the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Quantifying the Role of Sea Ice in Seasonal and Interannual Variability in Surface Stress and Mixed-Layer Entrainment in the Central Canada Basin
Abstract. Over the last few decades, the Beaufort Gyre region (BGR) has experienced significant ice loss, leading to changes in momentum transfer within the atmosphere-ice-ocean system. We evaluate monthly and interannual variability of surface stress in the BGR, and a sub-region (the Central Canada Basin, CCB) within it, from 2003–2023. We find an increase in surface-ocean stress magnitude in the CCB of 0.007 ± 0.001 N m−2 per decade, with the largest linear increases in November–December. We demonstrate that the shifting sea-ice regime, characterized by greater mobility and increased responsiveness to wind forcing, is the primary driver of increased surface-ocean stress, as opposed to changes in the wind field itself. Anomalously High-Stress Events (HSEs, characterized by stress magnitudes an order of magnitude larger than background levels) in the CCB occur most frequently in fall and have increased in frequency over 2003–2023, contributing to an increase in annual time-integrated surface stress of 0.54 ± 0.45 N m−2 per decade during HSEs. We examine the implications of increased sea-ice mobility and changing surface-ocean stresses on the ocean mixed layer using an energy-balance entrainment parameterization. We find that stress-driven entrainment can account for ∼65 % of October–March deepening, while buoyancy-driven processes contribute ∼7 % of deepening during freeze-up. Mechanical mixing during HSEs contributes to ∼28 % of all stress-induced mixing, even though HSEs occur less than 7 % of the time during the 2003–2023 period. In November and December, the only months with statistically significant increases in surface stress, we estimate that increased surface stress can account for ∼0.15 and 0.1 m year−1 more deepening in each month over 2003–2019. This is ∼20 % of the observed ∼0.7 and 0.6 m year−1 increased deepening in November and December over the CCB for 2003–2019, and we infer that other physical processes dominate the observed mixed-layer deepening trend. These results highlight how sea-ice decline and mobility enhance momentum transfer, suggesting a potential mechanism for a positive feedback via increased ocean-to-ice heat fluxes.
- Preprint
(14327 KB) - Metadata XML
-
Supplement
(4603 KB) - BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2026-3283', Anonymous Referee #1, 23 Sep 2026
-
RC2: 'Comment on egusphere-2026-3283', Ilker Fer, 24 Sep 2026
The manuscript presents an analysis of openly available datasets, including hydrographic profiles from ice-tethered profilers, wind data from ERA5 reanalysis, and remotely-sensed sea-ice velocity and concentration, and surface geostrophic currents. The study site is a subregion (central Canada Basin, CCB) of the Beaufort Sea gyre region. The methodological framework builds largely on previous studies using similar datasets from this region and is generally well referenced. The novel aspect of the work lies in its examination of the evolution of mixed‑layer properties and the relative roles of surface stress and convective forcing. In addition, the authors identify and quantify periods of elevated surface stress and assess their contribution to mixed‑layer evolution.
The manuscript is well written, and the figures are both informative and visually effective. The figures support the narrative well, and the narrative supports the results and conclusions. The manuscript will make an important contribution to the literature after some revisions.
Overall, I would argue that the manuscript is a better fit for Ocean Science than for The Cryosphere. I am probably biased because I serve as an editor for Ocean Science. Following the editor’s suggestions, the authors have introduced several statements intended to emphasize the relevance of the work to The Cryosphere. However, I do not find these additions particularly convincing. The authors could consider a transfer to Ocean Science.
Because I refer below to some of my own work, I prefer to sign this review and leave the question of relevance of these studies to the authors.
I find the bulk-mixed layer approach highly simplistic and associated with considerable uncertainty, although I also recognize that it may be adequate for this manuscript. I would encourage the authors to expand their discussion of these uncertainties and, more importantly, to provide sensitivity analyses for key empirical parameters. For example, it would be useful to assess how the results change when the coefficients m0 and C as well as the drag coefficient and Ekman used in the surface stress calculations (and therefore friction velocity), are varied by a factor of two. Even such simple experiments would provide a clearer picture of the robustness of the conclusions.
The structure of the manuscript could also be improved. At present, descriptions of datasets and methods are distributed across several sections. For example, Section 2 focuses on surface stress but also contains much of the data and methods description, while sea‑ice thickness observations from the BGOS moorings and the mixed‑layer depth estimates are introduced later in Section 4. I suggest consolidating the descriptions of datasets and methods into dedicated sections. Process‑specific calculations, such as mixed‑layer entrainment estimates, could remain within the results sections where they are discussed.
Finally, I find the literature review somewhat biased toward studies conducted in the Pacific sector of the Arctic Ocean. Several relevant studies from other regions and perspectives appear to be missing. For example, one of the central conclusions of this manuscript regarding the relative importance of convection versus entrainment from below the mixed layer is closely related to one of the key points listed in my 2017 paper, referenced below. A broader review of the existing literature would help place the present results into a more balanced context.
My specific comments follow.
Kind regards,
Ilker Fer, University of Bergen
Specific comments (in order of appearance)
Abstract- “HSEs… in the CCB …have increased in frequency over 2003–2023» This is only one event increase, if I understood the text correctly.
Abstract- the last statement referring to a “positive feedback” is not adequately supported by the discussion in the main text and does not appear among the findings summarized elsewhere in the manuscript. I therefore recommend either removing this statement from the abstract or expanding the discussion section to explicitly address and support this proposed feedback mechanism..
Li 10- “energy-balance entrainment parameterization” could be more accurate; this is a vertically integrated bulk mixed layer param based on …
Li 45- your review is biased to the Pacific sector of Arctic in general. Here, e.g., could consider: Graham, R.M. et al. Winter storms accelerate the demise of sea ice in the Atlantic sector of the Arctic Ocean. Sci Rep 9, 9222 (2019). https://doi.org/10.1038/s41598-019-45574-5
See also: Fer, I., et al. (2017, One-dimensional evolution of the upper water column in the Atlantic sector of the Arctic Ocean in winter, J. Geophys. Res. Oceans, 122, doi:10.1002/2016JC012431). One of their key points is “Increase in mixed-layer salinity from freezing (10%) is significantly less than that due to entrainment (90%)”, comparable to one of your conclusions.
Surface stress calculation, eq. 1-3 and drag coefficients: quadratic stress parameterization, rotation of uEk and under-ice drag coefficients all become suspect when there are significant buoyancy fluxes, such as from melting. See Reifenberg et al (2025, Turbulence Observations below Drifting Sea Ice: TKE Production and Dissipation in the Meltwater-Influenced Boundary Layer, J. Phys. Oceanogr., 55, 451-470, https://doi.org/10.1175/JPO-D-24-0102.1), particularly their discussion on drag coefficient, and also summary point 2 of their main finding, which can be linked to your findings.
Fig 1. Mention the source of data (ERA5)
Eq2. Please mention why you include the Ekman velocity in the reference velocity
Ekman depth of 20 m seems arbitrary as this will be dependent on the surface stress
Li 130- type on McPhee
Fig 3. I think you show the PDF and 80th percentile of stress using all individual daily estimates (not spatially averaged over CCB). You could clarify this in the caption.
Li 173. You could consider a comment related to the free-drift ratio
Lines 180-200, comparisons to Martin et al. results. If I recall correctly, it is a pan-Arctic study. Please clarify if you’re comparing CCB results to an Arctic average. This may also have a consequence on the discrepancy (line 195) you attribute to different analysis periods.
Li 221: if I understand correctly the increase in frequency of HSEs is from 6 to 7 per year. I am not sure whether this one extra event is significant (both in terms of number and also of impact). Also, I cannot follow how one additional event can increase the mean surface stress from 0.11 N/m2 by 0.54 N/m2 per year, which appears difficult to interpret, as the increase is approximately five times larger than the reported mean annual value. The text refers to a “time‑integrated surface stress,” but this is not reflected in the units provided. I suspect that I may be misunderstanding the quantity being reported, which suggests that additional clarification is needed.
Li 240: the definition of buoyancy frequency needs more care.
Is upper 50-m too restrictive? Profiles suggest N2max may be deeper
Li 263: please double check if this statement is correct: “the mixed layer has exhibited a shoaling tendency, which coincides with a decrease in freshwater … content”. If MLD is shoaling, I would have expected an increase in freshwater content.
Eq.5 and li 303 : According to the equation, if B>0 then ice is growing (i.e., brine release and convection). This contrasts with the sign of the second term in Eq.4. Please explicitly mention the sign convention where you write “… B can be positive (destabilizing buoyancy flux from ice growth?) or negative (stabilizing buoyancy flux from ice melting?)”. This is also in conflict with where you write “If B > 0, C is set to 1, while for B < 0, C is set to 0.05 (the release of dense brine leads to convective mixing)”. It should be the other way around, I think. Please carefully check through. Are the calculations correct?
For the convective case, C=0.05 is a representation of penetrative convection in the absence of mean shear (I think this should be better mentioned in the text). Stigebrandt uses 0.05, slightly increasing Farmer (1975)’s average value.
The statements: “During the ice growth season (October through May), wBe remains weakly positive and small» and “…while buoyancy-driven entrainment contributes only 7% (Figure 11a). Hence, surface momentum transfer plays a larger role..” may be sensitive to the choice of m0 and C.
Li 296: “m0 is a constant (empirical) mixing efficiency” It is not a mixing efficiency but an efficiency factor that scales the wind stirring effect. Please clarify.
m0 coefficient is tuned to 0.6 to simulate the climatological mean annual cycle of mixed layer T/S in the Baltic Sea. In his own application to the Arctic Ocean mixed layer, Stigebrandt (1981) applied m0 = 1.25. This roughly doubles your stress-based term. How are we going to deal with this uncertainty? It needs perhaps a recalculation and sufficient discussion.
Li 336: “Our estimates of shoaling as a result of ice melt are typically larger (∼3 m per month on average) than observed decreases in mixed-layer thickness in the melt season”… The stress part of your simplified bulk-mixed layer approach (Eq.4), will probably not be accurate when there are large stabilizing buoyancy fluxes. For example, measurements from the MOSAiC drift showed buoyancy effects in ocean-ice drag coefficient estimates [Fer et al JGR (2022), Upper-ocean turbulence structure and ocean-ice drag coefficient estimates using an ascending microstructure profiler during the MOSAiC drift. https://doi.org/10.1029/2022JC018751]
-------
Citation: https://doi.org/10.5194/egusphere-2026-3283-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 153 | 99 | 26 | 278 | 63 | 24 | 22 |
- HTML: 153
- PDF: 99
- XML: 26
- Total: 278
- Supplement: 63
- BibTeX: 24
- EndNote: 22
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Please see attached review