Coastal Wetland Responses to Sea Level Rise: The Losers and Winners Based on Hydro-Geomorphological Settings

Round 1

Reviewer 1 Report

I have carefully read the paper and overall, I think it is an interesting work, worth publishing. The study is very timely as coastal wetlands are threatened by sea level rise and anthropogenic pressures, in general.

Although I am not a native English speaker, I believe the English needs polishing throughout the text.

Overall, I think the data are well presented and explained, the results are supported by the data and the discussion is well established.

My suggestion is that the paper should be accepted with revisions.

You may find in attachment the annotated pdf with additional comments for the authors.

Comments for author File: Comments.pdf

Author Response

Thank you for your comments. We have made all the changes suggested. 

Reviewer 2 Report

This paper uses a machine learning approach to suggest how estuarine wetlands will transition under various sea-level rise scenarios. The choice of variables is poorly justified - there is the assumption that these are important but no detail is provided. They are at least measurable. There is very little process based explanation of the findings of the analysis. The authors seem to assume that the results are correct without really needing to account for the types of differences they find. Why are these results reasonable? There is some discussion of the upland transition near the end but it seems to be an afterthought. The results are complex and poorly presented with extensive repetition of numbers in the text. The presentation/language is very hard to follow at times - some diagrams and tables might make the results easier to digest. 

I was left wondering what the authors think would actually happen in the estuary. One key factor of time was not accounted for. How rapidly might these changes occur - many of the habitats were tree dominated - how do they adjust - clearly not instantly. I would like to see a more thorough examination of the limitations of the study e.g., the inability to distinguish brackish trees from salt marshes seems to be a major limitation but it is glossed over. Surely the brackish zone might b subject to squeeze? Is the salinity gradient not important?

Some comments are provided on the attached file. Note the pdf was converted to Word for comment and then back to pdf so there may be some format issues.

Comments for author File: Comments.pdf

Author Response

We appreciate your comments and suggestions. Please see the attched file which documented our point-to-point responses.

Lin 212: I agree that this is much better than a simple bathtub model but these are quite dramatic changes in water exchange - I think some discussion of how such SLR might adjust the tidal dynamics and thus the tidal planes would be of value (even if you don't make an adjustment acknowledge the assumptions or justify why it is small).

Thank you for your comment. We added “The approach does not, however, include any future modification to the tidal dynamics as a result of morphological adjustment to SLR, thus it remains a first-order approximation to future water levels inside the estuaries [57]” to the end of the paragraph.

Line 235: Something missing in this sentence - As xyz then what?

Thanks for your comment. The sentence was changed as “As the land cover classes are highly imbalanced, and most machine learning models tend to be more efficient and accurate in predicting the majority class than the minority class as they aim to minimize the overall error rate, [62], sampling techniques are often used to tackle the problem of extremely imbalanced data”.

Line 251: Three?

Yes. Changed to “Three SLR scenarios…”.

Line: 312: The insert photographs require annotation for the features mentioned to be clear to the reader.

Thanks for your comment. Features are annotated in the inserts.

Line 321: I may have lost track but have you actually states which one is which yet - I assume that scenario 1 is the lowest but clarification would be good. Especially as the finding in the next sentence seems a little but counterintuitive for the highest amount of rise.

Thanks for your suggestion. The sentence was revised as “… increased 1,256 ha (10.5%), 10,227 ha (89.9%), and 23,851 ha (209.8%) under low, moderate and high SLR scenarios”.

Line 322: This sentence is really not clear - comparable to what?

We revised the sentence as “In comparison to the areal gain in Water, the loss in terrestrial vegetation area was comparable under low scenario…”.

Line 328: Do you mean this is inconsequential? This terminology is odd.

Yes. Changed as suggested.

Line 360: Presumably this is due to topography/slope - if so I think you should say that here and previously when this first comes up

Thanks for your suggestions. We added “likely due to steeper slope …” in the sentence.

Line 371: It is confusing to state that something uneven is consistent - please rephrase.

Thanks for your comment. We revised the sentence as “… and the uneven areal changes in different wetland types was consistent between scenarios …”.

Line 430: These section are extremely hard to follow and really grasp as they just describe lots of data. Find a better way to display this information in a figure.

Thanks for your suggestion. We revised this section so it’s clearer.

Line 442: This needs to be introduced when you are discussing the data- provide more details on the actual topography. Is it really 'hydro' geomorphology or just slope? If there is a hydro component you should explain what you mean.

Thanks for your suggestion. We added more details.

Line 449: So how do you know it is hydro-geomorphology? You need to be much clearer about what you do and don't assume in your analysis - and provide some discussion about how these additional factors could influence or interact with the factors you included

Yes, as we used only hydro-geomorphological variables in the model. 

Line 449: Unclear what this means

Revised to “tide entrance conditions”.

Line 452: Hardly a complex model. I think you should look further into models that simulate wetland change

Revised as “Nevertheless, the results of our machine learning approach are broadly consistent with more data-demanding numerical models [73], such as the 1D Marsh Equilibrium Model (MEM), 2D SLAMM (the Sea Level Affecting Marshes Model), and more complex 3D Delft3D hydrodynamic model [82].

Line 468: Are these factors surrogates in any way for the other processes - e.g., distance and depth could reflect potential sediment supply?

We have not considered sedimentation and other factors when constructing Distance to water Edge and Tidal depth for the future scenarios.

Line 472: That actually includes brackish marsh - maybe it was the brackish components of the category that were not persistent? How would you know?

We changed to saltmarsh/swamp.

Line 499: This needs to be acknowledged much earlier

Thanks for your suggestion. We added this in the Introduction.

Author Response File: Author Response.docx