Analysis of Geological Hazard Susceptibility of Landslides in Muli County Based on Random Forest Algorithm

Round 1

Reviewer 1 Report

Needsview after major revision re

Comments for author File: Comments.pdf

Author Response

Thank you for your comments concerning our manuscript entitled " Analysis of geological hazard susceptibility of landslides in Muli County based on random forest algorithm " (Manuscript Number: sustainability-2159922). Those comments are all valuable and very helpful for revising and improving our paper, as well as the important guiding significance to our researches. We have studied comments carefully and have made correction which we hope meet with approval.

Author Response File: Author Response.pdf

Reviewer 2 Report

The manuscript entitled “Analysis of geological hazard susceptibility of landslides in Muli County based on random forest algorithm” that was submitted to “Sustainability” used a machine learning (ML) classifier to predict the landslide susceptibility assessment (LSA). The manuscript contains an interesting topic, but it requires several modifications. In this regard, the following comments are requested to be addressed by the authors:

Comment 1: The English of the paper is readable; however, I would suggest the authors to have it checked, preferably by a native English-speaking person, to avoid any mistakes.

Comment 2: The concluding remarks of the abstract are not well-written. It’s merely the repetition of the objectives and title of the manuscript. Please add the quantitative findings, methods limitations and model justification into the abstract.

Comment 3: The necessity & novelty of the manuscript should be presented and stressed in the “Introduction” section.

Comment 4: Provide a literature on the methods developed/applied on ML application in LSA in the “Introduction.” Using a table to demonstrate the advantage-disadvantage of these methods can be helpful. Towards the end, mention the superiority & repeat the novelty of your work.

Comment 5: A relevant source of subjectivity and uncertainty is introduced when splitting the input parameters into an arbitrary number of classes with random break values. These choices affect the results. Would you please describe your solution?

Comment 6: Please provide the table of hyper-parameters values of all algorithms.

Comment 7: Please add a subsection clearly articulating the main limitations, wider applicability of your methods, and findings in the “Discussion” section.

Comment 8: The authors should deepen the discussion.

Comment 9: The conclusion section needs uncertainty statements and policy implications of this study.

Comment 10: I would suggest that the authors review and include the following studies to improve the manuscript.

1. Azarafza, M., Azarafza, M., Akgün, H., Atkinson, P. M., & Derakhshani, R. (2021). Deep learning-based landslide susceptibility mapping. Scientific reports, 11(1), 1-16.

2. Nikoobakht, S., Azarafza, M., Akgün, H., & Derakhshani, R. (2022). Landslide susceptibility assessment by using convolutional neural network. Applied Sciences, 12(12), 5992.

 

3. Nanehkaran, Y.A., Pusatli, T., Chengyong, J., Chen, J., Cemiloglu, A., Azarafza, M., & Derakhshani, R. (2022). Application of Machine Learning Techniques for the Estimation of the Safety Factor in Slope Stability Analysis. Water, 14(22), 3743.

Author Response

Thank you for your comments concerning our manuscript entitled " Analysis of geological hazard susceptibility of landslides in Muli County based on random forest algorithm " (Manuscript Number: sustainability-2159922). Those comments are all valuable and very helpful for revising and improving our paper, as well as the important guiding significance to our researches. We have studied comments carefully and have made correction which we hope meet with approval.

Author Response File: Author Response.pdf

Reviewer 3 Report

The manuscript, entitled "Analysis of geological hazard susceptibility of landslides in Muli County based on random forest algorithm" (by Wu et al.) analyzed landslide data and 10 landslide-influencing factor data under the implementation of the random forest model to assess the landslide occurrence probability (susceptibility) in a county of China. Furthermore, model performance indicators were used to reflect the reliability of the applied model’s output.

General remarks

The manuscript deals with a topic of interest which makes it suitable for publication. However, I find severe issues that must be solved in order the manuscript to be considered for publication.

Major comments (among others)

Regarding the landslides, as it is shown in the sub-figures of Figure 3, not only the location but also the extent of the landslide events recorded by field surveys were recognized. So, it would be more representative and accurate for the authors to handle them considering their extents (representing as polygon features) and not just their locations (representing as point features).

Expert for the above landslide representation, relevant statements in the manuscript about the phenomenon is totally incorrect! First of all, which is the difference between "15 landslides" and "313 landslides"? Secondly (and most important) according to the widely accepted classification of landslide types proposed by Varnes (1978), debris flows constitute one of landslide types. However, according to statement, debris flows are not considered landslide type and are separated. What this separation was based on?

Concerning the content of the manuscript, the only element that is presented in a satisfactory degree and in detail within the manuscript is the data used for the landslide-influencing factors. For instance, among others, the applied model of random forest is described just by a single paragraph.

Furthermore, "Discussion" just focuses on highlighting, from one side, the applied random forest model and its advantages, and from the other side the agreement of the factor weight findings with those from other previous relevant studies. For instance, an interpretation of the results, or the limitations/assumptions of the study, are not at all provided.

The “Conclusions” section also just duplicates the numerical results of the analysis.

 

Except for the above, a few minor suggestions are provided so that the authors take them under consideration:

Minor suggestions

Extensive English language checking and editing are required for the entire manuscript.

Line 28: Double mention of "accuracy" with different percentage, 99.43% and 99.3%. Which is the correct? Revise it.

Lines 37-44: Too long sentence! Revise it or split it into shorter ones.

Lines 49-50: Not easily understood sentence. Revise it.

Line 51: "more influenced by" is more correct than "influenced by more". Revise it.

Lines 53-55: Revise this sentence.

Lines 58-59: "become more and more widely" what? applied, used? Revise it.

Line 63: Instead of "Algorithms model", "Models" is more appropriate. Replace it.

Line 69: "type", not "types".

Lines 70-71: Revise this sentence.

Line 83: Add "subjectively" after "determining".

Lines 91-96: Split the sentence into two: "...deep cut mountain landform. The territory...".

Line 113: This sub-heading is similar to the previous 2.1. Revise it.

Line 118: Instead of just "slope", "slope angle" is more representative. Consider it for the entire body text.

Lines 122-124: Reference for this statement?

Lines 203-204: Revise this part of sentence.

Line 228: If “ACC” is abbreviation, as first mention, explain it.

Lines 230-231: Duplicate sentences. Delete one of them.

Line 232: If "ACC" refers to accuracy, replace the second "accuracy" with "precision" (mentioned as the name of indicator).

Line 235: Similarly to the previous comment.

Lines 239-245: This paragraph fits better to the "Results" section.

Lines 248-258: This paragraph fits better to the "Discussion" section (specifically in 4.2 sub-section).

Line 266: The term "natural interruption method" refers to natural breaks (jenks) classification method provided in GIS software packages? Clarify that.

Lines 304-310: Too long sentence! Revise it.

Lines 332-349: As it was commented above, these two paragraphs just present again the results. They cannot be related to the "Conclusions" section.

Author Response

Thank you for your comments concerning our manuscript entitled " Analysis of geological hazard susceptibility of landslides in Muli County based on random forest algorithm " (Manuscript Number: sustainability-2159922). Those comments are all valuable and very helpful for revising and improving our paper, as well as the important guiding significance to our researches. We have studied comments carefully and have made correction which we hope meet with approval.

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

Dear Authors,

Considering your revision reply and recognizing your overall effort to improve the submitted manuscript, I eventually suggest its acceptance for publication.