Please see the attached PDF

Review of Verro et al, “How well do the regional atmospheric and oceanic models describe the Antarctic sea ice albedo?”

Given the ongoing losses in Arctic sea ice and the signs we are now seeing of similar processes getting underway across Antarctic sea ice as well, the manuscript here is topically well placed to explore the model descriptions of Antarctic sea ice albedo and certainly suitable for TC.

In general, the study is well executed, with a clear goal, and clearly written. I particularly appreciated the lucid descriptions of the various model albedo schemes which highlight the differences well for prospective users, as well as the multi-stage progression across spatial scales during the analysis.

I have relatively few comments to make here, although I consider the first point below as something the authors should take a close look at. Once a revision is made to address these comments, I trust that the manuscript will be suitable for publication in TC.

Major comments

While I can accept the decision to defer cloud impacts on sea ice albedo to future studies, it is on this point that the manuscript should be clearer. First, the typical difference between the clear and cloudy sky albedo should be noted, using e.g. Key (2001) as a reference. Then, if all data sources in the study are indeed clear-sky only, it should be made very clear how the models are coerced to only provide albedos consistent with clear-sky conditions.

Additionally, For the satellite data, the S2 and LS9 data are evidently clear-sky, though I would appreciate clearer details on the clear-sky atmospheric correction necessary to provide the surface reflectances (yes, actually these data are nadir-view directional snow reflectances – they can well be a good estimate for the view-integrated albedo, but you should be clear on the distinction). And for CLARA-A3, I think that the data there are available for various illumination conditions – which did you use here?

4.4 – it’s not clear how many drone flights contributed to the histograms in Fig 6? If from multiple days, what was the day-to-day albedo variability in the drone data? Were the flights made always over the same survey grid? Was the weather clear or cloudy – that would also change the observed snow albedo.

4.5. – you can calculate the mean SZA over the study areas as you know the S2 and LS9 overpass times; a first-order estimate for the albedo effect assuming non-melting snow would then easily be available from a lightweight albedo parameterization such as that of Gardner and Sharp (2010), please provide the assessment in the text as a yardstick for the reader.

Figs 10-12: It is my understanding that CLARA-A3 (not 3A as in some legends) albedos are either 5-day or monthly means, yet here the text refers to CLARA-A3 products from a specific day. Did you recompute your own daily version based on provided raw data?

324: 20 cm of snow is considered thin? From the optical (albedo) viewpoint, 10 cm is typically enough to effectively make the snowpack optically semi-infinite. The text does refer to this effect (for both ice and snow), but it would be nice to have quantified estimates here for typical depths required – and for the authors to consider if any of the results are affected. Also, while melt ponds are rarely encountered over Antarctic sea ice, it would be nice from the completeness viewpoint to recall that melt pond albedo is also not uniform, but depends on the depth of the pond and the properties of the underlying ice. Several appropriate references exist highlighting this effect.

Minor comments:

131: A +/-1% measurement uncertainty sounds very high for field conditions. Is this a manufacturer estimate?

fig 3: legend gets lost in the subplot c, please consider moving it outside of plot area.