Dear Dr. Cook et al.,

Thanks for your revised manuscript and response to the two reviewers’ comments. The response is thorough and well-reasoned, strengthening what I already considered to be a compelling initial submission. You have thoughtfully evaluated the continent-wide significance of a well-established ice-shelf process that has received limited attention in recent years – and never at such a large scale – but for which we are increasingly well poised to better understand its role in ice-shelf (in)stability. The manuscript does not position itself as not the final word on the matter and makes clear the roadmap for improved understanding of brine infiltration, which is also appreciated. The Supplementary Material is also well-considered. Below I have several minor comments for your consideration prior to a final decision on your manuscript.

Thanks again for considering The Cryosphere for your work,

Joe MacGregor
NASA GSFC

Throughout the manuscript, the terms “seawater infiltration” and “brine infiltration” appear to be used interchangeably, e.g., lines 5/29 vs. 5/30. I find the former more evocative (and still accurate), but the latter is in more common use in the literature. If they are indeed referring to exactly the same phenomenon, and I believe they are, then only one such term should be used (I personally prefer “seawater”).

Abstract: The abstract could be improved in a couple of places. In particular, the term “variable” is not quite right, as it places the emphasis more on its use in a model than the “presence of liquid water”, which is better characterized as a “physical property” of the ice column than a “variable”. Separately, the last two sentences are written in the passive voice and could be more direct, e.g., “We find that seawater infiltration…” and “Finally, we identify the most likely locations…”

2/23: I understand the addition of additional DEMs as a response to reviewer #1, but some of them are now based on older data, e.g., Bamber et al. (2009) uses a combination of ERS-1 and ICESat data. Further, I was surprised to see that Helm et al. (2014, TC), which uses CryoSat-2 data, was not included, as it is publicly available and uses next-generation radar-altimetry data as compared to Bamber et al. (2009) or Liu et al. (2015). Further, since the Bedmap-2 surface DEM is – as best I understand it – intended to improve upon Bamber et al. (2009), that older DEM appears deprecated relative to Bedmap-2. Please reconsider which surface DEMs are most essential to include in this study. If the Bamber et al. (2009) DEM is kept, then it might be better to refer to it as ERS-1/ICESat to keep it more consistent with the other shorthand names.

3/5: Aren’t “error estimate” and “bias” simply uncertainties? In particular, a bias is not really a bias if its sign is unknown. If a single term can be used for all of the above, please do so.

4/13-14: Admittedly, the nomenclature gets tricky, but aren’t “laterally unconfined ice shelves” simply “ice tongues”? I can’t see from Figure S3 the names of these particular regions in detail, but if the local ice features at the mentioned coasts are already termed ice shelves, then this comment can be ignored.

4/24: This is an interesting insight and useful context, but it’s not made clear what the root cause of this model/observation alignment is. Based on the Supplementary Figures, in particular that for ice thickness, I suspect that an ice-thickness gradient there is the underlying cause for the different in predicted infiltration extent. Is that correct? Either way, further explanation would be appreciated.

5/22: IceBridge

Dear Dr. Cook et al.,

Thanks again very much for your submission and for choosing The Cryosphere. I was intrigued by your study from the start, the reviewers essentially agreed and you revised it well. There appear to be several fruitful avenues of research that your initial mapping, modeling and analysis present. I've examined and am satisfied with your revision, and I'm pleased to accept it for publication.

Cheers,

Joe MacGregor