the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 11 Feb 2026
Model Assessment of Winter Extratropical Cyclone Short-Term Impacts on the Antarctic Marginal Ice Zone
Abstract. The Antarctic marginal ice zone (MIZ) is the transitional region between the Antarctic sea ice edge and the consolidated ice cover, which is characterised by the presence of ocean surface waves and relatively small ice floes, and is a region where atmospheric and oceanic processes strongly influence sea ice dynamics. Extratropical polar cyclones intensify these processes by amplifying wave activity, transporting heat and moisture, and driving sea ice drift across the MIZ. Here, the CICE sea ice model with a wave propagation module is used at a 0.25° resolution to analyse statistically the impact of ∼400 cyclones on the location of the Antarctic ice edge and MIZ width. Cyclone-driven winds cause sudden shifts of ∼20 km in the ice edge location, through both compaction and expansion, with expansion events more effective in early winter and compaction later. MIZ widening is primarily driven by short-lived, major wave-induced breakup events, which increase the MIZ width by ∼30 km on average, with greater breakup associated with greater widening. Extreme ice edge changes result from the combined influence of sea ice drift and thermodynamics, whereas extreme MIZ width changes are primarily governed by the presence or absence of wave activity. These findings highlight the varied yet pronounced responses of the Antarctic MIZ to extratropical cyclones and underscore the critical role of waves in shaping the boundary between the MIZ and the consolidated ice pack, reinforcing the need to account for wave impacts in future studies.
- Preprint
(14483 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-5209', Anonymous Referee #1, 07 Apr 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2025-5209/egusphere-2025-5209-RC1-supplement.pdfCitation: https://doi.org/
10.5194/egusphere-2025-5209-RC1 -
RC2: 'Comment on egusphere-2025-5209', Anonymous Referee #2, 02 Jul 2026
Title: “Model Assessment of Winter Extratropical Cyclone Short-Term Impacts on the Antarctic Marginal Ice Zone”
Authors: Noah S. Day, Siobhan P. O’Farrell, and Luke G. Bennetts
Short Summary:
This paper uses a standalone CICE6 configuration with a waves-in-ice module (CICE6-WIM) at 0.25° resolution to examine the synoptic-scale response of the Antarctic marginal ice zone (MIZ) to intense cyclones, analyzing hourly output for July–November of 2015–2018. The manuscript focuses on the evolution of sea ice properties in the MIZ and changes in MIZ width, presenting case studies of thermodynamic and dynamic processes related to sea ice drift and wave-induced ice breakup. They use an external cyclone-tracking algorithm to drive their modeling domain and to derive a statistical analysis. The study is a well-motivated extension of the 1° climatological configuration of Day et al. (2024) to the synoptic scale, and the focus on cyclone-driven ice processes is valuable. The configuration is clear, as well as the limitations of the model. My comments are about internal consistency with the Appendix and a few places where the wording could be tightened.
Introduction: Well-structured and citations are up to date. Just ensure that the cited preprints are updated to their published format.
Major comments:
Methods:
- The attenuation coefficients (c₁, c₂ in Eq. for α) and their values / SIC-scaling are not given, so the WIM cannot be reproduced from this paper alone. Please state the coefficients or reference the companion paper explicitly.
- The cyclone selection filter reduces 136,871 tracks to 396. Since all statistics rest on these 396 storms, the bias check described as "not shown" should be shown, at least in the supplement materials (spatial, central-pressure, and radius distributions of retained vs. discarded storms).
Case studies:
- The first case study and the appendix (C1) describe the same event in opposite ways. Please clarify:
- The main text emphasizes melt and says average ice age increases, but C1 says the same event shows "substantial new ice formation" and a decrease in ice age. The appendix (new ice, age decrease) matches Fig. C1.
- The main text mentions that the warm air is "heating the ocean surface," but Fig. C1b shows sea surface temperature moving only from about −1.899 to −1.902 °C (0.003 °C change), at the freezing point, that then cools. There is effectively no ocean warming. Please remove or reword this.
- The main text gives two different concentrations "at the time of crossing": ~80% and ~40% . Fig. C1d shows ~0.82–0.85, supporting 80%. The heat-uptake mechanism relies on the 40% value, which the figure does not support. Please reconcile or clearly state that 40% applies only to the outermost cell.
- The " basal melt" claim in the main text is stronger than the paper's own evidence. Fig. C1c shows ice thickness flat at ~0.15 m through the event, so basal loss is not visible.
- The word "controlling" overstates the case-1 result. §3.1 says thermodynamics is "controlling ice edge location," but Fig. 2h shows only that edge position and net melt/growth move together in time.
- Edge-motion attribution: make case 1 consistent with case 2.
- Case 2 (§3.2, Appendix C2) correctly ties edge motion to the meridional (v) wind. Case 1 instead argues the edge motion is thermodynamic because winds are "u-dominated" (C1). But Fig. C1 shows the v-wind still reaches ~−10 m s⁻¹, which is enough to move a thin (0.15 m), low-concentration edge. "u is larger" does not mean "v is negligible," since it is v that moves the edge north–south. Since the authors already apply the edge-velocity-vs-wind check in case 2, they should apply it to case 1 as well. It will either confirm a genuinely melt-dominated case or reveal a contribution from drift.
Statistical results:
- The "melting/freezing" event labels are defined by a freezing point, which is misleading. Tice > −1.8 °C defines "melting events," but Fig. D1b confirms −1.8 °C is the ocean freezing point, not the ice-surface melting point (~0 °C). So a "melting event" does not indicate surface melt, it indicates the ice surface sitting at or above the freezing point. Moreover, since the dominant melt mode is basal, a threshold on surface temperature does not track the process which is also why melt occurs during "freezing" events. Please rename these events or state explicitly that the threshold marks a surface thermal state, not surface melting or a growth/no-growth boundary.
In addition, in line 289: "Melt during freezing events exceeds July–August melt during melting events" is confusing and mainly reflects season, not the events. By November, basal melt is high regardless of the surface-temperature label. This actually shows the surface label does not capture basal melt, worth stating plainly, and rewording "heightened ice vulnerability" to name what is increasing.
- The seasonal reversal in ice-edge change (advance in Jul–Aug retreat by Nov, Fig. 8c) is a clear trend in the medians, but the interquartile ranges include zero every month, so it is not yet shown to be significant. Since the claim is a sign change across the season, a bootstrap testing whether the July and November medians differ from each other and from zero would substantiate it.
- 10b shows major-breakup widening and minor-breakup narrowing largely cancel, so the net cyclone effect on MIZ width is near zero each month. Could this be stated more clearly in the results? This would also right-size case study 3: major breakup is common but case-3-scale widening is a tail event (Fig. 10a), not the typical response.
- With only two predictors, a random forest is a heavy choice, and its smooth background surface can give an impression of confidence not supported by the underlying data density, particularly the narrow-MIZ hotspot, which rests on few points. Additionally, the two predictors are not independent, since a wider MIZ attenuates more wave energy before it reaches breakable ice, so edge Hs and MIZ width jointly control breakup. This collinearity is likely why the relative contributions are challenging to quantify and is worth stating explicitly. If the random forest is retained, please report its validation (train/test error, per-region sample sizes) so the result is not overread.
- The authors note that the CICE6 welding rate is orders of magnitude below the Roach et al. (2018b) estimate, which adds uncertainty to MIZ-narrowing timescales. Could you go one step further? Since the low rate means the model under-welds, the modelled reconsolidation (e.g., case study 3) is plausibly biased slow or incomplete. Conversely, if reconsolidation is in practice dominated by new-floe formation, the welding error may not affect the result. A one-line statement of which process dominates the narrowing, and in which direction the low welding rate biases it, would resolve this.
Minor comments:
- Line 84: Please correct to: elastic-viscous-plastic rheology
- Line 176: "During the cyclone passage, the ice edge from 60.4° S to 61.1° S" is missing a verb
- Lines 209–211: the wave-fracturing sentence is duplicated.
- Line 249: "Melting events occurs" -->"occur."
- Line 359: "leave the the MIZ" -->typo
- Line 373: "a the mixed-layer" -->typos
- Line 450: Probably means decreases in mean sea ice thickness and age.
- "Frazil and congelation at comparable rates" (§3.1) overstates Fig. 2g, where frazil is initially ~3× congelation.
- Report the actual correlation (r = −0.75, already in Fig. 11b) in the text rather than "moderate to strong."
Citation: https://doi.org/10.5194/egusphere-2025-5209-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 879 | 656 | 77 | 1,612 | 178 | 194 |
- HTML: 879
- PDF: 656
- XML: 77
- Total: 1,612
- BibTeX: 178
- EndNote: 194
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1