate, you are discussing only the polar SAT. The PI runs shows similar multidecadal variability
192-196 max surface air warming .. is always constrained by the melting temp of the ice surface..
While I can follow this argument for Greenland and higher CO2 levels, where it is at least true in summer, this is almost completely irrelevant for Antarctica. Even in the highest scenarios the melting is restricted to coastal areas. The high elevation areas of the ice sheet are and will be far away from the melting temp and are obviously accumulating happily mass (see fig.7c). Give a better reason!
200 does the physics of sea ice depend really on the cumulated emissions or rather the Arctic SAT, which is linearly related to GSAT and the cumulated emissions? Please give a physically plausible reasoning!
220 How does the sign change of the GrIS mass contribution relate to the time, when the GSAT anomaly becomes negative?
section 3.6 The effect of the residual mass loss on the climate is not shown at all. If it is negligible, great, than please explicitly state this. Showing ZE-0 also in in the climate plots particularly in the south would remove remaining doubts.
339 THe typo
362 here or somewhere else a small discussion would be helpful, that GrSMB does not lead to more ice production for negative GSAT anomalies (Fig. 7d). Please discuss the mechanism(s) behind this.
Â