Thank you for the revised manuscript submission entitled "The impact of snow depth, snow density and ice density on sea ice thickness retrieval from satellite radar altimetry: results from the ESA-CCI Sea Ice ECV Project Round Robin Exercise”. I appreciate the author's efforts to address the reviewers comments, though in some instances I feel that more could have been done. Below I list several suggested improvements to the paper. One major comment is that you really need to include a discussion of freeboard uncertainties, a concern raised by all 3 reviewers. These uncertainties will impact your validation of total freeboard and ice draft. I think one thing you could do is clearly state in the introduction you are not focused on validating freeboard, but are taking existing published freeboard data (that are subject to a large range of uncertainties) and looking at the impact of snow depth, snow density and ice density on ice thickness and ice draft. To me the manuscript still requires a lot of streamlining to make the results/discussion easier to follow and make your key points known.

Also, I have to agree somewhat with reviewer 3 that it's questionable how novel this study is as many of the results have been discussed in other studies. As a response you mention that you show the limitations of using a RA freeboard product with a grid resolution of 50 to 100 km for sea ice thickness retrieval.
I don't see that you have clearly shown this, so perhaps you need to focus a discussion on that conclusion. The same with your other novel points you mention in the response. Make these more clear within the manuscript, perhaps by restructuring the manuscript to clearly highlight what is novel. I also question the need for the snow depth intercomparison as other studies have already looked at W99 vs OIB vs AMSR-2. Is it not possible to refer to those studies and discuss impacts, rather than discussing it again here? What new information are you presenting regarding the accuracy of the W99 data set?

I do find the abstract still is a bit too long and should be shortened to summarize the main points
Introduction:
page 2, seems you should make it clear that ERS1/2 RA and Envisat RA-2 do not in fact provide hemispheric coverage of ice thickness, large parts of the Arctic are not sampled.
page 3, line 19, this paragraph seems to pertain to all the different techniques for measuring ice thickness, but obviously laser altimetry and ULS measurements are not based on assumptions of radar wave penetration, perhaps rewrite that part for clarity.
page 4, line 1-9, I think in here you should somehow make it clear you are not focused on a consistency check of freeboard, I often find the discussion becomes too unclear as to what you are doing.
page 4, line 9, I don't follow your reasoning for not using the most recent OIB data.
page 4, line 14, remove e.g.
Section 2.1
page 5, first paragraph, I think in here you should mention the resolution of RA and RA-2.
page 5, line 19, so about 20-150 measurements are used to create the average?
page 5 lines 23-30, I do find statements such as "the majority of RA-2 sea ice freeboard values are in a reasonable range". Please be more specific (same with the snow depth statements). What is defined as reasonable?
page 5, lines 26-29, reviewers have raised concerns about the use of the W99 data in regions that were extrapolated. I feel that this needs to be more explicitly addressed in the study. Both reviewers 1 and 2 have suggested to remove that discussion. If you are going to keep it in, you need to state here that while you are showing the results, they cannot be trusted.
page 7, lines 21-29, this seems out of place as you do not discuss freeboard uncertainties in the RA/RA-2 data. Since you are not validating sea ice freeboard, I think you need to reframe this discussion slightly to make that point clear.
page 10, line 5-6, since the paper is already rather long, you could remove the first two sentences.
Figure 4, I don't understand your figures, are you saying that Figure 4a and Figure 4c are AMSR-E snow depths only? Then why say the are for CryoVEx in the caption? Also, I'm surprised you start the discussion with Fram Strait since W99 doesn't have data there. Seems better to start with Figure 5 discussion (moved to Figure 4) and then go back to Fram Strait, together with your responses (revisions) to the reviewers concerns about the Fram Strait results.

page 10, line 15-19, I wouldn't say OIB underestimates W99, rather the other way around (i.e. W99 overestimates snow depth compared to OIB).
page 10, lines 22-28, this belongs with a discussion of the figure about the Fram Strait results not here.

page 11, line 26-28, Indeed a climatology will not capture any interannual variability, thus one doesn't expect W99 to match with yearly variations in snow depth seen in OIB.
Table 4, specify over what months these summaries are for. Same with Table 5
page 13, line 15, wouldn't you expect a seasonal change in density? So couldn't that be part of the explanation?
page 13, lines 18-19, I don't understand this reasoning here. Are you saying that for the FYI parts when you use the AMSR-2 snow depths, you should also use FYI density for retrieval of RA-2 derived draft?
page 13, line 24, should specify here too what ice density data are used.
page 13, line 31, should define what is meant by the Arctic Ocean here (i.e. what areas are sampled - this should have gone into the description of the coverage of RA).
page 14, lines 13-16, awkward sentence, rewrite. Also, a limit to this comparison is the W99 snow depth data is an extrapolation.
The Discussion section seems to need a bit of reworking. The first part of is goes back to the ULS comparison, which should have been included when presenting those results. The section should really start with the discussion starting on page 15, line 7. Also, please try to better separate out the draft vs OIB comparisons in the discussion.
A paper that should be referenced within this work is the recent one by Zygmuntowska et al., 2014 (TC).
page 19, recommendation 1) I'm surprised at the recommendation to use the W99. As the authors are aware, the Arctic Ocean is now covered by more FYI than MYI, but also that atmospheric conditions are changing and the W99 does not capture interannual variability in precipitation. Thus, while it may be the only thing we have right now, I would think a recommendation towards developing an interannual varying snow depth product would be of more value.
page 19, recommendation 2) I don't feel that you have quantified that the impact of sea ice density is as large as the impact of snow depth. Other studies have suggested this, and I agree, but I think you should try to quantify that in a table or figure. What ice type product do you recommend using?

I am pleased to accept tc-2014-23 "The impact of snow depth, snow density and ice density on sea ice thickness retrieval from satellite radar altimetry: Results from the ESA-CCI Sea Ice ECV Project Round Robin Exercise" for publication in The Cryosphere.