Dear Sindhu Ramanath Tarekere and co-authors,

Your manuscript has now received two reviews. They are both constructive and supportive of your work. One review is especially thorough and provides many important points for attention. Thank you for your responses to the reviews. I invite you to submit your revised manuscript. I look forward to reading it.

Cheers,
Kristin Poinar

Dear Sindhu Ramanath Tarekere and co-authors,

Your revised manuscript has now been evaluated by two peers and received one review document, which highlights the manuscript's strengths while pointing out many remaining areas for improvement. I continue to believe that the analysis being presented is novel and worthy of presentation, but both reviews (including the extreme brevity of Report #1) indicate to me that the manuscript is difficult to understand. I have some suggestions, below, along with supporting all recommendations made by Referee #2 in their report.

My primary concern is the writing style and organization of this paper. It is a long manuscript with many things to present, which makes organization very important for getting the main ideas across.

Therefore, I suggest the use of topic sentences that begin each paragraph with a clear and important idea. The lack of topic sentences, and instead a great many paragraphs begun with details, make it difficult and meandering read through the findings in this manuscript's current version. Your findings will be better absorbed by your readers, and thus your results will have more impact, by being presented with up-front topic sentences.

I have specific line-by-line suggestions as well.

L9 - The median distance (reported in the abstract and elsewhere) is too precise, need to remove significant figures to something appropriate.

L83 and elsewhere - Suggest removal of all use of "ground truth" since this term is misleading and false; replace with "labels" or "manual delineations", which are accurate terms

L229 - "As observed by Mohajerani et al. (2021), the failure scenarios of the neural network are relatively consistent compared to those of a human."
This is ambiguous: it could mean that the failure scenarios are similar to human failure scenarios, or that the failure scenarios are more consistent than of a human. Please rephrase to better communicate the intended meaning.

Section 4.2 / Results of LOCO - I underscore the reviewer's point that it takes careful reading to tell which non-interferometric predictor variables were helpful versus extraneous. Section 4.2 needs to be written more clearly, with results-stating topic sentences for each paragraph.

Table 3 is also difficult to extract information from, perhaps because all features subsets are written out in long phrases. You might consider a format like the following, with simple abbreviations (RC, etc.) defined in the table caption:
Network 1 | Features present: RC, IC, ... | Features absent: DEM, T, V, ...

Figure 11 - Neither the colorbars nor the caption names what quantity is being shown in these panels. Please fix this omission.


In your revision, please address these points as well as the complete list of helpful suggestions provided by Referee #2. I am hopeful that you will take the time to complete another thoughtful revision, with good attention to organization of each paragraph, in order to make this important manuscript easily readable by the scientific community.

Cheers,
Kristin Poinar

Dear authors,

Thank you for your revised manuscript. Your revisions have improved the presentation and quality of the manuscript. I now find it much easier to follow and understand the procedures, choices, analyses, and findings substantially better. I have only the following comments that would need to be addressed before your manuscript can be published. They are all relatively minor and should be attainable by the authors without much trouble. Thus, I give you an early congratulations.



Abstract: Line 10 word "only" - Please consider removing this word. Having read the rest of the paper, I understand why it’s here — but alone in the abstract, one is left wondering “What else is there available than Re and Im?”.

Line 20-21: Consider defining grounding line (GL) in its second use to avoid conflict with the citation in parentheses.

Line 27: Fix capitalization in citation MacGREGOR

Line 56: What you refer to as "the growing volume of SAR acquisitions suited to detect the grounding line" ought to be cited. I couldn’t find that these sensors are ever discussed explicitly - there is only the implicit assumption that readers are familiar with NISAR, S1c, etc. (I did find a mention in the Conclusion that used "including", implying a non-exhaustive list. There needs to be explicit mention here in the Introduction, too, and the list should be complete.)

Figure 1: What are the two black arrows? Please define in the caption.

Figure 2: Need to define in the caption what the legend abbreviations ERS, SEN, and TSX denote (it is inferrable to most readers, but needs to be explicit for everyone).

Figure 10: Caption has a missing word, "which displayed"

Line 315: possible typo, says "the" but might mean "then"

Figures A1 and A2 - Please add a location, latitude/longitude labels, and/or a scale bar to this figure, as you did for Figure 1.


I am hopeful that you will be able to enact these small improvements. My thanks again for your previous efforts and success at creating a better manuscript of an important study.

Cheers,
Kristin Poinar

Dear authors,

Thank you for your revisions. I am now happy to accept this paper for publication.

I find a small error in Figure A2.a, which is that some of the labels on the color bar have become obscured. This was perfect in a previous version of the figure, so I am leaving this manuscript open for "technical corrections" to give you an opportunity to fix these labels. This is the only change requested.

Congratulations on the acceptance of your manuscript.

Cheers,
Kristin Poinar