Advancing Landslide Susceptibility Mapping in the Medea Region Using a Hybrid Metaheuristic ANFIS Approach
, Alexandru-Ionut Petrisor 2,3,4,5,*, Lyes Mohamed Baba Ali 6 and Abdelghani Leghouchi 7
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThis article proposes a novel hybrid Metaheuristic model for the spatial prediction of landslide susceptibility in Medea, combining the Adaptive Neuro-Fuzzy Inference System (ANFIS) with four novel optimization algorithms. It is an interesting paper; however, it needs some corrections and several moments should be clarified, from my point of view.
- Authors claim that traditional machine learning methods have some limitations, such as time-consuming, over-fitting and under-fitting. However, these are inherent issues that are closely related to the data and machine learning model used. As far as I know, it’s not something that can be solved by a hybrid approach. In other words, hybrid methods can improve training efficiency and prediction accuracy, but is not the problem addressed as mentioned in the paper. Moreover, why the four optimization algorithms mentioned above are chosen should be carefully stated in terms of what problem is to be solved, rather than a generalized problem. In this case, authors should clearly point out the motivation of this work. In fact, landslide susceptibility mapping is a traditional task and a large number of modeling approaches have been proposed and used, just a simple traditional method combination work is not enough for academic publication in the journal.
- Figure 1: How to define high susceptibility and low susceptibility? How are these data used in LSM? Does the susceptibility have an effect on the results?
- A detailed description of the landslide information should be provided; the current description is too simple. Moreover, it is better to provide landslide site and image maps.
- The description of the hybrid method is confusing. How can different methods be combined? What are the advantages of combining methods?
- Figures must be improved. For example, landslide points should be placed on the LSM map.
Comments on the Quality of English LanguageMinor editing of English language required
Author Response
Dear Reviewer #1,
Thank you for the deep and constructive revision of our manuscript. Please find attached a detailed response.
The Authors
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsThe manuscript titled "Advancing Landslide Susceptibility Mapping in Medea Region Using Hybrid Metaheuristic ANFIS Approach" submitted by Debiche, Fatiha et al., is well within the scope of the journal but not with the special issue of the Land journal "Remote Sensing Application in Landslide Detection and Assessment" since the article nowhere mentions the application of remote sensing for the detection of landslides except in one part where it is written that the detection of Historical Landslide Locations was done using Google Earth images. The work deals with the development of a novel advanced hybrid machine learning model designed to assess landslide susceptibility effectively. The manuscript is well-written, the results are presented in a average way, but it lacks highlighting its importance and novelty.
What are the limitations of your methodology? What are suggestions for future work?
Specific comments:
A nice study area map (showing the location of the landslide) will make the manuscript more appealing.
Which software is used for photogrammetric or remote sensing processing since the special issue refers to Remote Sensing Application in Landslide Detection and Assessment? May be I have missed it.
Formatting errors are there. Try to keep the representation uniform over the manuscript.
A large number of triggering landslide parameters such as Precipitation, Land cover, Distance to stream and Distance to the road do not have any data on how and in what way the spatial data were collected. Which map, what scale?
Since it was stated that the lithology layer (X1) was established using data from a 1:50,000-scale geologic map, how were the subdivisions defined since it is certainly not found on the geologic map.
The paper stated that "The most appropriate model was integrated into ArcGIS, and optimal input maps were divided into 30x30 m cells using the fishnet tool, which is available in ArcToolbox. The ANFIS parameters of the best model were applied to analyze optimal input layers and provide landslide susceptibility classes for each cell, resulting in the preparation of the landslide susceptibility map for the study area". This needs to be described in detail, as the defined methodology applied in I do not agree with the authors that it is possible to define everything this way.
Author Response
Dear Reviewer #2,
Thank you for the deep and constructive revision of our manuscript. Please find attached a detailed response.
The Authors
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsDear authors,
I have added several notes on an annotated version of your pdf. However, the main problem is that the geological and mathematical parts of the manuscript are not correlated. Finally, the manuscript appears like two separate papers. It is clear in the results section, which is rather superficial. So, my judgment is a major revision.
Comments for author File: Comments.pdf
Quality of written English is in general of good staus but a final edit is needed.
Author Response
Dear Reviewer #3,
Thank you for the deep and constructive revision of our manuscript. Please find attached a detailed response.
The Authors
Author Response File: Author Response.pdf
Reviewer 4 Report
Comments and Suggestions for AuthorsDear Authors,
thank you for your article entitled “Advancing Landslide Susceptibility Mapping in Medea Region Using Hybrid Metaheuristic ANFIS Approach”. It could represent a valuable contribution to the research on the topic of landslide susceptibility. Unfortunately, there are some sections of the manuscript not discussed and/or presented properly and the paper cannot be accepted for the publication in the present form.
A severe issue concerns the figures and related captions, which need to be revised. Some suggestions follow:
- Concerning coordinate labels along the map frame, consider increasing their font size to make them readable at print size;
- Adding a hillshade map in all figures would greatly improve the readability of presented maps, highlighting existing relationships between morphological features and reported data;
- Concerning legends, consider increasing their font size to make them readable at print size; some of them are missing a title and the unit of measure should always be stated; please also check the legend of figure 1, which appears to report inventoried data instead of high/low susceptibility sectors;
- Concerning captions, consider modifying them to better convey information. A reference to data source is often missing and should be implemented.
- Some figures among 2-9 (see, for instance, figures 2, 6, 7) don’t depict the distribution of causative factors, rather than input data used to derive it. Please check and correct the datum accordingly.
Another issue concerns the lack of references, which affects the overall quality of the work. An example is the “2.1 – Case study” paragraph, where information needs to be supported by adequate references. The same comment applies for data sources, which are sometimes omitted (see, for instance, data about geological and lithological features of the study area). Please also add a reference to the work of Jang (line 257) and to Smith's classification (line 478).
Concerning equations, please check their numeration (equation at line 386 and the first one in table 1 are both assigned the number “18”). Furthermore, several symbols are barely readable.
Regarding causative factors, more information and references are required to justify adopted classification (see paragraph 3.1. Database compilation).
Finally, regarding the use of English, there are a few minor errors, typing mistakes, and repetitions. It is recommended a thorough rereading to fix these issues.
Best regards
Comments on the Quality of English LanguageMinor editing of English language required
Author Response
Dear Reviewer #4,
Thank you for the deep and constructive revision of our manuscript. Please find attached a detailed response.
The Authors
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Comments and Suggestions for AuthorsIn general, the authors have attended most of my suggestions. The replies and revision are both a good effort from the authors.
- Figure 1: It is better to provide a base map for the study area, as it is too simple now.
-Section 3.7: It is better to provide results from different methods, rather than just comparing accuracy.
- A high precision value does not mean that the results of a model are reasonable and realistic. It has an important relationship with data selection and model training. Results are very specific to the given datasets, varying with the dataset. Authors should pay attention to this point.
Comments on the Quality of English LanguageMinor editing of English language required
Author Response
Thank you very much for your new comments, and for the time and efforts invested in reviewing our revised work. Please find attached the explanation of addressing them.
Author Response File: Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsEverything is OK now. The authors responded and accepted all my suggestions.
Author Response
Thank you very much for your new comments, and for the time and efforts invested in reviewing our revised work. Please find attached the explanation of addressing them.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsDear authors,
In the previous version of your contribution, I consider that the used mathematic methods are not explained. This problem remains also in the new version.
Comments on the Quality of English LanguageMinor editing in written English is appropriate.
Author Response
Thank you very much for your new comments, and for the time and efforts invested in reviewing our revised work. Please find attached the explanation of addressing them.
Author Response File: Author Response.pdf
Reviewer 4 Report
Comments and Suggestions for AuthorsDear authors,
thanks for the reviewed version of the article “Advancing Landslide Susceptibility Mapping in Medea Region Using Hybrid Metaheuristic ANFIS Approach”. The overall quality of the work has greatly improved, but there are still some minor issues and the paper cannot be accepted for the publication in the present form.
Concerning figure 2, according to the text it should convey information about landslides designated as training and validation data. The legend within the figure should be corrected accordingly.
Concerning references, the works of Jang (1993; line 313) and Smith (1986; line 535) should be added to the reference list.
Concerning equations, please check numeration (for example, equation at line 442 and the first one in table 2 are both assigned the number “18”).
Finally, along the text a few typing mistakes are present (see, for example, the word “Eelevation” in the caption of figure 5).
Best regards
Comments on the Quality of English LanguageMinor editing of English language is required
Author Response
Thank you very much for your new comments, and for the time and efforts invested in reviewing our revised work. Please find attached the explanation of addressing them.
Author Response File: Author Response.pdf
