the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 07 Aug 2025
Quantifying the interplay of Meltwater and Ice-Albedo Feedbacks in the Arctic Ice-Ocean System
Abstract. Sea ice melting generates multiple feedbacks through meltwater release and open water expansion. Due to the tight coupling of the ice-ocean system, these feedbacks are challenging to quantify independently. We employ a well-validated one-dimensional coupled sea ice-ocean model, by removing meltwater or keeping sea ice constant during the melting season, to quantify the independent effects of meltwater and ice-albedo feedbacks on the Arctic ice-ocean system. The experiments reveal the following: (1) Meltwater-induced strong stratification can insulate a portion of solar radiation into the Near Surface Temperature Maximum (NSTM), generating a negative feedback with a feedback factor of -0.19 (i.e., 19 % ice melting reduction). (2) The ice-albedo positive feedback factor is +0.41 (i.e., 41 % ice melting amplification). (3) These two feedback processes exhibit nonlinear interdependence: switching off the ice-albedo feedback reduces the meltwater feedback strength to -0.09, while eliminating meltwater effects enhances the ice-albedo feedback to +0.46. Meltwater effects persist until the following freezing season. The NSTM insulated by meltwater in summer suppresses ice formation during winter in strongly stratified regions. In the weakly stratified western Nansen Basin, summer meltwater release plays an important role in preventing the upward mixing of Atlantic warm water. The meltwater feedback in the ice-ocean system is more pronounced in experiments with thinner initial sea ice, indicating that as Arctic sea ice will continue to decline and Atlantification will intensify in the future, the impact of meltwater on the ice-ocean system is expected to become increasingly significant.
- Preprint
(4001 KB) - Metadata XML
-
Supplement
(5612 KB) - BibTeX
- EndNote
Status: open (until 08 Oct 2025)
- RC1: 'Comment on egusphere-2025-3030', Anonymous Referee #1, 19 Sep 2025 reply
-
RC2: 'Comment on egusphere-2025-3030', Anonymous Referee #2, 26 Sep 2025
reply
General Comments:
The overall objective of the study is to quantify the independent effects of meltwater and ice-albedo feedbacks on the Arctic ice-ocean system. Specifically, the work proposes to anser three questions:
- To what extent do the meltwater and ice-albedo feedbacks influence ice melting during summer?
- What is the contribution of the meltwater (ice-albedo) feedback if the ice-albedo (meltwater) feedback is not involved?
- What are the regional differences in the impact of meltwater on the ice-ocean system?
These research objectives are within the scope of The Cryosphere and, to the best of my knowledge, are novel. The difficulty of the research arises from the strong coupling between the feedbacks of meltwater and ice albedo, which prevents a clear differentiation of the contribution of each. To solve this problem, the authors propose a methodology based on a series of simulations where feedbacks are selectively activated. Independent contributions are then encoded in feedback factors. I consider the methodology appropriate, implemented in a rigorous way and well documented. The supplementary material reinforces its validity with an extensive comparison between the model and observations. However, I believe that its description would benefit from the inclusion of the 1D dynamic equations either in the main text or in the supplementary material.
The presentation and discussion of the results raised my main concerns with the manuscript. Model results are quantitatively presented but, in main opinion, poorly interpreted from the physical point of view. In some sections (see below), the explanation of some model results would require a qualitative interpretation or a physical hypothesis. The quantities are only indicated and at most justified with references or general assessments. Brief physical explanations would be very informative.
I consider that a revised version of the manuscript could be suitable for publication in The Cryosphere.
Specific Comments:
L 55: In winter, brine rejection caused by ice formation leads to the upward entrainment of heat from the NSTM and AWW and impedes winter ice formation consequently (Smith et al., 2018; Steele et al., 2011; Timmermans et al., 2017), which is a negative feedback known as the ice production-entrainment feedback (Goosse et al., 2018). My understanding is that the ice production-entrainment feedback inhibits vertical heat transfer, promoting more ice production. could you confirm this point?
L 72 Zhang et al., (2023) demonstrated that the removal of meltwater increases ice melt by 17% Could you add briefly the physical mechanism proposed by Zhang et al. to justify this effect?
L 75 Indicate the differentiated aspects considered in the work between meltwater and icea albedo feedbacks. There is a strong relationship between both and it is not clear their intrinsic and differentiated components.
L 102 I would suggest to include the mathematical expression of the 1D model.
L 116: A boundary condition is always required to get a particular solution to the differential equation. Please, clarify the sentence No boundary condition is applied at the bottom of the model.
L 175 Based ON the feedback factor (γ) framework proposed by Goosse…
L 177 Total response by Total Response
L 179 (Reference Response, RR)
Figure 3. Please define the acronyms in the figure caption. The figure is cited before they are defined in the text.
L 264-265 the brine rejection process during winter creates higher ocean-ice heat flux, which in turn inhibits ice formation. Could you briefly clarify the mechanism of this? As I understand it, brine rejection contributes to the formation of the cold halocline that isolates the upper Arctic layers from the warm waters of the Atlantic. It would favor ice formation in this case. Could you clarify your assement?
L 270 the removal of meltwater results in a 0.19 m increase in ice formation compared to the CTRL run. Why? How do you physically interpret this result?
L 273-275 The noMWIA run also shows less ice formation compared to the noIA run (Figure 5b) in station NS2, NS4 and NS5, but no sea ice melting is observed during winter, further demonstrating the importance of meltwater in weakly stratified regions. What is the physical explanation for this result?
L 277 This suggests that sea ice retreat during summer can promote ice formation in winter. Can you give a physical reason based on your model? Is it just a geometric effect of having more ocean surface available to freeze.
L 291 the NSTM, because the heat absorbed by the ocean can mix sufficiently within the upper 30 m. This explanation would disagree with observations. The NSTM layer can be observed at this depth in the Canada Basin (Jackson 2010). Provide an explanation to the discrepancy between the model and observations.
L 297 …agree well with the observed values…
Figure 6.Add month labels to the figures corresponding to station BS5, similarly to those of NS2
Figures 6 a,e,I,m,c,g,k and o. The discontinuity in the MLD seems an artifact of the surface tracking algorith due to the appearance of a new water mass, the meltwater. A gradient tracking algorithm could identify this as the bottom of the mixed layer. There is no discontinuity in the rest of simulations because the meltwater input is cancelled. Could you clarify this point?
L 305 How do you explain this big impact of meltwater in the mixed layer depth during the freezing season? Is there any observational or 3D model result that corroborates it?
L 390 This is a process similar to the noMW run, where less meltwater entering the ocean accelerates sea ice melting. Meltwater from ponds, sooner or later, flows into the Ocean. This might slow or delay the formation of a meltwater layer, but probably not inhibit it like in the study.
L 455 it delays winter ice melting by slowing surface cooling. Should be warming instead cooling?
Model code and software
1D-model Haohao Zhang https://github.com/HaohZhang/1D-model
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 1,814 | 50 | 15 | 1,879 | 22 | 59 | 58 |
- HTML: 1,814
- PDF: 50
- XML: 15
- Total: 1,879
- Supplement: 22
- BibTeX: 59
- EndNote: 58
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
I find that this is an interesting study that provides insights into the role of seasonal sea ice melt(water) on the evolution of sea ice and the upper ocean across the deep Arctic basin. It uses a 1D coupled ice-ocean model that has been used in earlier somewhat similar studies. It nicely examines the sensitivity of the coupled system to seasonal sea ice meltwater release and ice-albedo feedback. It is also generally well written. And my only "objections" in this first round are just that authors should make clear of the "caveats" of such a simplified approach when extrapolating the results. I elaborate on these points below. Beyond that I think that this work fits perfectly in the scope of TC, and is publishable with minor revisions given the authors also reflect on the points I raised below in the Discussion and Conclusions of the revised manuscript and place the work properly in context of references given.
Generic comments
The Introduction could benefit from noting on the very small scale meltwater layers that form below sea ice (Smith et al., 2023; Salganik et al., 2023a), and aren't necessarily resolved by the model setup, and are perhaps more stable than in the model setting(?). Salganik et al. (2023a) also provides a direct estimate of the effect of such meltwater layers on the summer melt of (level) sea ice, albeit from very short-term observations, which might be interesting to compare to. Further Perovich et al (2021) and Smith et al. (2025) discuss the fate of sea ice meltwater based on direct observations, and latter indicates that 10% of the meltwater does not enter the ocean, but that would have a small impact on your results, would it? I would think the findings from these recent studies would be good to introduce and used to place this work better in context.
From what I understand, the work conducted assumes all ice is level ice, and this is obviously not the case in reality. Deformed ice (ridges, rubble etc.), can easily have an areal fraction of >30% (e.g. Brenner et al., 2021, see their Suppl. Material), and this is known i) affect (limit) the spreading of sea ice meltwater and that ii) ridges melt much faster than level ice (e.g. Salganik et al., 2023ab) - so I would perhaps at least note this fact and be somewhat careful when extrapolating the results, as the "meltwater" layer effect likely applies only to a fraction of the ice cover - not the full ice cover - given thicker ice might melt much more rapidly, so the overall effect might be much less than proposed here.
Lines ~40-50: I am not an oceanographer, but I find this "freshwater" budget presented here, somewhat misleading. When you look at the salinity profiles you use as initial conditions, there is a large freshwater inventory already before sea ice melt onset, which from my understanding is primarily meteoric water (river + net P/E). This meteoric inventory is order of 10+ meters (e.g. Bauch et al., 1995; Dodd et al., 2012) or even more in places, that has accumulated over time and to me its this freshwater component that sets the scene for the over all layering in the Arctic Ocean, not the seasonal sea ice melt. Its the former (aided by sea ice formation) that creates the halocline that is the barrier between AW and the surface. To my understanding sea ice melt creates the seasonal shallower mixed layer. And with the exception on parts of Nansen basin that do not have effective sources of river runoff, as you also note. Sea ice formation also plays a crucial role in distributing this freshwater. Seasonal sea ice meltwater is then probably more important for the NSTM, given how deep solar radiation can penetrate the ocean. Anyway, I felt that this part needs some clarification in this regard.
Albedo - I think the main text should include a brief description of how the model treats sea ice albedo (also add a panel of albedo in Figure 4), with reference to the Appendix. Given that you try to single out the effect the of albedo, this should be more thoroughly explained in Section 2, and showing it in a panel in Figure 4 would also be very helpful to assess the albedo evolution in the model.
I would like that you assess better how the above factors affect the generalization of your results in the revised manuscript. The Conclusion now lacks any appreciation of how the "real world" might differ from the simplified model experiments, nor compare to earlier findings, I would find that to be appropriate for the benefit of the reader (its also rather superficial in the Discussion as well). It would also be useful to advise future studies what could potentially be done better/different.
Appreciate the illustrations esp. Figs 2 and 10. Nice work.
Specific comments
L101 - in this section (someplace in the main text) I think it would be good to briefly note on how the surface albedo of ice/ocean is treated in the model, especially for the ice. This is probably important when considering the relative impact of meltwater or albedo.
L143 - snow depth, how does this evolve over time, does it reappear in fall? I did not quite understand that, but perhaps you can add a sentence to clarify. While not so important for melt, snow will definitely also impact ice growth.
L145-160, Fig 1 & 2. Grouping of stations and initial ocean conditions - what caught my eye is the profile for NA-5, this seem very different than anything around in (NA or AM), and thus wonder, have you checked other profiles in this area to see whether this is a recurring type of profile. I would have based on location usually expected something more like the rest of the NA or AM profiles. NS-1 is what I would say typical Transpolar drift profile you find in AM/NA, and here using the choice of geography to categorise the station is a bit misleading. But I assume a single station does not make any difference to the final interpretation for the regions.
L167-169 - Somewhat confusing, first you say that you use "two subregions" but the map shows one single area with the red line in the map. Is the forcing averaged over the whole area and that is used as forcing at each site.
L170-171 - Later you state this external FW flux is very small compared to sea ice meltwater. If you run the experiments with zero external freshwater are the results any different? I also find that this term is a bit awkward, given that any advective FW (relative to ice motion) would not be only at the surface, but in the whole upper water column rather than a continuous flux at the very surface. The magnitude when averaged over the Arctic maybe is as such reasonable, but the way its implemented not.
L171 - how does this value compare typical sea ice meltwater fluxes? And if understand correctly this is never changed in the four experiments?
L181 - do you mean "sea ice meltwater discharge" or does this also mean the "external freshwater" is zero, please clarify.
L184 - you mean "sea ice meltwater flux" - just to be consistent in the use of terms
L187-188 "results are presented in Section 3."
L191 - specify whether its only sea ice meltwater set to zero, or also external meltwater. Just to be sure. Thank you.
L205 - please specify what meltwater
L226-227 - I assume SWpenetrate is the the energy transmitted through the ice to the ocean? I assume the more correct term is then "transmitted".
L230 - Fig. 4. - I wonder if it would be useful to show the panels in the following order, SIT, SIC, Fsw, SFW, given the former two drive the latter.. and could add a weak horizontal line at zero in current panel d. Also in panel d, indicate the "external freshwater flux". And the legend would preferably be placed in the top panel.
L236 - change to "initial ocean profiles"
L238 - change to "impact of initial ocean profiles"
L238-240 - In light of this, I would suggest you add station NS-2 to figure 4, so the left hand side is for BS-5 and right hand side fir NS-2.
L239-240, the word "condition"*2 can be deleted.
L280 - Fig 5. I would think it wo8uld be much simpler if you grouped the stations by the same color as in Fig 5, I think the within group variation is only large for the NS stations (and perhaps the one AM station), so would be probably enough to use fewer colors. Applies also to Fig 7 and 9, especially in 9 they dots are overlain and does not make a difference showing each station with a different color in my opinion.
L320 - Fig 6 - a closer look at the typical types of ocean profiles (Fig. 2), it would be informative to show here also one station that is "in between", and to me this would perhaps be AM-1 or AM-2. In terms of less surface freshening than BS-2, but also a different initial heat content than BS-2. I would think this is useful to have in the main text.
L386 - Stating "is reliable" sounds overconvincing yourself. Rather, I would rather phrase this in some more insightful way, how you can "tease out" the possible contribution of different factors. But I would still have some doubts that using e.g. a uniform forcing all across the Arctic, might not be representative in regions with much atmospheric activity and strong synoptic events, which are taken out by averaging the forcing? Then especially again the Atlantic sector., e.g. Graham et al. (2019).
L187-393 - As noted in the generic comments the work of Perovich et al and Smith et al are relevant mention here. Also the fact that with melt ponds you significantly increase the transmission of solar radiation to the ocean and not only absorption to ponds/ice and decrease the albedo. Refer to Nicolaus et al. (2012).
L400-401 - Duarte et al. (2020) point towards synoptic events being important in the Nansen basin region, how does your "arctic wide averaged forcing" mean for this type of single-events? And that single-event ocean heat fluxes (up to 400 Wm2, see Duarte et al. 2020 and references therein) are important in this region (see also Graham et al., 2019).
L419-420 - How does this relate to the observations of Lind et al. (2018) and Skagseth et al. (2020)? Are these examples of conditions that could prevail in the Eurasian basin in the future? How are they captured in the model experiments, please elaborate.
L396-423 - In general this seems to point me to the fact that advective terms can be very important in the Nansen basin case of this work? Omitting those, could possibly distort the results presented here signficantly? Given that ice is always transported into this region with the Transpolar Drift often replacing melted ice, providing more potential for meltwater sources, and heat is also continuously also transported with Atlantic boundary current "replacing" lost ocean heat.Conclusions
L443 - "well-validated" is subjective, and should be deleted here IMHO.
L444 - "sea-ice meltwater" - again specify for the benefit of the reader (applies to whole manuscript).
Conclusions in general - As noted in the generic comments the results need to be better in context of possible shortcomings I noted in the generic comments, e.g. in relation to the fact you only represent the whole ice cover as level ice?, relative to how albedo is treated in the model (and relates to observed albedo), uniform atmospheric forcing vs possibly very regional conditions (esp. in storm tracks in the NS region), and sea-ice meltwater balance in the model vs. observations etc. And what are your recommendations for improving this in future work?
References
Bauch, D., Schlosser, P., & Fairbanks, R. G. (1995). Freshwater balance and the sources of deep and bottom waters in the Arctic Ocean inferred from the distribution of H218O. Progress in Oceanography, 35(1), 53–80. https://doi.org/10.1016/0079-6611(95)00005-2
Brenner, S., Rainville, L., Thomson, J., Cole, S., & Lee, C. (2021). Comparing Observations and Parameterizations of Ice‐Ocean Drag Through an Annual Cycle Across the Beaufort Sea. Journal of Geophysical Research: Oceans, 126(4). https://doi.org/10.1029/2020JC016977
Dodd, P. A., et al. (2012). The freshwater composition of the Fram Strait outflow derived from a decade of tracer measurements. Journal of Geophysical Research: Oceans, 117(C11), n/a-n/a. https://doi.org/10.1029/2012JC008011
Duarte, P., et al. (2020). Warm Atlantic Water Explains Observed Sea Ice Melt Rates North of Svalbard. Journal of Geophysical Research: Oceans, 125(8). https://doi.org/10.1029/2019JC015662
Graham, R. M., et al. (2019). Winter storms accelerate the demise of sea ice in the Atlantic sector of the Arctic Ocean. Scientific Reports, 9(1), 9222. https://doi.org/10.1038/s41598-019-45574-5
Lind, S., Ingvaldsen, R. B., & Furevik, T. (2018). Arctic warming hotspot in the northern Barents Sea linked to declining sea-ice import. Nature Climate Change, 8(7), 634–639. https://doi.org/10.1038/s41558-018-0205-y
Nicolaus, M., Katlein, C., Maslanik, J., & Hendricks, S. (2012). Changes in Arctic sea ice result in increasing light transmittance and absorption. Geophysical Research Letters, 39, L24501. https://doi.org/10.1029/2012GL053738
Perovich, D., Smith, M., Light, B., and Webster, M.: Meltwater sources and sinks for multiyear Arctic sea ice in summer, The Cryosphere, 15, 4517–4525, https://doi.org/10.5194/tc-15-4517-2021, 2021.
Smith, M. M., et al. (2023). Thin and transient meltwater layers and false bottoms in the Arctic sea ice pack—Recent insights on these historically overlooked features. Elem Sci Anth, 11(1), 1–41. https://doi.org/10.1525/elementa.2023.00025
Smith, M. M., et al. (2025). Formation and fate of freshwater on an ice floe in the Central Arctic. The Cryosphere, 19(2), 619–644. https://doi.org/10.5194/tc-19-619-2025
Salganik, E., et al. (2023a). Temporal evolution of under-ice meltwater layers and false bottoms and their impact on summer Arctic sea ice mass balance. Elementa: Science of the Anthropocene, 11(1). https://doi.org/10.1525/elementa.2022.00035
Salganik, E., et al. (2023b). Observations of preferential summer melt of Arctic sea-ice ridge keels from repeated multibeam sonar surveys. The Cryosphere, 17(11), 4873–4887. https://doi.org/10.5194/tc-17-4873-2023
Skagseth, Ø., et al. Reduced efficiency of the Barents Sea cooling machine. Nature Climate Change, 10(7), 661–666. https://doi.org/10.1038/s41558-020-0772-6
Other relevant literature that might be useful to include
Van Straaten, C., Lique, C., & Kolodziejcyk, N. (2025). The Life Cycle of the Low Salinity Lenses at the Surface of the Arctic Ocean. Journal of Geophysical Research: Oceans, 130(4), e2024JC021699. https://doi.org/10.1029/2024JC021699