Dr. Palm. The second round of review comments have been returned to me, with both reviewers thanking you for the efforts made to improve the paper after the first round of reviews. However, both reviewers still have significant concerns that need to be satisfactorily addressed prior to publication. Although this is the second round of reviews, I am willing to allow another round of revisions if you wish. I would suggest that you address all of the reviewers ongoing concerns, including the following.

Reviewer #1:
The main concern is that you need to substantiate the claim that MERRA-2 is too moist/cold. One possible approach is to conduct a more thorough comparison of MERRA-2 RH with observations, and make sure consistent RH parameters are used in the different data sets. In addition, please consider the following as well.

Figure 2: what is the source of the AWS data? What is their name and elevation? Why are those ‘randomly’?

Figure 3 a and 3b require significant improvement to address the following concerns. They basically show a big cloud of data and do not support any of the results. The authors should average the data in time and show e.g. monthly averages in comparison to stations. Also, what is the source of the station data? Are they quality controlled? Are they corrected for RH at low temperatures? Drawing conclusions from a comparison at PE (which is not a really typical ‘Antarctic’ station, since it is sheltered by topography, and blowing snow does not occur frequently) only does not provide evidence of any Antarctic-wide signal. If the authors want to justify this claim, it will require the inclusion of more weather stations.
Also, even more importantly, does MERRA-2 show RH with respect to ice or to water? I suspect it’s with respect to ice, and I know ERA-Interim shows RH with respect to water (even at freezing temperatures). This will largely explain the difference between the two reanalyses. Please check thoroughly and make sure they are consistent. If they are inconsistent, I would argue that the claim that MERRA-2 is too moist is no longer justifiable.

Table I: please include uncertainties for each year.

Line 499: How is 5% justified? The authors should carefully compare MERRA-2 to many AWS stations in Antarctica, and derive an uncertainty/error from that – see above.

Reviewer #2:
The major concern is that procedures to estimates the sublimation rates involve considerable errors and its evaluation is not sufficient to date.

Please address the following concerns:
Once the sublimation is generated in the specific grid, the properties of air, such as the temperature and the humidity, change and then flow into the next downstream grid. Obviously, amount of sublimation decreases along the flow except for the case of a drastic temperature rise. Please confirm whether this process is taken into account in the procedure described in the paper. If it is not, it seems likely that you overestimate sublimation. Could this be the cause of the difference with previous research.

There is a concern that the procedure to evaluate the total error is no relevant. Probably more careful analysis is needed. At a minimum, estimated errors of +/-50 % should be declared in the abstract as well.

Thank you ,

Philip Marsh

Line 212: there is a need to better explain why the same algorithm for cirrus clouds is applicable for the larger blowing snow particles.

Dear Dr. Palm - thank you for your detailed responses to both reviewers comments during both rounds of review. Although the reviewers still raise significant questions concerning some of your methods and results, I believe that the paper is ready to be published.

Thanks again,

Philip Marsh