This manuscript describes the foraging behaviour of Weddell seals in the Ross Sea, using oceanographic and behavioural data logged using miniaturized loggers attached to the seals. While no ground-breaking result is obtained in this study, the data is of value and the analysis deserves its own publication conditional to major revisions.

I have in particular one major concern about the quality control of data. The authors cite in many places the publication of Yoon and Lee 2021. This publication is written in Korean in what appears to be a corean journal. This does not follow international standards and I am unable to follow what is written there. I believe the authors should treat this publication as a technical report, and assume the reader is unable to utilize it. For this reason, the current work should provide more extensive information about the different corrections that have been estimated and applied (Step 1 to Step 3 in section 2.2). The authors do not seem to be aware of the work of Siegelman et al 2019 either, which provides several recommandations for quality-control including density removal and thermal cell effect corrections.

The analysis of water masses appears very rough. Figure 3 is very hard to read, and it has some strange features such as the MSW being “stuck” at 300m for most of the period. It would be nice to show some profiles and/or sections to get a better sense of what you are trying to show. One wonder also how much of your results depend on the different spatial sampling between the two years. I suggest the authors refine their analysis of hydrographic data to produce more specific results.

Figure 6 is amongst the less informative I have every seen. That the mixed layer is deeper in winter than in summer shouldn’t come as a surprise for anyone even remotely interested in oceanography. This cannot reflect the main novelty of this work. The authors need to clarify what is the main novelty of this study.

Minor comment:

l. 35: this sentence seems to imply you are talking about “climate” change, because that’s what you are describing earlier on, and maybe also because you use the word “adapt”. Yet, the changes you describe in the text are related to diurnal/seasonal variability only. The abstract needs to be clarified.

l. 109-110: this accuracy numbers seem overly optimistic. See Siegelman et al. 2019 for recent estimates of accuracies.

This manuscript describes environmental attributes that Weddell seals appear to favor for foraging within the Ross Sea, using satellite-linked relay loggers and accelerometers to document prey capture attempts. I think the authors have looked at the animals’ foraging ecology from multiple different aspects for a comprehensive view of activities. I hope to see this published after some issues are addressed. There were many areas (identified in below comments) where there was insufficient detail to thoroughly understand methodology and how the authors had performed data processing. Overall, it was also difficult to discern the novelty of this study relative to other work that is referenced throughout the text where Weddell seals were tagged over the winter in the Ross Sea to determine important water masses they associate with. I think one very cool thing about this study that could be emphasized quite a bit more, is that these authors actually have prey capture events to compare between daytime/nighttime, season, and water masses. Prey capture and foraging success if often implied in these foraging ecology studies; however, to my knowledge this has not actually been measured before in Weddell seals overwinter. I also think the results here could be better into broader context with other Weddell seal foraging studies that have been conducted.

Line 84-85: Would rephrase as ‘Weddell seals are the deepest diving phocid with the exception of the elephant seals’ (both southern and northern elephant seals dive deeper than Weddells)

Line 110: Should make clear whether all records were transmitted via ARGOS satellite, or whether some of these instruments were recovered.

Line 110: Since sex is used as a cofactor in model building (as authors state later in the Methods section) the sample sizes of male to female should be put somewhere in this paragraph.

Line 116-118: It is unclear whether this is a SMRU accelerometer. Or, whether this was a separate accelerometer that was attached alongside the SMRU tag. In either case, it needs to also be made clear what make and model the accelerometer was.

Line 124: Which logger divided the dive into 3 segments? Are the authors back to talking about the SMRU tag? The rest of this paragraph is confusing and should be clarified. Did the SMRU tags really divide each dive into 3 segments, as these instruments typically provide 4 inflection points within each dive. It sounds like that is the case here and the authors then did further processing by interpolating X number of midpoints and then the authors divided the dive into 3 segments: descent, bottom, and ascent. The ‘dive threshold’ also needs to be defined: is it that only dives >X m were retained in the dataset?

Line 195: How were dives that exceeded the IBCSO bathymetry (seals diving deeper than the ‘bottom’) treated?

Lines 199-200: change ‘<’ to ‘>’ for dives durations > 5760 and dive depths > 906 being excluded.

Line 200: Would clarify that this is ‘vertical travel speeds exceeding 5.1’ (unless authors have also put a filter on horizontal distance traveled)

Line 219: Were these models run with REML or ML?

Line 219: There is no statement of model validation (checking for homoscedasticity etc).

Line 240: This whole paragraph incorporates a lot of discussion points into the Results section. These sentences especially that reference other works would be more appropriate in the Discussion section

Line 257: Percentage of dives made in MSW ?

Figure 2. Could the points be color coded by water mass? It is difficult to interpret.

Figure 3. I like seeing the depths of the seals and the depths of the water masses together; and the x-axis being categorical (month) for the boxplots makes sense. It is unclear how to make sense of the depth of the water masses with this x-axis. Is ‘Mar’ equivalent to March 1 for the continuous variable plotted for water mass? This should be made more clear.

Figure 6 seems very general without a lot of information given that the AASW deepening is well known. I wonder if this might be better portrayed if the dive record of one seal is overlain on top of the schematic to show dive depth profiles (&with prey capture attempts marked) across a few days in March relative to the AASW; and dive depth of that same seal for a few days in July relative to AASW to show a representative example of the seal avoiding AASW if it is less preferred. Otherwise, this figure could probably be omitted.

Tables: It seems odd that some of the variables stated to have a large impact on behaviors had very high p-values in the models (for example Table 1. Sex has a p value of 0.23 and year had a P value of 0.893 – did it really improve model fit enough to stay in the best fit model?). This is generally considered to be one of the drawbacks of stepwise approaches to model selection; or it can result from differences in ML versus REML methods.

Line 277-278: Instead of ‘variations in’ would clarify which direction these shifts in behavior went in daytime versus night (greater proportion benthic dives, depths etc).

I also thought the Results said there was no difference in number of dives (i.e., foraging frequencies)?

In general, there were two things missing (for me) from the Discussion section. First, I think the most novel aspect of this paper is – that while there have been other Weddell seal tagging studies within the Ross Sea also looking at water masses that the animals associate with --- to my knowledge, this has never been paired with the addition of the accelerometers for prey capture events. This validates a lot of the ecological theories that have always been applied given the assumption that the animals are foraging a lot more in certain areas. It also highlights that even with a similar number of dives during the nighttime, animals are capturing less prey even though the animals (& prey) are likely shallower in the water column and it should be potentially less costly for the animals. That’s pretty interesting! I am also aware of studies using accelerometry to document prey capture events in Weddell seals in the summer, but I am not aware of any such studies in the winter. I think some summer studies could be referenced for comparison between the breeding season, summer, and winter. I think more emphasis could be put on how the prey capture attempts validates important aspects of daily and seasonal foraging ecology.

The other is I think that this would benefit for some discussion comparing the findings from this study with others that have tagged Weddell seals in the Ross Sea (were findings the same? – implying consistency across longer timespans? Or were some aspects different?). And also beyond the Ross Sea to put into context.