|This is my second round of reviews for this manuscript. In my first review of the original manuscript, my main concerns regarded (i) the reasoning being the choice of the ST criteria, (ii) the lack of direct evaluation of the methods, and (iii) improving the manuscript's readability. |
The authors have answered all my initial comments, and I find that the revised manuscript addresses well these three main concerns.
The introduction and method description were greatly improved. The result section is now clearer and presented in a more linear way. I appreciate the addition of Figures 3, 5, and 7.
I still have a few remarks regarding editing and text clarifications, and one main comment which I trust the authors will be able to address. Overall, I believe the manuscript is close to being fit for publication, and would recommend acceptance after some minor revisions.
Since the choice of the DR criterion used for computing the ST criterion is, to some extent, somewhat arbitrary, it would be nice to add 1 or 2 sentences discussing this choice & how it should be optimized in the last section of the manuscript. You already touched on this topic by indicating that the ST criterion could be slightly changed to provide a better match with the DR criterion. This raises two questions:
- Should one aim for a match between both methods?
- If yes, are there particular reasons or limitations to why the adjustment of the ST criterion, following your own recommendations, could not be included in this manuscript as well? Adding an attempted “optimal ST criterion” for future users could strengthen your manuscript even more.
L. 6 and 338: Could you clarify what is meant by “robust” here? Most likely is done in the core text, but if mentioned in the abstract it should be clear what the criteria are for readers.
L. 28-30: This is confusing: you first refer to the advection of shelf water onto more saline waters as a halocline formation mechanism. Yet in the following sentences, you state that this mechanism is supported by 2 other studies, but then cite the formation of halocline via ice melting on top of warm AW. Although there are similarities in these mechanisms, they are not exactly the same. In the first, the input of freshwater comes from the fresher shelf waters, themselves fed by continental runoff and/or ice melt. In the latter, only direct ice melt on top of saltier waters is at play. Please clarify this statement.
L. 40: This is again confusing. I understand what you mean, but calling the PWW both a part of the halocline and a halostad layer can appear contradictory. You could consider saying that the halocline in the Canada Basin is constituted of PHW, in which we distinguish a halocline layer formed by the PSW, which overlays a cold halostad formed by PWW, and below the PHW is found the strongly stratified LHW of Atlantic-origin. Adding a schematic figure of what these layers look like in a T-S profile could also help. You could point to the local stability maxima as well.
L. 50: typo “ist”
L. 59: This sentence should rather be merged with line 52, here it is a little off-topic.
L. 67: “among others by others” typo
L. 86-87: This sentence should be moved to the end of the previous paragraph (l. 59), as it results from emphasizing the importance of the halocline in the Arctic system.
L.88: The manuscript outline is long and too detailed. Consider shortening this part, to no more than 1 sentence per section (1-2 lines) without detailing each sub-section. You might also not need to acknowledge your reviewers’ suggestions.
L. 148: “as demonstrated below” What does this refer to? This section? A figure?
Fig. 3: Nice figure.
- tick labels are occasionally overlaying each other, consider fixing that
- panel label and plot lines are occasionally overlaying each other, consider fixing that
- if possible, would be great to have a profile of density as well.
L. 263-310. This section is interesting but quite dense. Consider breaking it down into several paragraphs, e.g., by adding a line break l. 271 (before “In addition”) and l. 283 (when transitioning to the Canada Basin).
L. 275-277: Somewhat convoluted phrasing. In general, it is easier when sentences are straight to the point. In this case, you could just say that “the ST method also detects more frequent cases of halocline BD below 120 m compared to the DR method.”
L. 282: Any comments on this? You don’t make any use of Fig. 7c otherwise.
L. 312-316: This is merely a suggestion, and you may choose or not to follow it: since you have already explained your concepts in section 2, you can consider removing the reminder of these concepts at the beginning of your subsections in section 3. Here, for example, lines 312-318 mostly repeat information provided in section 2.
Since the paper is not overly long, you can however decide to maintain these reminders if you prefer to do so.
L. 335: Redefine all acronyms in the summary section
L. 374: This is a bit colloquial (and rather pessimistic wording). Consider saying rather that e.g., “mismatch rates seem to be notably varying across ITPs tested. This suggests that further testing and refining are needed in order to account for regional and temporal variability of CHS characteristics”
L. 381: I would suggest rephrasing your very last sentence so that it highlights the potential of AI methods without discarding too harshly traditional methods (including your newly proposed one).