The Relevance of Geotechnical-Unit Characterization for Landslide-Susceptibility Mapping with SHALSTAB

Round 1

Reviewer 1 Report

Dear Authors,

The manuscript is focused on the application of a well known physically-based model (SHALSTAB) for landslide susceptibility evaluation on a study area prone to landslides (Jaguar creek basin, Brazil).

One of the main disadvantages of the early version of SHALSTAB was to integrate the geotechnical parameters as a constant value for the whole area under study. The authors acknowledged this limitation, as well as in many other studies where this subject was tackled, however, they additionally brought an intensive research – given the number of model’ runs and scenarios performed – to provide some findings regarding the response of the models’ outputs when the geotechnical parameters are discretized by each lithology/soil type.

The paper fits the scope of the journal, and is well written and well structured. Moreover, it provides new findings that can be incorporated in further research on the topic. In my opinion, the manuscript can be accepted as it is.

Kind regards

Author Response

A: We appreciate your kind and insightful comments and suggestion for acceptance of our paper.

Reviewer 2 Report

The authors addressed an important point in landslide susceptibility mapping: the systematic lack of spatialized geotechnical information. It is indeed true that geotechnical parameters are often overlooked or homogenized despite their fundamental role in controlling slope stability. I must praise the authors for putting forward the need to consider geotechnical parameters. Perhaps, to improve the impact of the work, the authors could better discuss the minimum amount of data necessary to make a difference in model results. For example, how many data points per square km could be considered sufficient depending on the spatial heterogeneity of the study area? How can borehole data be averaged over depth and how will the uncertainties in parameter determination affect the reliability of the result? 

Author Response

A: We appreciate your kind and insightful comments and suggestions for improving our paper. Here we see three questions for this improvement: (1) the minimum number of the data points per square km for sufficient representation of the spatial heterogeneity of the study area; (2) the representability of borehole data over the soil depth; and (3) effects of uncertainties in parameter determination about the results’ reliability. We added this information to the text. Furthermore, we added a paragraph in the discussion session where we address the question of the uncertainties.

In order to follow your suggestion (Question – 1), we added a new paragraph in the discussion session. We can say that there is a possibility where the minimum sampling number for determining the soil cohesion could be different from that for soil friction angle. The choice of the minimum number of sampling points requires careful consideration. Even though we understand this importance, we did not concern it, because our objective was just to show the effect of geotechnical unit consideration on mapping performance with SHALSTAB. Please see the modified paragraph:

The choice of the minimum number of sampling points requires careful consideration. This is one of the most critical issues for field workers who want to be time-efficient while at the same time achieving precise results. This number may depend on (i) soil types within an adopted soil classification, (ii) soil parameters for analysis; and (iii) other physical conditions such as soil covers (vegetation) and soil depth. In any case, adopting geotechnical units in future assessments will surely enable the generation of more reliable landslide susceptibility maps.

 

The Question – 2 is about borehole use. In Michel (2015), the borehole test was carried out in order to characterize the soil mechanical properties of the regions of the possible slipping plane. Our paper as well as Michel (2015) did not intend to characterize the whole profile of a soil layer with the parameters’ data obtained only at one depth. Our interest was to have soil mechanical characteristics at the possible slipping depth. In order to clarify our interest, we put more explanation within the first paragraph of the item 2.2. Please see the modified paragraph.

The input parameters ϕ, c, and z were obtained from Michel [41], which carried out soil samplings and Borehole tests at 20 points distributed across the 5 geotechnical units (Figure 2). Note that the sampling depth for the borehole tests in Michel [41] was near the possible slipping plane. This was done because our interest was to have soil mechanical characteristics at the possible slipping depth, and we did not intend to characterize the entire soil layer profile.

The uncertainty effects on result reliability (Question – 3) must be one of the most important themes in the geosciences like hydrology, geomorphology and hydrogeomorphology. This theme should be investigated more in detail. As mentioned above, our objective was to show the effect of geotechnical unit consideration on mapping performance with SHALSTAB. Therefore, we added one paragraph about the uncertainty issue at the end of the discussion session. Please see the added paragraph.

Although the present study showed the effect of geotechnical unit consideration on mapping performance with SHALSTAB, it did not investigate how the uncertainties in parameter determination affect the reliability of the results. This theme should be investigated more in detail in future studies. How uncertainty affects results reliability are among the most central challenges in geosciences in general, including hydrology, geomorphology and hydrogeomorphology [47,48]. The uncertainty depends on (i) the nature of each parameter (model input data), (ii) the methodological procedures used to collect data in the field and to analyze in the laboratory, (iii), the technical level of the involved researchers, (iv) the adopted models, among other variables. Therefore, when feasible, uncertainty evaluations should be carried out together with sensitive analysis like in Michel et al. [26]. The results presented here advance this direction, showing how results vary when different assumptions are considered in each scenario.

Reviewer 3 Report

This paper is about to provide a new look for the application of SHALSTAB model. In view of the fact that most of the original applications of the SHALSTAB model used homogeneous soils, the authors do some efforts to incorporate geotechnical properties as a variable in contrast to the original applications. In theory, this effort will improve the accuracy of landslide susceptibility mapping based on the SHALSTAB model. However, there still exist some critical issues should be addressed before I can recommend this article. In view of the above considerations, I recommended this manuscript should be Seriously Revised at present situation. General and specific comments are given below.

GENERAL COMEMNTS

1. More details of the algorithm of SHALSTAB model should be provided. After over twenty years development, many researchers have issued various SHALSTAB models with different performance, or even led to mixed results. So, the version of SHALSTAB models is the key. Although the author mentioned in the Section 2.3, “The present work adopted the modified algorithm proposed by Michel et al. (2013, 2021) which performs a spatial discretization of the geotechnical values of each unit,” more details should be provided in the text. 

2. The details (including background and assumptions) regard Equation (1) should be address. Currently, the Section 2.1 is too brief to accept. Some important assumptions such as that there is no overland flow, no significant deep drainage, and no significant flow in the bedrock is not clearly state. When we mentioned the q in the Equation (1), it is the effective precipitation (rainfall minus evapotranspiration). Although the “uniform recharge rate” is acceptable in Line 128, this expression is difficult to establish a direct association with precipitation. The T, transmissivity of soil is the saturated conductivity times soil thickness (precisely, the vertical integral of the saturated conductivity). Notice that you have submitted the thickness of the soil in Line 173. However, you should provide the saturated conductivity either so as to match the parament required of Equation (1).

3. The meaning of log q/T, “instability thresholds” is not well clarified. For example, what is the meaning in Figue 4 for the area marked with red corresponding to “-3.1 <log q/T <-2.8”? From the author’s current statement, the potential readers are very likely to be misled. Frankly speaking, the authors should not expect all potential readers to have a good theoretical knowledge of the SHALSTAB model. Therefore, how to clarify what is the meaning of the output of the model in a simple and understandable way is of the importance that the author should pursue. For example, if we have estimated the transmissivity to be about 65 m2/day, hence a value of “-3.1 <log q/T <-2.8” means that the steady state rainfall (note that precisely should be effective precipitation, i.e., the rainfall minus evapotranspiration) was 51.35 ~102.7 mm/day. While how much the true average steady state rainfall in this area should be carefully studied. In short, the area marked with red corresponding to “-3.1 <log q/T <-2.8” is actually represents the newly-increased unstable area in the area under the range of rainfall (51.35 ~102.7 mm/day), relative to the lower level of precipitation classification (“-3.1 <log q/T <-2.8”). All in all, the author is expected to expressed analysis results of the model to potential readers in a more plain language.

4. For DPIUA method. This method is proposed by authors, so I suggest it should be highlighted in both Section 2.5 and Section Conclusions. Atually, there are only two curves: The first one is the “percentage of cumulative area”, the second one is the “percentage of cumulative area of landslides”. The further of the two curves situated are, the better the performance of the SHALSTAB model. Also, you can express the same meaning at the perspective of “points”. For each log(q/T) category, there exist two points corresponding to this category, the first one is “percentage of cumulative area”, the second one is the “percentage of cumulative area of landslides”. The further of the two points situated are, the better the performance of the SHALSTAB model. So, the DPIUA presents an average difference between two types points. The higher DPIUA, the better performance of model.

5. The conclusion is too weak. It is recommended to divide the conclusions into several points marked as numbers. Some consolidate points of view in the Section Discussion can be regarded as conclusions. For example, you compare with 13 different scenarios, so which one has the best performance (most matching the actual landslide inventory) should be well discussed and then regard as one of the conclusions. Please seriously rewrite the conclusions. More detail requirements, please refer the guidance for authors by MDPI.

6. Strengthen background research. Regarding SHALSTAB model, there are several latest papers closely related to your current study. The following references, but not limited, can be well analysed for the consideration:
[1].     Guimarães, R.F.; Machado, W.P.; De Carvalho, O.A.; Montgomery, D.R.; Gomes, R.A.T.; Greenberg, H.M.; Cataldi, M.; Mendonça, P.C. Determination of Areas Susceptible to Landsliding Using Spatial Patterns of Rainfall from Tropical Rainfall Measuring Mission Data, Rio de Janeiro, Brazil. ISPRS Int. Geo-Inf. 2017, 6, 289.
[2]. Gomes, R.A.T.; Guimaraes, R.F.; de Carvalho, O.A.; Fernandes, N.F.; do Amaral, E.V. Combining Spatial Models for Shallow Landslides and Debris-Flows Prediction. Remote Sensing 2013, 5, 2219-2237, doi:10.3390/rs5052219.
[3]. Cando-Jácome, M.; Martínez-Graña, A. Determination of Primary and Secondary Lahar Flow Paths of the Fuego Volcano (Guatemala) Using Morphometric Parameters. Remote Sensing 2019, 11, 727.
[4]. Konig, T.; Kux, H.J.H.; Mendes, R.M. Shalstab mathematical model and WorldView-2 satellite images to identification of landslide-susceptible areas. Natural Hazards 2019, 97, 1127-1149, doi:10.1007/s11069-019-03691-4.
[5]. Vieira, B.C.; Fernandes, N.F.; Augusto, O.; Martins, T.D.; Montgomery, D.R. Assessing shallow landslide hazards using the TRIGRS and SHALSTAB models, Serra do Mar, Brazil. Environ. Earth Sci. 2018, 77, doi:10.1007/s12665-018-7436-0.

SPECIFIC COMMENTS

7. The abbreviation of “SHALSTAB” should be defined at the first appear in text. This abbreviation stands for “SHAllow Landslide STABle” which is designed for mapping shallow landslide potential.

8. The reference Michel et al. (2013, 2021) should be note as “(in Portuguese)”.

9. Table 2 and Table 3. As mentioned in General comment 2, the saturated conductivity should be provided. 

10. Low quality of Figure 6 and Figure 9. Please use professional software to generate both of them instead of simple tools such as defaulted templates provided by Excel. 

 

Comments for author File: Comments.doc

Author Response

A: We partially agree with this comment. Even though the SHALSTAB has been used for over 20 years, the model’s structure continues to be further developed. As Dietrich and Montgomery (1998) commented, the original model is very simple. But we used its modified version developed by Michel et al. (2013, 2021). Since many papers recently applied the SHALSTAB, explanations of the algorithm and their theory are easily accessible. Hence, we believe that it is unnecessary to present this model's detailed theory and algorithm. For example, nowadays, the hydrological model HEC-HMS and the hydrodynamic model HEC-RAS are so popular that researchers must know their theory and papers that used these models do not explain them in detail. Hence we are sure that the explanation in the present paper, i.e. the content presented in section 2.1 is enough for the readers as they can find more detailed information in previous studies. The equation (1) exactly explains the essential theory of SHALSTAB. Since we cited Montgomery and Dietrich's (1994) original papers, the readers can seek detailed information there. However, following the suggestion, we modified the first paragraph in the section 2.4. Please see the section 2.4.

             The present work adopted the modified algorithm proposed by Michel et al. [25,26], which performs a spatial discretization of the geotechnical values of each unit. This modified version permits different values of soil characteristics such as c, ϕ and z. Besides Michel et al. [25,26], Sbroglia et al. [28] also obtained a good performance when using this modified version.

 

1. The details (including background and assumptions) regard Equation (1) should be address. Currently, the Section 2.1 is too brief to accept. Some important assumptions such as that there is no overland flow, no significant deep drainage, and no significant flow in the bedrock is not clearly state. When we mentioned the q in the Equation (1), it is the effective precipitation (rainfall minus evapotranspiration). Although the “uniform recharge rate” is acceptable in Line 128, this expression is difficult to establish a direct association with precipitation. The T, transmissivity of soil is the saturated conductivity times soil thickness (precisely, the vertical integral of the saturated conductivity). Notice that you have submitted the thickness of the soil in Line 173. However, you should provide the saturated conductivity either so as to match the parament required of Equation (1).

A: Thank you for your suggestions. We believe that, in the community of hydrology, geomorphology and hydrogeomorphology, the term transmissivity (T) is well established as the saturated hydraulic conductivity times soil layer thickness. We added this definition to the paper. Regarding the second question about the overland flow, we require further clarifications. Should we discuss here the Hortonian or Dunne type? The evapotranspiration is presented. In this case, the reviewer wanted to say the potential evapotranspiration or the real one? We modified section 2.1 as below:

SHALSTAB is a deterministic model directed to the identification of sites predisposed to shallow landslides. This model results from combining the infinite slope stability model and a hydrological model [7]. Thus, SHALSTAB determines the areas susceptible to landslides based on the ratio between recharge rate and soil transmissivity sufficient to cause a slope destabilization (Equation 1). Here, soil transmissivity is defined as the product of saturated hydraulic conductivity and depth to a restrictive layer [40]. Detailed descriptions of the SHALSTAB model can be obtained in [7,41]

 

 

2. The meaning of log q/T, “instability thresholds” is not well clarified. For example, what is the meaning in Figue 4 for the area marked with red corresponding to “-3.1 <log q/T <-2.8”? From the author’s current statement, the potential readers are very likely to be misled. Frankly speaking, the authors should not expect all potential readers to have a good theoretical knowledge of the SHALSTAB model. Therefore, how to clarify what is the meaning of the output of the model in a simple and understandable way is of the importance that the author should pursue. For example, if we have estimated the transmissivity to be about 65 m2/day, hence a value of “-3.1 <log q/T <-2.8” means that the steady state rainfall (note that precisely should be effective precipitation, i.e., the rainfall minus evapotranspiration) was 51.35 ~102.7 mm/day. While how much the true average steady state rainfall in this area should be carefully studied. In short, the area marked with red corresponding to “-3.1 <log q/T <-2.8” is actually represents the newly-increased unstable area in the area under the range of rainfall (51.35 ~102.7 mm/day), relative to the lower level of precipitation classification (“-3.1 <log q/T <-2.8”). All in all, the author is expected to expressed analysis results of the model to potential readers in a more plain language.

A: Thanks for the suggestions. Here we would like to emphasize that the SHALSTAB is a very widespread model. Furthermore, MDPI articles encourage the articles to be as direct as possible. Hence, we believe it is more advantageous for the readers to read the model’s manual. However, following your suggestion, we added further explanations in section 2.5 and added a new table.

In this study, we used the instability thresholds as in the original software [7]. SHALSTAB estimates the saturated proportion of soil thickness and originally classifies seven classes of instability, as shown in Table 6. From the value of q/T, the degree of instability for each cell in the study area is calculated. The model results are given on a logarithmic scale because of the very small values obtained for q/T.

 

Table 6 – Instability thresholds adopted in this study [7].

ID	Classes
1	Unconditionally instable
2	log q/T< -3.1
3	-3.1 < log q/T < -2.8
4	-2.8 < log q/T< -2.5
5	-2.5 < log q/T< -2.2
6	-2.2 < log q/T
7	Unconditionally stable

 

3. For DPIUA method. This method is proposed by authors, so I suggest it should be highlighted in both Section 2.5 and Section Conclusions. Atually, there are only two curves: The first one is the “percentage of cumulative area”, the second one is the “percentage of cumulative area of landslides”. The further of the two curves situated are, the better the performance of the SHALSTAB model. Also, you can express the same meaning at the perspective of “points”. For each log(q/T) category, there exist two points corresponding to this category, the first one is “percentage of cumulative area”, the second one is the “percentage of cumulative area of landslides”. The further of the two points situated are, the better the performance of the SHALSTAB model. So, the DPIUA presents an average difference between two types points. The higher DPIUA, the better performance of model.

A: We appreciate the suggestion to highlight the DPIUA in the text more. Agreeing with this suggestion, we added some explanation in the section 2.5 and the results. We also highlighted it in the conclusions.

The elaborated graphical presentation (Figure 9) presents two curves with values for each log(q/T) category: The first one is the “percentage of cumulative area”, the second one is the “percentage of cumulative area of landslides”. The further these curves are, the higher the DPIUA is, and the better the performance of the SHALSTAB model.

4. The conclusion is too weak. It is recommended to divide the conclusions into several points marked as numbers. Some consolidate points of view in the Section Discussion can be regarded as conclusions. For example, you compare with 13 different scenarios, so which one has the best performance (most matching the actual landslide inventory) should be well discussed and then regard as one of the conclusions. Please seriously rewrite the conclusions. More detail requirements, please refer the guidance for authors by MDPI.

A: We agree with the suggestion. Therefore, the section Conclusions was rewritten. Please find bellow the modified version

The SHALSTAB model has been frequently applied to assess shallow landslide susceptibility. In developing countries, where geotechnical unit characterization is difficult due to lack of financial support, many studies measure the soil mechanical properties at few points inside study areas or do not measure at all and use the data available in the SHALSTAB tutorial [32]. Therefore, the present study demonstrated the importance of considering the geotechnical unit characterization on landslide mapping performance by considering the case of the Jaguar creek basin, southern Brazil. In general, results showed that considering sampling points to characterize the geotechnical units provides better results.

To evaluate the model performance, the present study proposed a new index called DPIUA, which allows evaluating the model robustness according to different instability thresholds (Table 6 and Figure 8). Through using several performance evaluation methods, the benefits generated by the characterization of the geotechnical units were verified. These advantages were evident as a result of the better performance resulting from the context in which the soil parameters were discretized into various geotechnical units.

Thus, in summary, future assessments must consider greater cartographic precision to obtain the variability of the geotechnical units in the application of the SHALSTAB model and also a sufficient amount of in-situ soil sampling points to characterize these units. In this regard, it should be noted that the choice of the minimum number of sampling points requires careful consideration. This is one of the most critical issues for field workers who want to be time-efficient while at the same time achieving precise results. This number may depend on (i) soil types within an adopted soil classification, (ii) soil parameters for analysis; and (iii) other physical conditions such as soil covers (vegetation) and soil depth. In any case, adopting geotechnical units in future assessments will surely enable the generation of more reliable landslide susceptibility maps.

5. Strengthen background research. Regarding SHALSTAB model, there are several latest papers closely related to your current study. The following references, but not limited, can be well analysed for the consideration:
   [1].     Guimarães, R.F.; Machado, W.P.; De Carvalho, O.A.; Montgomery, D.R.; Gomes, R.A.T.; Greenberg, H.M.; Cataldi, M.; Mendonça, P.C. Determination of Areas Susceptible to Landsliding Using Spatial Patterns of Rainfall from Tropical Rainfall Measuring Mission Data, Rio de Janeiro, Brazil. ISPRS Int. Geo-Inf. 2017, 6, 289.
   [2]. Gomes, R.A.T.; Guimaraes, R.F.; de Carvalho, O.A.; Fernandes, N.F.; do Amaral, E.V. Combining Spatial Models for Shallow Landslides and Debris-Flows Prediction. Remote Sensing 2013, 5, 2219-2237, doi:10.3390/rs5052219.
   [3]. Cando-Jácome, M.; Martínez-Graña, A. Determination of Primary and Secondary Lahar Flow Paths of the Fuego Volcano (Guatemala) Using Morphometric Parameters. Remote Sensing 2019, 11, 727.
   [4]. Konig, T.; Kux, H.J.H.; Mendes, R.M. Shalstab mathematical model and WorldView-2 satellite images to identification of landslide-susceptible areas. Natural Hazards 2019, 97, 1127-1149, doi:10.1007/s11069-019-03691-4.
   [5]. Vieira, B.C.; Fernandes, N.F.; Augusto, O.; Martins, T.D.; Montgomery, D.R. Assessing shallow landslide hazards using the TRIGRS and SHALSTAB models, Serra do Mar, Brazil. Environ. Earth Sci. 2018, 77, doi:10.1007/s12665-018-7436-0.

A: Thank you very much for informing various papers. As the papers [4] and [5] have been already used for references, we newly cited only the papers [1] to [3]. These are all cited in the item Introduction.

 

SPECIFIC COMMENTS

6. The abbreviation of “SHALSTAB” should be defined at the first appear in text. This abbreviation stands for “SHAllow Landslide STABle” which is designed for mapping shallow landslide potential.

A: Thank you for advice. We put this information in the item Introduction.

 

7. The reference Michel et al. (2013, 2021) should be note as “(in Portuguese)”.

A: Thank you for the correction. We used the Mendeley software to format the references using the sytle recommended by the journal. We will make sure that this information is included in the final version according to the regulations of MDPI

 

8. Table 2 and Table 3. As mentioned in General comment 2, the saturated conductivity should be provided. 

A: The values of saturated conductivity were presented in Tables 2 and 3.

 

9. Low quality of Figure 6 and Figure 9. Please use professional software to generate both of them instead of simple tools such as defaulted templates provided by Excel. 

A: The quality of Figures 6 and 9 was improved as requested.

 

Round 2

Reviewer 3 Report

In this round the authors response SIX general comments and FOUR specific comments. Although the author did not fully follow some GENERAL COMMENTS, they also gave acceptable explanations. I noticed that the manuscript has experienced steady improvement in terms of readability. The thought connotation, framework, and scientific English writing of this manuscript reach the average quality of the papers in GeoHazards.