Assessing Landslide Susceptibility in the Northern Stretch of Arun Tectonic Window, Nepal
Round 1
Reviewer 1 Report
Dear Authors,
in my opinion, the article is rather relevant and interesting. However I suggest creating the Discussion section in the paper, to underline the novelty of your findings compared to previous researches.
Good luck!
Author Response
We are very grateful for the constructive comments and suggestions from the three reviewers. The manuscript has been revised accordingly based on these suggestions.
The major corrections and revisions have been marked in red in the revised manuscript.
Responses to Reviewer #1
Dear Authors,
in my opinion, the article is rather relevant and interesting. However, I suggest creating the Discussion section in the paper, to underline the novelty of your findings compared to previous researches.
Good luck!
Reply: We thank the reviewer for the suggestion. In the revised manuscript a “Discussion” section has been added, including the success rating curve and discussion about various relevant previous research.
Reviewer 2 Report
The paper proposes two well-known methods for assessing landslides susceptibility in an area of Nepal. The approach is well known, but the results could be interesting for local authorities, as few studies are present. I suggest some very important revisions:
In the following my comments:
1) Remove the term vulnerability, as it applies to buildings, you are performing susceptibility analysis.
2) It is not clear why these events are threatning, are there roads, or trails or towns? Please add some comments and insert them in the maps.
3) I assume that you are studying susceptibility for the triggering, hence you have to analyze the causing factors only for the crown areas of the landslides.
4) What kind of landslides are you referring? I suspect that you have rockfalls evolving in debris flows. Please refer to Hungr et al. classification.
5) Usually landslide datasets are divided in training and testing, you can do that and finally you could use a ROC for evaluating the adopted approaches.
6) In the results you could show the frequency classes for each parameter with histograms rather than descibing them.
7) The methods comparison could also be shown in terms of map difference of the relative classes, discussing the differences and similarities.
8)Fig. 1 change the order of c, b, a and in a insert an hillshade to provide some 3d shape. Insert some toponyms and critical elements (roads, towns and so on...) The same applies to figs. 4-10.
9) The paper has very few reference, considering that many authors worked on the topic, you could add these and some of the reference within:
Dai et al. 2005 https://doi.org/10.1016/S0013-7952(01)00093-X
Reichenbach et al. 2018 https://doi.org/10.1016/j.earscirev.2018.03.001
Forte et al. (2019): https://doi.org/10.3390/geosciences9100412
Yan et al. 2019 https://doi.org/10.1016/j.geomorph.2018.10.024
Author Response
We are very grateful for the constructive comments and suggestions from the three reviewers. The manuscript has been revised accordingly based on these suggestions.
The major corrections and revisions have been marked in red in the revised manuscript.
Responses to Reviewer #2
The paper proposes two well-known methods for assessing landslides susceptibility in an area of Nepal. The approach is well known, but the results could be interesting for local authorities, as few studies are present. I suggest some very important revisions:
In the following my comments:
1) Remove the term vulnerability, as it applies to buildings, you are performing susceptibility analysis.
Reply: We have removed the term vulnerability (by replacing it with susceptibility) in the revised manuscript.
2) It is not clear why these events are threatening, are there roads, or trails or towns? Please add some comments and insert them in the maps.
Reply: Indeed, some of these events are threats to various infrastructure, mainly villages and roads. Their locations have been added in the map (Fig. 3~10) and a comment added in the text.
3) I assume that you are studying susceptibility for the triggering, hence you have to analyze the causing factors only for the crown areas of the landslides.
Reply: In this study, the entire area is examined. Landslide source area was only considered during the preparation of landslide polygons. The debris run out and deposition area has already been excluded in the landslide polygon used for the analysis and generation of the susceptibility map. The factors which are not at the landslide crown also play a role in the occurrence of landslide, for example, streams are not at the landslide but may alter the pore water pressure of the slope triggering landslide.
4) What kind of landslides are you referring? I suspect that you have rockfalls evolving in debris flows. Please refer to Hungr et al. classification.
Reply: Indeed, there are various kinds of landslides in the study area, rockfalls, debris flows included. In the field investigation, obviously it is not possible to backtrack the exact characteristics of each landslide, any evident significant traces of sliding surfaces and debris materials are considered (without distinguishing the precise type), this information has been clarified in the manuscript.
5) Usually, landslide datasets are divided in training and testing, you can do that and finally you could use a ROC for evaluating the adopted approaches.
Reply: Indeed, we agree that there are multiple methods possible including the receiver operating characteristic (ROC) curve; we have mentioned some studies in the “Discussion” Section. In the present study the sample size is relatively small with 74 landslides, using some of the landslide as training data and using the remaining as testing data might not produce reliable results. Of course, this is still something that can explored in the future studies. We have offered some brief discussion in the “Discussion” Section.
6) In the results you could show the frequency classes for each parameter with histograms rather than describing them.
Reply: Since the information about the frequency class for each parameter is already included in each factor map (Figs. 4 to 10), we feel that plotting the histograms might be redundant, this is why at times the values are briefly mentioned in the text.
7) The methods comparison could also be shown in terms of map difference of the relative classes, discussing the differences and similarities.
Reply: The differences and similarities in the susceptibility maps are discussed on Line 270 to 279 in Section 4 when the results are presented. Since the map of relative class is same for both methods, as same factors and same classes are considered, the difference is the only in the landslide susceptibility index as the two methods adopt different approaches to calculate the landslide susceptibility index, so the susceptibility maps are different.
8) Fig. 1 change the order of c, b, a and in a insert an hillshade to provide some 3d shape. Insert some toponyms and critical elements (roads, towns and so on...) The same applies to figs. 4-10.
Reply: Reordering of images have been done along with addition of hillshade effect in Fig. 1. The hillshade effect is also in landslide inventory map (Fig. 3). The trails and villages have been added in all relevant figures suggested (see Comment 2).
9) The paper has very few reference, considering that many authors worked on the topic, you could add these and some of the reference within:
Dai et al. 2005 https://doi.org/10.1016/S0013-7952(01)00093-X
Reichenbach et al. 2018 https://doi.org/10.1016/j.earscirev.2018.03.001
Forte et al. (2019): https://doi.org/10.3390/geosciences9100412
Yan et al. 2019 https://doi.org/10.1016/j.geomorph.2018.10.024
Reply: We thank the reviewer to draw our attention to these important studies. The above-mentioned references have been added (Ref [9], [11], [12], [13]).
Reviewer 3 Report
This paper is of great significance to carry out susceptibility assessment research on landslides in eastern Nepal. However, the following issues need to be paid attention to:
1、 When evaluating the susceptibility of landslide, the paper should first explain how to select the evaluation unit of landslide and its rationality. According to the content of the paper, the author choose pixels as the evaluation unit. But if so, how to obtain the "slope shape" index in section 4.1.3? Because a pixel corresponds to a slope, it cannot represent the shape of convex and concave, which are contradictory.
- What GIS tools used should be introduced in the paper?
- The value range of C in equation 3 and the spatial correlation between this value and landslide should be introduced.
- It should be explained why stream proximity has an important impact on slope stability.
- Whether the paper can consider not using the factor of stream power index, because it has little impact on landslide.
- The study area in this paper is in the area with frequent tectonic activities. Therefore, in the geological factors, the geological structure (such as fault) should be considered as the influence factor of landslide.
- Some landslide density in the conclusion are not completely consistent with table 2 and table 3 in the text. Please check carefully.
Author Response
We are very grateful for the constructive comments and suggestions from the three reviewers. The manuscript has been revised accordingly based on these suggestions.
The major corrections and revisions have been marked in red in the revised manuscript.
Responses to Reviewer #3
This paper is of great significance to carry out susceptibility assessment research on landslides in eastern Nepal. However, the following issues need to be paid attention to:
- When evaluating the susceptibility of landslide, the paper should first explain how to select the evaluation unit of landslide and its rationality. According to the content of the paper, the author choose pixels as the evaluation unit. But if so, how to obtain the "slope shape" index in section 4.1.3? Because a pixel corresponds to a slope, it cannot represent the shape of convex and concave, which are contradictory.
Reply: Digital Elevation Model (DEM) is composed of many pixels. Each pixel has its own x coordinate y-coordinate and z coordinate (elevation). It can be considered that each pixel resembles a (map) point (i.e., a certain small area). The area covered by a landslide is introduced into the google map and later GIS based on the field survey of its location and area, a landslide is represented by many pixels. The slope shape or the curvature cannot be determined with a single pixel, rather, it is calculated from adjacent pixels; for example, the slope can be calculated between two pixels, based on their x, y, and z-coordinates. The curvature is calculated the by taking the derivatives of DEM. The processing procedure in the present study is very common in geological investigations (as in many cited papers); it is well known that conventional GIS tools have sophisticated computation capacity to deal with such geometric calculations,
- What GIS tools used should be introduced in the paper?
Reply: QGIS and ArcGIS 10.7 (ESRI, USA) are used in the present study, which has been mentioned in the revised manuscript.
- The value range of C in equation 3 and the spatial correlation between this value and landslide should be introduced.
Reply: Theoretically the value of C has no confined range, i.e., it can range from positive infinity to negative infinity, such extreme cases represent practically meaningless scenarios such as all the areas are landslides, or no landslide at all. Therefore, it is not mentioned in the text (just as in many other published studies). However, it is mentioned in the revised manuscript (Line 151~152) that if the value of C is positive then there is positive association of that factor class with landslide occurrence, and if the value is negative then there is no association of the factor class with the landslide occurrence.
- It should be explained why stream proximity has an important impact on slope stability.
Reply: The proximity of the slopes to drainage is one of the important factors influencing stability. The degree of saturation of slopes play important role in slope stability (Yalcin and Bulut 2007; Yalcin 2008). Streams can adversely affect slope stability by toe-incision or by saturating the lower part of hillslope material due to increase in water level (Gokceoglu and Aksoy 1996; Kanungo et al. 2020). It has been elaborated in the revised manuscript along with the supporting references.
- Whether the paper can consider not using the factor of stream power index, because it has little impact on landslide.
Reply: The gradient of streams in the study area is very high and the rainfall intensity is extreme; in such scenario, bank incision and material transport are very common, which can influence the occurrence of the landslides; therefore, the factor stream power index (SPI) is commonly used in studies carried out in the areas where the stream gradient is very high, for example, SPI is considered in the studies carried out by Poudyal et al. 2010; Kayastha et al. 2012; Park et al. 2013; Chen et al. 2017 which are cited in the manuscript, therefore in the authors’ opinion, it is worth considering this factor for this case study.
- The study area in this paper is in the area with frequent tectonic activities. Therefore, in the geological factors, the geological structure (such as fault) should be considered as the influence factor of landslide.
Reply: The major geological structure in the study area is the Main Central Thrust (MCT). The precise location of the MCT in this area is still debated or controversial (Macfarlane et al. 1992; Gupta et al. 2010; Mottram et al. 2011; Adhikari et al. 2021; Li et al. 2022), therefore it may not be possible to accurately consider this factor at this stage of the study. In addition, the MCT is inactive in many parts of Himalayas including in this area (Schelling 1992; McDougall et al. 1993; Rajendran et al. 2015; Ansari 2021; Ansari 2022); the slope failures due to geological structures are generally not observed so far in the study area; hence the factors related to geological structure like distance from fault is not considered in this study. But it might be worthy of further investigations when more information becomes available.
- Some landslide density in the conclusion are not completely consistent with table 2 and table 3 in the text. Please check carefully.
Reply: We have carefully checked the values to ensure the consistency in the revised manuscript. It is noted that in the conclusion the result of both methods are mentioned, therefore it is presented as a range, i.e., 0.06%~0.10% in the very low susceptibility zone, and 0.71% ~0.89% in the high susceptibility zone.
Additional references cited in this reply
Adhikari, D., Silwal, C. B., & Paudel, L. P. (2021). Review of the Geology of the Arun-Tamor Region, Eastern Nepal: Present Understandings, Controversies and Research Gaps. Journal of Institute of Science and Technology, 26(2), 79-97.
Ansari K. (2022). The main Himalayan thrust and geometrical parameters. (Accessed on 25 February 2022) https://researchoutreach.org/community-content/main-himalayan-thrust-geometrical-parameters/
Ansari, K. (2021). Review of the geometric model parameters of the main Himalayan thrust. Structural Geology and Tectonics Field Guidebook—Volume 1, 305-323.
Li, R., Ao, S., Xiao, W., Schulmann, K., Mao, Q., Song, D., Tan, Z., Wang, H., Bhandari, S. (2022). Tectonic juxtaposition of two independent Paleoproterozoic arcs by Cenozoic duplexing in the Arun Tectonic Window of the eastern Nepalese Himalaya. Frontiers in Earth Science, 573.
Macfarlane, A. M., Hodges, K. V., & Lux, D. (1992). A structural analysis of the Main Central thrust zone, Langtang National Park, central Nepal Himalaya. Geological Society of America Bulletin, 104(11), 1389-1402.
McDougall, J. W., Hussain, A., & Yeats, R. S. (1993). The Main Boundary Thrust and propagation of deformation into the foreland fold-and-thrust belt in northern Pakistan near the Indus River. Geological Society, London, Special Publications, 74(1), 581-588.
Mottram, C. M., Harris, N. B., Parrish, R. R., Argles, T. W., Warren, C. J., & Gupta, S. (2011). Shedding light on the Main Central Thrust Controversy, Sikkim Himalaya. Journal of Himalayan Earth Sceinces, 44(1), 60.
Rajendran, C. P., John, B., & Rajendran, K. (2015). Medieval pulse of great earthquakes in the central Himalaya: Viewing past activities on the frontal thrust. Journal of Geophysical Research: Solid Earth, 120(3), 1623-1641.
Schelling, D. (1992). The tectonostratigraphy and structure of the eastern Nepal Himalaya. Tectonics, 11(5), 925-943.
Round 2
Reviewer 2 Report
Although the authors did not implement all my suggestions, they satisfactorly discussed them and modified the manuscript accordingly.
For me the paper can be accepted.
