Please could you submit your revised manuscript. I will then decide if this needs to be reviewer by either reviewer 3 or an alternative reviewer. I would like to see the revisions in the manuscript so I can assess for myself if your findings are significant or not.

Many thanks for your detailed responses to all three reviewers. I can see that this has likely greatly improved the flow of the paper.

With very best regards,
Jenny

I asked for a second review of this revised manuscript. The review is not favorable, and I would have to reject the paper if the major points raised are not addressed. I would like to hear your opinion. Especially if you consider that you can improve the paper. Many of the issues are related to writing style, however the main point that the presentation of the data is not backing up the scientific results claimed is one that I was also concerned about.


Specific points

All line numbers refer to the track change manuscript.

Abstract line 4: either "the cause of the opening is not yet determined" or "the cause of the opening is much debated".

I would caution you not to use T3b, T4 and T5 in the abstract as this is jargon not many readers may understand.

Page 4 Line 124: "so not for us" is colloquial.

Page 6 line 134: "is only directly available"

Page 6 line 165: "so-called" is colloquial.

Page 16 line 287: "could also be warm air"
line 298: "snow layer" duplicated

Page 19: line 398: "so it is more likely that the"
line 399, this sentence is awkward

Thank you for your kind acknowledgements fully recognising the support you have recieved.

I would like to add that I agree with the authors regarding an unfortunate tone in reviewer 2's review refering to two groups as 'not clever'. I would like to take this opportunity to say I hope this was an oversight in the reviewers behalf and not aggressive, and I would highly recommend checking language for such aggressively dismissive statements in the future. The authors have suggested more appropriate language that would have conveyed the same meaning with out the loaded implications of superiority.

Jenny

hank you for the improvement in the clarity of this paper.
While I agree with the reviewers that your findings are limited, I am happy to see that you have refocused the paper to show the limitations and the results your analysis supports.

Your revision is a substantial rewrite, and as you have included new analysis I have taken careful consideration of the changes. Most papers that are in their third round of comments are at the point of only needing technical revisions, however with new results presented the paper requires a full review, which I have done myself. I find that revision is required.

You have developed a method to forecast polynya opening over Maud Rise based, and have demonstrated that this forecast does not have false positives in the satellite record. It is also useful to know if there are false negatives, and what the occurance of this is. Looking at figure 5 I think there maybe some small overlap where the standard deviation in brightness temperature of the three channels is similar to climatology for some polynya events. Can you quantify this?

I am concerned about the additional analysis presented in section 4, where you are exploring if AVHRR can be used to identify physical features of the upper ocean or ice that may be associated with polynya opening. In particular, I do not see how the surface brightness temperature can be related directly to ocean properties under the ice in winter. See my simple reasoning for this in the specific comments below. The only finding you have is that there is correspondence with wind curl suggesting T4 decreases under opening favorable winds and T45 is correlated to opening. I think the only thing you can really say out of this is that the brightness temperature changes associated with polynya opening are probably just showing the lead opening that is a precursor to polynya forming. It is very important that you are clear about the limitations of your findings here. Really there is no indication that AVHRR can tell you anything more than that there is a change in ice concentration. My take home message from this is that AVHRR and other satellite remote sensing have to be used together with oceanographic data to identify mechanisms causing the polynya to open. Section 4 needs a tighter focus to ensure readers are not misled that you can say anything about the interior ocean under ice with AVHRR. Your finding appears to be that you might be able to identify opening, which might be a precursor to the polynya open, but can not say anything about mechanisms creating sensible heat polynyas. This is an important distinction if it is the case. What I get out of this section is somewhat of an understanding of why the forecast method might work, and that it might be limited to polynya's that require wind driven opening as a precursor to their formation. Thinking this way refocusses section 4 away from over-reach to try to identify mechanisms of polynya formation to understanding why the forecast might work for this particular polynya. Which does make my question in the paragraph above important, as we could ask does every detected reduction in variability of the various brightness temperature measures (which appears to be related to wind driven opening) result in a polynya?

I highly recommend you read the literature on detecting leads in AVHRR. Here is a starting point:
Lindsay, R. W., & Rothrock, D. A. (1995). Arctic sea ice leads from advanced very high resolution radiometer images. Journal of Geophysical Research: Oceans, 100(C3), 4533-4544.
You should also consider what the brightness temperature is telling you about the skin temperature in the sensor footprint. There is also a ice surface product based on MODIS, that uses similar information.
Lindsay, R. W., & Rothrock, D. A. (1994). Arctic sea ice surface temperature from AVHRR. Journal of Climate, 7(1), 174-183.
Hall, D. K., Key, J. R., Casey, K. A., Riggs, G. A., & Cavalieri, D. J. (2004). Sea ice surface temperature product from MODIS. IEEE transactions on geoscience and remote sensing, 42(5), 1076-1087.

It will greatly help the manuscript if you are clear about what the values you use in your forecast are actually measuring.

I have some specific comments. Line numbers are based on the track change version of the manuscript.

Title: I would suggest you do not need 'upcoming' in the title, you are already pointing to early detection.

line 7: "we find in fact 30 polynyas". I would remove "in fact".

Line 110-115 (end of section 2.1): The float surfacing indicates open water. A lead is likely to have water at the freezing point, and will only have water close to the surface in summer at higher than freezing. Also, I am not understanding the last sentence in this paragraph. The float tries to surface every 10 days. If it surfaces twice this indicates open water on both those dates, but not that there was open water 10 days prior to when it surfaced.

The sea ice concentration time series is not referenced where you first introduce it. Also it is my understanding that the NSIDC and Comiso bootstrap algorithms are different. In the conclusion you appear to refer to the NSIDC data set as the bootstrap data. Please clarify and make sure the appropriate data citation is included in the paper.

line 225, you do not need the second question mark in this paragraph.

line 251-254: "hypothesised that oscillations in the infrared brightness temperature, especially in T4, in the days before the polynya opens might reflect oceanic convective movements. Their argument is that as the warm water is being upwelled, more heat is going through the ice." The base of the ice is always at freezing point. So it is salinity changes that affects the temperature of the base. This will vary very little compared to the top surface. The heat flux through the ice is indeed from warm bottom to cold top surface prior to the onset of surface melt, but the rate will be influenced more by the atmospheric temperature than the ocean temperature. The increase in brightness temperature can be due to enhanced open water, so it would be very hard to dissociate the effect of opening from upwelling, and based on my reasoning above I actually think any upwelling signal would be much smaller than that due to changes in the ice thickness within the satellite footprint. I do think the mooring analysis is interesting, and the correspondence of opening and upwelling (or eddy transfer of warm water) is interesting. In the context of the remote sensing I am not convinced that the variability in T4 is associated with the water property changes. Your discussion is suggesting the same thing, that the signal is likely leads opening. I feel it could be more clearly pointed out that correlation is not causation, and the upwelling/eddy presence is happening at the same time as ice opening. I actually do not understand how you might be able to detect ocean warming under the ice in the skin temperature. This is possible in the open ocean, but if there is ice present the temperature changes at the surface due to heat flux change from the ocean will be much more subtle and maybe lower than the resolution of the sensor. I suggest the title of this section (4.1) is misleading, because you do not show that variation in T4 is directly related to upwelling, or that there is even a statistically significant relationship. In fact you even say "The T4 data are too patchy to robustly determine the increase in brightness temperature that corresponds to upwelling". What your findings do indicate is summed up in lines 312-314 that there is a slight negative correlation between wind favorable to upwelling and surface temperature change (temperature reducing). And you point out that the mooring data shows warm/salty water closer to the surface prior to polynya events, but these might be eddies rather than wind driven upwelling. Essentially the oceanographic data is suggesting that there should not be a direct correspondence between upper ocean warming and ice opening, which is what you show. My feeling is that it is pretty obvious you can not use skin temperature to determine what is happening at depth in the ice covered ocean in winter. It is possible that you need both warm water and wind driven opening to create a polynya and these are not necessarily happening at the same time, it might just be the coincidence of a warm core eddy passing as the ice opens.

Line 278: "On the three moorings, all hydrographic sensors are in the CDW (symbols on Fig. 7 and supplementary Table B2), but the shallowest onesare above the core and can therefore be used to observe potential upwelling." Confusing, you first say there are no sensors outside the CDW then you say there are.

line 299, I think you need a comma after "In the polynya years we detected"

line 313: "lead opening", I would caution you just to say opening of the ice, for the reasons stated above.

line 319" We now investigate whether these leads can be detected" might be "We now investigate whether this ice opening can be detected"
... and in other places

Line 415: "The next step would be to check whether these criteria are still valid for other Antarctic open ocean polynyas,but also coastal polynyas(e.g. the Amundsen Sea Polynya, Randall-Goodwin et al., 2015) or even Arctic polynyas (e.g. North Water and North Green-land polynyas, Preußer et al., 2015; Ludwig et al., 2019)."
"but also" should be "including". I would also remove "even" from "even Arctic polynyas".

Figures: You will want to take a careful review of these. Figure 3 has part of a bounding box visible around the map that overlaps a color bar title. Figure 2 has labels that are more designed for a slide show than a publication, consider if a legend would be appropriate rather than the boxed labels.

Data citations are required for the AVHRR and NSIDC data used in the paper. Please make sure these are provided in the acknowledgements.

Dear Dr. Henzé and co-authors

This paper is now at a point where the limitations of the forecast method are identified and I am more confident that you are presenting the results fairly. Thank you for your patience in the review process, however it was required in this case given the limitations of your findings.

I do have to correct one point in your letter. However I see that in the manuscript you do not have this error, so no edits are required to the manuscript. You state that the base of the ice can be at temperatures higher than the freezing point. I guess it might be possible to super heat ice, but practically less realistic. I see that in the manuscript you correctly state that basal ice melt reduces the ice thickness which changes the heat conduction through the ice and hence the surface temperature.

line 242: "We verify this hypothesis in this section." The results are actually not that conclusive, I would change verify to test.

line 265 and table 1: Please identify where you obtained the atmospheric parameters.

line 272: "we cannot conclude regarding the usefulness of T4 as an indicator that a polynya is being triggered by upwelling." Grammar is a little off here. Try "we cannot conclude if T4 indicates a polynya is triggered by upwelling."

Please do take care to check your manuscript for typos. It has gone through a lot of change over the 4 or 5 iterations during the review process and I did notice a need for grammatical editing every time new text was introduced. I appreciate that you have made the changes I requested to ensure consistency with previous changes and clarity.

Best regards,
Jenny