Certainty Factor Analyses and Spatiotemporal Characteristics of Landslide Evolution: Case Studies in the Chishan River Watershed in Taiwan

Round 1

Reviewer 1 Report

This paper uses a different approach to landslide investigation. Although the structure of the paper is well observed, it needs the following corrections to improve:

1. It is better to mention the criteria affecting landslides well in the abstract, materials and methods and results. NDVI can be considered as an effective indicator along with other criteria.

2. Numerical results should be stated in the abstract.

3. This research does not have a specific evaluation method, it is better to use an evaluation method.

4. In the introduction, it is better to state the previous research on the research topic. Also, research innovation should be stated at the end of the introduction.

5. In the materials section, the effective criteria map, as well as the landslide distribution map, should be given.

6. In fact, the name of the method used is a certainty factor and not a sensitivity analysis. It is better to use another method of sensitivity analysis.

7. It is better to use hotspot analysis to study landslides in different years and also Moran index can be used to study the type of landslide distribution.

8. Improve the quality of charts.

9. The evaluation method should be expressed on the results.

10. Have a deeper discussion of the results by citing sources.

Author Response

Dear ISPRS International Journal of Geo-Information

My sincere appreciation for the valuable comments from the reviewers. The responses for all comments are listed in the following sentences.

Chunhung WU

***

Response for the comments of reviewer 1

Comments: English language and style are fine/minor spell check required

Response: The revised manuscript had been carefully checked.

Comments: This paper uses a different approach to landslide investigation. Although the structure of the paper is well observed, it needs the following corrections to improve:

1. It is better to mention the criteria affecting landslides well in the abstract, materials and methods and results. NDVI can be considered as an effective indicator along with other criteria.

Response: Appreciation for the valuable comments. The materials and processes for producing the annual landslide inventories during 1999-2017 were added in the 3.1 Materials.

Comments: 2. Numerical results should be stated in the abstract.

Response: Appreciation for the valuable comments. We had revised the abstract in the manuscript.

Comments:3. This research does not have a specific evaluation method; it is better to use an evaluation method.

Response: Appreciation for the valuable comments. I must apologize for that we don’t really understand the comments very well. The landslide evaluation methods were frequently used in building the landslide susceptibility model or evaluation of landslide hazard (If I was wrong, please let me know). The study focused on analyzing the relationship between geomorphologic factors and landslide evolution by using certainty factor method, exploring the spatiotemporal distribution and patterns of landslide evolution hotspot by using emerging hotspot analysis, and estimating the landslide clustering strength by using Z-values. The method of certainty factor, emerging hotspot analysis, and Z-value had been explained in detail in 3. Materials and Method. My apology for the comments again.

Comments:4. In the introduction, it is better to state the previous research on the research topic. Also, research innovation should be stated at the end of the introduction.

Response: Appreciation for the valuable comments. We had added the sentence about the previous research using the spatiotemporal analysis on landslide distribution, and also added the sentence about the innovation in this research in 1. Introduction.

Comments: 5. In the materials section, the effective criteria map, as well as the landslide distribution map, should be given.

Response: Appreciation for the valuable comments. We had added the landslide distribution and landslide susceptibility maps in 3.1 Materials.

Comments:6. In fact, the name of the method used is a certainty factor and not a sensitivity analysis. It is better to use another method of sensitivity analysis.

Response: Appreciation for the valuable comments. The sensitivity analysis in the title of manuscript had been changed to Certainty factor analyses. The sensitivity analysis had been changed “certainty factor analyses” in the revised manuscript.

Comments:7. It is better to use hotspot analysis to study landslides in different years and also Moran index can be used to study the type of landslide distribution.

Response: Appreciation for the valuable comments. The local outlier analysis by using the local Anselin Moran's I index had been added in 3.3. Local outlier analysis and 4.3 Result of local outlier analysis in the revised manuscript.

Comments:8. Improve the quality of charts.

Response: Appreciation for the valuable comments. We had improved the quality of figures used in the revised manuscript. We enlarged the size of figures and the DPI of figures was 600 DPI.

Comments:9. The evaluation method should be expressed on the results.

Response: Appreciation for the valuable comments. We had re-write the first sentence in each paragraph in 4. Result to explain methodology and materials used in the paragraph.

Comments:10. Have a deeper discussion of the results by citing sources.

Response: Appreciation for the valuable comments. The target of Discussion in the revised manuscript is to explain how to use the Z-value as an index to assess the landslide recovery in different scale (watershed or sub-watershed) or different size (small, middle, and large landslide) in different time period. Using the Z-value to quantify and assess the landslide clustering strength of long-term landslide evolution was a very important finding in this study, and no other method which can provide the quantified data of evolution, clustering strength, and recovery of landslide had been suggested or published. We had added the evaluating method of landslide recovery time from other studies in the 5. Discussion.

Author Response File: Author Response.pdf

Reviewer 2 Report

 

This manuscript analyzed the landslides caused by the Taiwan Chichi earthquake in 1999, the typhoon in 2009, and the subsequent landslide evolution in a several watershed in a region of Taiwan. It is helpful to understand the temporal and spatial distribution of landslides in this area. The specific issues are as follows: (1) The resolution of satellite images used to make landslide inventory maps at different times varies greatly, some 10 meters, some 2 meters, how this affects the results should be analyzed in the discussion section. (2) In the discussion section, the author should compare the data, methods, results, etc. with the previous studies to analyze the advantages and disadvantages of this work, and need to provide some prospects. Currently, the authors have only compared the result from this study to the result from the 2005 Kashmir earthquake. In fact, there are more and better studies in other areas, such as the 2008 Wenchuan earthquake and researchers' studies on the evolution of the landslides in the 1999 Chichi earthquake, etc. That is, the current Discussion section is too weak, and some sections should move to the Results and Analysis section. A real discussion and more content should be added to this section.

 

Author Response

Dear ISPRS International Journal of Geo-Information

My sincere appreciation for the valuable comments from the reviewers. The responses for all comments are listed in the following sentences.

Chunhung WU

***

Response for the comments of reviewer 2

Comments: Moderate English changes required

Response: The revised manuscript had been carefully checked.

Comments: This manuscript analyzed the landslides caused by the Taiwan Chichi earthquake in 1999, the typhoon in 2009, and the subsequent landslide evolution in a several watershed in a region of Taiwan. It is helpful to understand the temporal and spatial distribution of landslides in this area. The specific issues are as follows: (1) The resolution of satellite images used to make landslide inventory maps at different times varies greatly, some 10 meters, some 2 meters, how this affects the results should be analyzed in the discussion section.

Response : Appreciation for the valuable comments. The revision had been added in the 5. Discussion. 

Comments: (2) In the discussion section, the author should compare the data, methods, results, etc. with the previous studies to analyze the advantages and disadvantages of this work, and need to provide some prospects. Currently, the authors have only compared the result from this study to the result from the 2005 Kashmir earthquake. In fact, there are more and better studies in other areas, such as the 2008 Wenchuan earthquake and researchers' studies on the evolution of the landslides in the 1999 Chichi earthquake, etc. That is, the current Discussion section is too weak, and some sections should move to the Results and Analysis section. A real discussion and more content should be added to this section.

Response: Appreciation for the valuable comments. We had revised the Discussion paragraph in the manuscript. Using Z-value to assess the temporal change of landslide clustering strength was a key finding in this research, and also a quantified and convincible way to assess the landslide recovery time. The analysis of landslide clustering by using Z-value can be proceeded in different scale (watershed or sub-watershed), different size (small, middle, and large landslide) in different time period. We had moved the sentences about the temporal change of Z-value in small, middle, and large landslide to 4.5 Clustering strength of small, middle, and lrage landslide. We focused on how to use the Z-value to assess the oscillation period in the Chishan river watershed and in the upstream watershed, and the influences of spatial resolution of the landslide-identified images to the average Z-value during 1999-2017 in the Discussion paragraph.

Author Response File: Author Response.pdf

Reviewer 3 Report

The paper is really interesting and well presented. Figures and tables area clear and support the reader in understand the main concepts.  

The references used in the introduction section can be increased in number due to the high scientific production about topic as the one reported on lines 39-42. 

In the section 2. Research Area could be useful to increase the description including the characterization fo the area also in terms of meteoclimatica data, if available. In this way section 4.1 can be better understand and the results obtained compared with the characterization. 

Caption of Table 1 add a point at the end.

Author Response

Dear ISPRS International Journal of Geo-Information

My sincere appreciation for the valuable comments from the reviewers. The responses for all comments are listed in the following sentences.

Chunhung WU

***

Comments: English language and style are fine/minor spell check required

Response: The revised manuscript had been carefully checked.

Comments: The paper is really interesting and well presented. Figures and tables area clear and support the reader in understand the main concepts.  

The references used in the introduction section can be increased in number due to the high scientific production about topic as the one reported on lines 39-42.

Response: Appreciation for the valuable comments. We had added more reference in the revised manuscript.

Comments: In the section 2. Research Area could be useful to increase the description including the characterization for the area also in terms of meteoclimatica data, if available.

Response: We had added the sentence about the study area, Chishan river watershed, in 2. Research Area.

Comments: In this way section 4.1 can be better understand and the results obtained compared with the characterization. 

Response: We had added some sentences to compare the rainfall characteristic duing 1950-1998 and 1999-2017 and describe the rainfall characteristic during 1999-2017. The rainfall characteristic in the Chishan river watershed during 1999-2017 was plenty of rainfall with high rainfall intensity and concentrated in the rainy seasons.

Caption of Table 1 add a point at the end.

Response: Appreciation for the valuable comments. The revision had been revised in the Caption of Table 1 in the manuscript.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors have responded to all comments. Therefore, my decision is to accept the article.