the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 04 Jun 2026
Interannual variability of the winter sea ice edge in the Southern Ocean tuned by topography and oceanic transport
Abstract. The large-scale interannual variability of sea ice concentration in the Southern Ocean is largely controlled by atmospheric dynamics. By contrast, based on satellite observations, we show here that the local amplitude of interannual variations of the winter sea ice edge position is mainly modulated by the ocean bottom topography and oceanic processes. The standard deviation of the latitude of the ice edge displays substantial variations as a function of longitude, with prominent sharp peaks covering only a few degrees of longitude close to the main topographic features of the Southern Ocean. There, mesoscale eddy activity and the variability of the Antarctic Circumpolar Current jets are large, influencing both oceanic heat transport and sea ice velocity, thereby leading to large interannual changes in the position of the ice edge. Owing to such bathymetric control, these regions showing high variability in the winter ice edge position have remained relatively stable over recent decades, despite ample changes observed in other characteristics of the sea ice cover during the same period. Eddy-rich global sea ice-ocean models based on NEMO-SI3, both forced by ERA5 surface fluxes or coupled with the atmospheric model IFS, can reproduce the sharp peaks in the variability of the ice edge position, indicating that they adequately simulate the dominant influence of topography on currents and eddy activity. However, this requires a realistic mean ice edge position; otherwise, model biases can displace the ice edge away from regions of strong eddy activity and therefore distort the interannual variability.
- Preprint
(2738 KB) - Metadata XML
-
Supplement
(1472 KB) - BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2026-1823', Anonymous Referee #1, 22 Jul 2026
-
RC2: 'Comment on egusphere-2026-1823', Anonymous Referee #2, 08 Aug 2026
The article offers fairly persuasive evidence that the wintertime sea ice edge position variability is related to ocean bathymetry and is therefore mostly steady through time. The authors try to delve deeper into the controls on the sea ice edge variability and find that further attribution by winds can essentially be ruled out. They argue instead that ocean eddies are important. The strengths of the work are its novel first-order findings related to ocean bathymetery, the survey of other possible relationships, and the literature review. The weaknesses are a lack of statistical significance in correlations presented and limited evidence that ocean eddies really do matter.
General comments.
It looks like the regions of above average variance are also regions where the ice edge is least parallel to latitude circles. It would be worth trying to correct for this angle with respect to the latitude circle, so ice edge variability is perpendicular to the ice edge. In other words, reduce the variance by the angle of the ice edge with respect to latitude circles. I don't believe this will eliminate the range in ice edge standard deviation, but I expect it will reduce it.
Many correlations are given throughout the paper, such as for sea ice standar deviation between various time periods and between models and observations. I am suspicious that some relationships presented as different or described as strong versus weak would not pass rigorous stastical tests because there are so few degrees of freedom across the longitudes in Fig 1b. It looks like maybe 30 at best, which is probably not enough for such short records. The discussion of results is somewhat frustrating because this issue is not confronted. In my view, the discussion is still valuable as a potential set of drivers, but the authors should be frank with the reader on which ones turn out to be robust. Despite the appearance of poor stastical significance, the paper frequently claims that relationships are strong or large (i.e., Lines 337-8, 348-9 and 380). I'm especially skeptical of a relationship between geostrophic current and EKE that is claimed at Line 380-1.
The paper says that that the high resolution model's ice edge variability is more highly correlated with the observed ice edge variability than a low resolution model counterpart. However, the difference between the corrrelation is pretty modest. Are they significantly different? Even if so, doesn't the modest difference actually call into question the claim that topography drives variance? The other three high resolution model results are marginally convincing, again due to questions of statistical significance.
The authors give some counter evidence to the "classic view" which they say is sea ice variability is controlled by the atmosphere on in interannual timescales and the ocean on longer timescales. They argue that they've shown the ocean is a player on the shorter, interannual, timescales too. I'm not convinced they've shown it. If the relationship with ice edge and eddies were more convincingly shown, I'd be more persuaded. I was surprised that the authors don't provide an analysis of the eddy heat flux in the ocean to bolster this argument. Is it because they don't have it from observations? An estimate from an ocean reanalysis could be used. I'd also be interested in seeing it in the high resolution coupled model.
Specific comment
Line 41 and 160 and elsewhere have strange scientific notation that is missing a "times" between the numbers. It is unfortunate if this is the journal standard.
Line 157 compares two things with different units and yet says one is larger than the other. Was a percent used?
Line 306 I don't think this logic is reasonable. It would be okay without "thus" in line 306 or with "thus" replaced by "however".
Line 337-8 I'm calling this out again because I don't see this relationship as strong or perhaps even significant.
Line 341 Where is this shown? If not shown, the text should say "not shown" and instead provide a citation.
Citation: https://doi.org/10.5194/egusphere-2026-1823-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 112 | 38 | 17 | 167 | 28 | 11 | 9 |
- HTML: 112
- PDF: 38
- XML: 17
- Total: 167
- Supplement: 28
- BibTeX: 11
- EndNote: 9
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Goosse et al. is a well-written, engaging paper that offers evidence that the role of the ocean in controlling sea ice edge variability is not limited to long timescales, adding an interesting alternative to the classical view of the differing roles of the atmosphere and ocean for controlling interannual sea ice edge variability. This paper explores the pattern of local interannual winter sea ice edge variability, showing how it is represented in observations and in a few high resolution model simulations. The pattern of sea ice variability in observations has prominent sharp peaks that are shown to be controlled mainly by oceanic processes shaped by bathymetry. This pattern of sea ice edge variability has been stable over the past few decades, despite observed changes in sea ice. The two high-resolution model simulations both do a relatively good job reproducing the pattern of interannual variability. This indicates that the pattern of interannual variability is not dependent on the specific atmospheric conditions. Then they look at the specific mechanisms that may be controlling this pattern of variability, highlighting the role of mesoscale eddy activity, ocean heat transport, and the relative location of the Polar Front.
Overall, I think these results are novel and would be of interest to the geophysical community. The methods are sound, and the mixture of observations and model output provides a useful comparison of results and highlights specific model biases. I have a few comments and suggestions which are outlined below, but I recommend this paper for publication after minor revisions.
General comments:
Specific comments:
L56-57: This sentence is a bit confusing, do you mean the variability of the ice edge within a sector of longitudes only covering a few hundred kms or do you mean the variability of ice edge moving latitudinally a few hundred kms?
L80: Do you have any thoughts on whether these mechanisms to apply to other months, especially months where the varaibility is higher?
L103-105: It might be nice to have a Supplementary figure showing the climatology differences from each product across the full period, and reduced down to only the common period. This would support your assumption.
L124-125: Why did you choose a period different from the ERA5-constrained simulation?
L157-164: I found this paragraph a bit confusing. It could be helpful to include a Supplementary figure or schematic accompanying this description.
L197-199: Is this only applicable in the regions of large bathymetric obstacles? Notably, the longitudinal range between 100-150° also has lower standard deviation but has seen a northward shift in ice edge.
L317-319: Why did you use this definition instead of one using water mass characteristics/temperature gradient?
L321-322: Please clarify this sentence, do you mean "far enough away from the front"?
Fig 1a: It may be helpful to write “ice edge mean” for black line to highlight that this is the mean ice edge.
Fig 1a: I would suggest converting longitude labeling on the map to 0-360° labeling to make the comparison with the ice edge variability plot in 1b easier.
Fig. 6 It would be helpful to include longitude markers ranging from 0-360° to align with other figures.
Technical corrections:
Fig 1a and Fig 6: Missing colorbar unit labels.