Arndt et al provide a very valuable and interesting data set and this definitely merit publication in TC. Overall the the paper is fairly well written, however, while attempting for brevity (which is a plus), there are however lack of clarity that would need to be improved by providing some more detail throughout the manuscript, and some critical self-evaluation of the fundamental assumption that are made. I think the required revisions are largely technical and minor (although numerous), and thus constitute minor revisions before acceptance.

While I give only a few more generic comments here, the authors are provided detailed comments and suggestions in the annotated manuscript attached to this review. I would expect a point by point response also to those comments with the revised manuscript. 

Generic comments:

Use of sea ice and snow ice in the text - I would suggest to follow the following practice: Sea ice as noun = sea ice. Sea ice as adjective (modifier of noun) = sea-ice. Two examples: Climate warming has affected the extent of sea ice. Climate warming has affected the sea-ice extent. Using "snow-ice" throughout, would IMHO improve the readability.

Introductuction/Discussion/Summary: I suggest that the authors give a bit more quantitative results from previous work on the importance and contributions of snow-ice and superimposed ice to Antarctic sea ice mass balance (some of this literature is noted in the annotated ms.), this would also place the current work better in context of existing work by Lange, Jeffries, Maksym, Ackley etc. It could be also helpful to note on how WS compares to other sectors of the Southern Ocean to place the results in broader context.

Throughout (abstract, discussion, summary), the assumption on the fact that begative freeboard will always cause flooding when using the simple sea-ice model is fundamental to all the results presented, and the caveats of this assumption should be clearly noted and discussed. There are even studies that looked into this e.g. Ackley et al. 2020, Wever et al., 2021, and I would feel this needs to be better assessed.

It would also be good with an assessment what the pros and cons are of different type of buoys used for such studies, if flooding could be directly detected with e.g. a thermistor-string type buoy (like Ackley et al., 2020).

I also struggled what the purpose was to use two models, if I understand correctly the SNOWPACK model would provide both snow-ice, superimposed ice and snow depth in one go, what is the purpose of using the very simple thermodynamic model? There is neither much (enough) discussion on how the two models compare in e.g. producing flooding (snow-ice) and why there are differences. Maybe this is simply a case to elaborate a bit better on the justification on how the two models complement each other.

I would also have liked to see the ice thickness results, now these are not shown at all, although one key factor triggering flooding is that the ice is thin enough. As such it would be good to also show evolution of ice thickness alongside the snow-related parameters. Here  the assumption of constant vs. variable ocean heat flux could highlighted in the Summary. I also note that Ackley et al (2017) summarize ocean heat flux observations, and for WS they appear much larger than what is used in this study (at least for parts of the year). 

Data availability: Reference to ice core data are given in several places, but I do not see this data is listed with the data sets?

This study aimed to try to understand some of the complex snow processes that occur on Antarctic sea ice such as snow-ice formation and super imposed ice in order to improve snow models and also satellite measurements of sea ice volume. They used multiple Snow Buoys and insitu core data from multiple ship campaigns over the years in the Weddell Sea. The snow buoys tell the amount of snow accumulations and melting that occurs throughout the year, and the cores give the amount of snow ice and super imposed ice. A 1-d sea ice model and a more sophisticated snow model (SNOWPAK) are used to model the amount of snow, and super imposed ice and ice formation that is taking place at the buoy locations as they drift throughout th4e Weddell Sea. They found that the snow models were not the best at producing the amount of snow and ice compared to observations, which is likely caused by missing processes that are taking place within the ice. They also determined that more snow/ice formation is occurring in the western weddell sea and at lower latitudes.

While this paper was well written and easy to follow, I wish there were more details on the individual models and processes there in. What ways could the models improve to improve the snow-ice and super imposed ice estimates? How do the models produce the snow-ice?

More details like this could improve the paper. What would happen if you change the densities and salinities of the snow and ice in the models, will the snow depth be more similar to the buoy snow depths? More details like this would help.

Otherwise, I really enjoyed this study, and I think It makes a great addition to our understanding of the Antarctic snowpack.