Karst Collapse Risk Zonation and Evaluation in Wuhan, China Based on Analytic Hierarchy Process, Logistic Regression, and InSAR Angular Distortion Approaches

Round 1

Reviewer 1 Report

1. How to choose the two models(analytical hierarchy process (AHP) and logistic regression (LR) models)and what is the scientific basis? Could you give a fuller explanation in the manuscript?
2. What is the main theoretical contribution of this paper? Can it be clearer in abstract part?
3. The manuscript shows both models performed well in identifying high-risk zones with only a 3% discrepancy in area. Could you discuss the reason for the difference, model factor or data factor?
4. Whether the theory in this article is applicable to the research of other regions in the world, what factors need to be specially considered, and whether it can be discussed.
5. It is recommended that the authors search and reasonably supplement the latest literature in recent years, especially in 2021.

 

Author Response

We especially would like to thank you for giving us many substantive suggestions to improve the manuscript. We have carefully addressed all comments item by item. The changes are marked in yellow in the revised manuscript and are detailed in our reply letter. We hope that our revised version satisfies all of your concerns.

Author Response File: Author Response.docx

Reviewer 2 Report

The main goal of the research was to find the areas hazarded by karst collapse in Wuhan in China. The research methodology was based on: analytic hierarchy process (AHP), logistic regression (LR), and InSAR monitoring. Analysis done enable the authors to develop two models for karst collapse risk zonation. First model was based on AHP method and the second one was based on LR method. The results obtained were compared with weighted 38 angular distortion (WAD) method derived by incorporating InSAR deformation. As a final result of the investigation was a map with estimated probability of karst collapse.

I think that presented research are interesting case study of InSAR application to support modelling the zones prone to karts collapse with use AHP and LR methods. Proposed method may be of interest to a wide audience. However the novelty of applied methods is limited. Analytic hierarchy process and logistic regression has been used in order to asses caving and sinkholes modeling. The same as InSAR to observe sinkholes or crack occurrence.

Nevertheless the article is well written, graphic is are well done and legible. To summarize, I think the presented research and results obtained are convincing and worth publication.

Author Response

We especially would like to thank you for approving our work and giving us substantive comments. We have addressed the general comments carefully. We hope that our response can satisfy all of your concerns. Thank you.

Author Response File: Author Response.docx

Reviewer 3 Report

Manuscript ID: remotesensing-1470710

 

Title: Karst collapse risk zonation and evaluation in Wuhan, China based on analytic hierarchy process, logistic regression, and InSAR angular distortion approaches

 

Authors: Jiyuan Hu, Mahdi Motagh, Fen Qin, Jianchen Zhang, Wenhao Wu, and Yakun Han

 

In this manuscript authors present a detailed assessment of risk zones related to karst collapse phenomena in the area of the mega-city of Wuhan in China by the qualitative analytical hierarchy process (AHP) and the semi-quantitative logistic regression (LR) models.  Five types of criterion layers including karst geology conditions, overburden conditions, hydrogeological conditions, karst surface subsidence conditions, and anthropological activities were considered by authors into both models. Moreover, both models performed well in identifying high-risk areas with a highly correlated spatial distribution. Comparison and statistics of sinkhole presence and absence are performed by authors by locating successfully areas with higher karst collapse risk situation in the district of Wuhan mega-city.

 

In this form the manuscript doesn’t fit the high-quality standards of MDPI-Remote Sensing, I therefore suggest some minor changes:

As general remarks:

I suggest authors to summarize the abstract moving some sentence in other sections. In my opinion is too long.

 

In particular:

• At line 113 – Figure 1 - the writes of upper right rectangular box of Figure 1 (a) are difficult to be read (too little).
• From line 152 to 154 – Authors use the units m³/d as physical dimension of parameter C32 – it is not clear what does the physical unit d represents.
• From line 159 to line 165 – Authors use the A-DinSAR SBAS-Stamps method to detect the subsidence of the studied area (vertical and horizontal components), but as widely discussed in literature the A-DinSAR derived displacement rates should be compared with GNSS data and eventually also spirit levelling data to be validated/confirmed, cause the A-DinSAR is not geo-referenced and sensitive as well as the GNSS. Concerning the vertical velocities, the spatial geodesy derived ones should be connected/compared with high-precision levelling data. So at least a comparison with position time series of GNSS permanent or semi-permanent station present in the study area if available should be considered when estimating horizontal and vertical signals due to subsidence. I suggest Authors to add some explaining sentences and corresponding references.
• A line 195 – Table 1 – At the right side of the table the word environment is truncated.
• At line 265 – I think that 90*90 m² is more correct that 90*90 m.
• At line 305 – 90m*90m can also be indicated as 90*90 m² the discrepancy with line 265 is evident.
• At line 318 – I think that a reference of the SPSS® - software is necessary following the style of the journal Remote Sensing.
• At lines 344-345 - Authors should explain better why for the two sub-criteria described the condition (sig <0.05) is not satisfied. This sentence is not clear.
• At line 355 – Is not clear what does the data in round brackets (1:50,000) represent (chart scale?) some more word is necessary.
• At line 552 – I think that the word modes should be replaced with the word models, or the sentence need to be reorganised.

Author Response

We especially would like to thank you for giving us many substantive suggestions to improve the manuscript. We have carefully addressed all comments item by item. The changes are marked in yellow in the revised manuscript and are detailed in our reply letter. We hope that our revised version satisfies all of your concerns. Thank you so much.

Author Response File: Author Response.docx