the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 08 Mar 2022
Hazard Assessment of Earthquake-Induced Landslides Based on a Mechanical Slope Unit Extraction Method, A Case in Ghana
Abstract. Slope unit extraction is integral to earthquake-induced landslide analysis. The conventional watershed and hydrological slope unit extraction methods are precarious with a sudden change in slope gradient along the flow direction, which result in slope unit heterogeneity, conjoint slopes, and boundary defects of the extracted slope unit. This paper addresses this research gap by proposing a mechanical slope unit extraction method that combines watershed points, hydrological, and segmentation methods. This proposed method defines a slope unit as a closed homogeneous space of points overlaid by a mesh having a variance in the slope gradient along its flow direction. The method extracts and uses 3D points to solve slope heterogeneity defects associated with the conventional watershed methods, segmentation to solve boundary defects, and considers the slope pattern and incident ray at a depth to estimate the possibility of earthquake-induced landslides. Ghana (West Africa) is selected to test the proposed slope unit extraction method. The result shows that the method overcame boundary problems, heterogeneity, sudden gradient change, and conjoint slope unit defects associated with the conventional watershed and hydrological method and shows a uniform slope unit for landslide analysis in Ghana. The landslide prediction rate of Ghana also presents 70.9 % landslide inventory, giving an estimated threshold displacement of 9 cm.
- Preprint
(6642 KB) - Metadata XML
- BibTeX
- EndNote
Status: closed
-
RC1: 'Review for paper NHESS‐2022‐43', Anonymous Referee #1, 14 Apr 2022
Review for paper NHESSâ2022â43: “Hazard Assessment of EarthquakeâInduced Landslides Based on a
Mechanical Slope Unit Extraction Method, A Case in Ghana” by Antwi Buah et alii
Dear Authors,
I have evaluated the manuscript “Hazard Assessment of EarthquakeâInduced Landslides Based on a
Mechanical Slope Unit Extraction Method, A Case in Ghana”.
The authors analyse as a case study the whole territory of Ghana proposing a consolidated workflow that
goes through a definition of safety factors for each analysis unit from which a coseismic displacement is
computed by means of the Newmark rigid block method through the quantification critical acceleration K. The
authors systematically analyse the critical aspect in the definitions of coseismic displacement, introducing
factors that account for nonâinfinite slope failure or nonârigid block displacement.
Nevertheless, several critical aspects were found. Given the title of the work, I would have expected a
detailed analysis of the proposed slope unit extraction method, also in relation to the other methods listed
in the manuscript. The authors stated that “Ghana (West Africa) is selected to test the proposed slope unit
extraction method”, however, most of the manuscript deals with an overview of coseismic displacement
analysis methods. The mechanical slope unit extraction method is poorly explained, and the strength and
weaknesses of the method are not explored. I missed what the attribute “mechanical” is referred to in the
proposed method, and which of the considered mechanical parameters are included in the slope unit’s
definition.
In addition, in the summarised approach, it is not clear if slope units or grid cells were adopted as analysis
units. In my opinion, this study seems to have been conducted on grid cell analysis units and not on SU. In the
workflow in Fig. 2 the analysis adopted the slope eight H and α and β angles of Fig.1, however, the images
reported in Fig. 7 and the following variables (e.g., slope angle) seem to contradict it.
According to the Flowchart reported in Fig. 2, the proposed framework aims to provide a "Displacement
prediction for slope units.", however, what is understood from the method is that the slope units are used
only as an area in which calculating the Prediction rate rather than a unit of analysis for evaluating the
coseismic displacement.
If so, how the improvement of SU delineation implemented in the proposed approach can improve the
prediction rate should be better discussed. Are the different slope unit extraction methods more important
than the adopted SU area?
A quantitative analysis of the differences introduced by the three approaches is missing, and there is a
limited quantitative analysis except for the assessment of prediction rates. Only a general comment on the
number and extent of the resulting slope units (lines 306â312) is reported. The parameters considered for
the calculation of the slope units are not clear (e.g., threshold of the accumulation surface of the initial
flow, minimum and maximum surface for the slope units).
Noticeably the manuscript poorly discussed the validation of the computed displacement with
respect to the landslide catalogues. What about the existing or available inventories? I suppose that “scars”
reported in the prediction rate table refers to some landslide catalogue? Are the ones reported in Fig. 10 from
an eventâbased inventory or generic landslides catalogue? The authors report a generic comparison with
the susceptibility map of Ghana which seems to express no more than the spatial relationship between
steep slopes and landslides. If these catalogues are available, are the prediction rate and failure rates
compared to seismically induced landslides?
Regarding the adopted seismic input, how the seismic loading in terms of PGA was selected to 0.13g? Is it
representative of Ghana's seismicity? Are they referred to as a seismic hazard map? Is it comparable to
some of the earthquakes available in the record? Is the seismic input assumed in the analysis the same in
the whole Ghana territory? The authors state that Eq. (20) is “used to determine the extent of earthquake
vibration that can trigger slope displacement in Ghana”: Does the analysis consider a specific seismic
event? Please, clarify. Map of seismic distribution of historical earthquakes, as well as the inventory of ground failure, would help the reader understand.
The manuscript style is not adequately organized in styles and formatting. Figure and not sequentially cited
in the text and often placed in the wrong chapters. Many of them are not useful or simplistic (e.g., Fig. 4 or 5).
Reference to figures and tables in the text are not consecutive, captions not informative and not self-standing.
Tables 2, 3 and 5 can be combined or eliminated. Some of the reported equations are pleonastic
(e.g., Eq. 1, Eq. 7, Eq. 19) and can be substituted by reference.
Despite the adopted techniques being appropriate, the knowledge gap is not clarified or mentioned, and the
expected scientific progress is not clarified by analyses and conclusions. I do not understand the improvement
posed by the proposed SU method for the definition of scenarios or analysis of seismic induction. Regardless
of the specific interest in the case study here analyzed, the work conducts a systematic analysis of the
sensitivity of one or more models and of different mechanical parameters, which has already been the
subject of numerous studies in the literature, therefore, this does not add relevant scientific elements for the
engineeringâgeological community.
I think the subject of the article is within the scope of the journal, but the main objective stated in the
abstract and introduction was not deeply addressed in the manuscript. Major limitations regarding the SU
delineation and the consequent coseismic displacement analysis were found. The mechanical slope unit
extraction method is poorly detailed, as well is not clear the use of SU for displacement analysis.
Given these aspects, the paper would benefit from a deep systematic revision of the presentation of results and
proposed approach with respect to the claimed aim before consideration for publication in Natural Hazard
and Earth System Science. Please note the specific comments included. I hope that you will find the
comments to be of use to you.
Specific Comments
1) The Introduction chapter is unbalanced on the literature empirical relations available for the calculation
of the coseismic displacement and of the different solutions for landslides stability analysis, with respect to
slope units’ delineation approaches.
2) The authors stated from line 80 that the main aspect of the model used and proposed has three distinct
features compared to the others and include: i) the SU delineation, ii) the consideration of poreâwater
distribution and iii) the GIS computation of Fs and the ky to avoid iterative errors:
The first important aspect is the SU definition approach that should solve slope heterogeneity defects is not
adequately presented and is not clear and easy to understand by the general audience. Secondarily, the role
and the areal constraints of pore pressure (hydrostatic?) lacks in the whole manuscript. Is it considered
parametrically? Does the SF of Fig. 7 account for the ru pore pressure ratio reported in the methods
section? Finally, the areal GISâbased quantification of SF and Ky appears to be a commonly used approach in
the scientific community.
3) The rigid and flexible block effect is considered important and treated analytically, however, the
contribution of cohesion in earth shallow landslides is considered irrelevant, which are mainly governed by
the effect of apparent and mechanical cohesion of unsaturated media. A critical comment about this topic
would improve the manuscript. In addition, eq. (8) by Saygili & Rathje, 2009 is dependent on cohesion.
Which values have been adopted?
4) Are the landslides plotted in Figures seismically induced? In some figures are reported locations of
“failure areas and catalogue” that are not explained and/or not fully considered in the validation of results.
5) Regarding the Prediction rate I’d spend more effort in discussing the general concept and the meaning of
the two curves for the validation of the results. With a few landslides’ observations, I supposed the Pr is
overbalanced by a large number of negatives. I’ll express the success rate as the ratio between true positive
rate and false-positive rate. Would the success rate for true positive be more informative? What about the
effect of strength parameters on the lone True Positives (S1)?
6) Concepts expressed in methods are often repeated in the results section. Technical language is often less
precise (see Detailed comments). The introduction of the chapter “Seismic activity of Ghana” is mainly
focused on African landslides, however, a detailed analysis of the available landslide catalogue lacks. I would
suggest changing the title or rephrasing the introduction to the chapter.
7) The manuscript is not adequately cared for in styles and formatting. Figure and not sequentially cited in the text
and often placed in the wrong sections. Many of them are not useful or simplistic. Reference to figures and
table are not consecutive, captions not informative and not selfâstanding. Please consider a detailed review
according to Journal standards.
Detailed Comments
Line 20: Please change the typos sentence that is supposed to be “in order to”
Line 36: Newmark instead of Newark
Line 60: Please deeply explain the limitation in reflecting morphological features
Line 101: Do the authors referred to river thalweg with the term “crevasses”?
Line 116: It is not tensile. It corresponds to the shear stress component parallel to the failure surface
Line 121: “m” should be formatted in Italic
Line 139: References are cited twice
Line 143: Please rephrase the definition of Ky(g)
Line 190: The list of coordinates can be omitted in the text since are included in the figures.
Lines 192â194: Please use the International System of Units
Line 148: Reference of DCF is missing. The factor is reported sometimes in subscript, please uniform it.
Line 205: The Romanche Transform fault and earthquake epicentres should be located on Map.
Line 217: Reference to figure missing
Line 227: Unit weight cannot be considered a strength parameter
Line 258: Annum
Line 323: “Ninetyânine per cent correlation is obtained”. I missed the presentation on this correlation
analysis.
Line 337: Ts/Tm areal or lithological distribution is not discussed.
Table 2: In the percentage column 1/3 has been changed to 3/Jan by the corrector
Comment on References and Figures
âMost of the figures are not informative (Es. Fig. 4 or Fig. 5) and can be delated.
âFig. 7a the redâgreen colour bar looks counterintuitive to express the SF of slopes.
âFig. 14 caption is wrong, looks like a repetition of fig 13.
âTable 2 can be merged with Table 3 since it is not informative. Areal percentage distribution of the two
major complexes is not indicative. I’d report it for the lithological unit in Table 3.
âFig. 5 can be merged with fig. 6.
âFig. 6 can be improved including the relief map reported in Appendix Fig.A1.
âFig. 9 It is difficult to appreciate the difference between the three adopted models.
âFig. 12 Change "Hydrolical"
âBu et alii 2019 is not listed in the referenceCitation: https://doi.org/10.5194/nhess-2022-43-RC1 - AC1: 'Reply on RC1', Yingbin Zhang, 21 Jul 2022
-
CC1: 'Comment on nhess-2022-43', Dongliang Huang, 23 Jun 2022
I have evaluated the manuscript "Hazard Assessment of Earthquake-Induced Landslides Based on a Mechanical Slope Unit Extraction Method, A Case in Ghana." The authors methodically dissect the fundamental viewpoint of the meanings of coseismic dislodging, presenting factors that record an infinite slope failure or flexible sliding block failure, as well as proposing a mechanical slope unit extraction method. However, perspectives are found that need verification or correction.
- The manuscript is not adequately organized in style and formatting.
- The introduction must adequately present the proposed slope unit method and the gap in the displacement method used.
- Eliminating the boundary and heterogeneity effect of the hydrological slope unit extraction method are the main innovations of the newly proposed slope unit extraction method in the manuscript; however, they have not been adequately explained in the manuscript. An explanation of the heterogeneity and boundary effect will sufficiently improve the manuscript.
- Fig. 4 can be deleted because it is not serving any purpose.
- Line 148, The Depth Correction Factor (DCF), should be cited and well written.
- Fig.5 should be modified to contain the geological details of the study area (Ghana).
- The heading "seismic activities of Ghana ", seems to be more focused on seismic activities in Africa rather than Ghana.
- Table 2, Table.3, and Table.4 can be merged and summarized because they seem to be repetitive and not very informative.
- Fig.6 and Fig .7 should be modified to reflect the actual safety factor and critical acceleration of Ghana.
- The authors should also check the typing errors in the manuscript, for example, Line 34 and 36, Newmark and not Newark. The authors need to check some of the grammatical errors and correct them.
- Fig.16 isn't informative and can be deleted because the highest elevation of the study area is already at line 191 of the manuscript.
Apart from these few comments, I think the subject and principle of the manuscript are within the scope of the journal. Minor revisions are needed to improve the quality of this manuscript.
Citation: https://doi.org/10.5194/nhess-2022-43-CC1 - AC4: 'Reply on CC1', Yingbin Zhang, 21 Jul 2022
-
RC2: 'Comment on nhess-2022-43', Anonymous Referee #2, 23 Jun 2022
In this paper, the authors claim that available methods used to generate slope units have some significant drawbacks. To address those issues, they propose a technique called mechanical slope unit extraction. Specifically, they focus on an area in Ghana where they claim that earthquake-triggered landslides could provide a good case study to prove the validity of their approach.
Overall, slope units are widely used in landslide susceptibility and hazard assessment as an alternative landscape partition to grid-cells. In this regard, the manuscript has a reasonably good motivation. However, there are some fairly large points that I believe need further consideration. Below, I outline two main points to address before giving some detailed line-by-line comments.
Main comments:
1- Unfortunately, the manuscript was not written well and this effect my revision as a whole because I had a hard time following the flow and messages the authors are trying to convey. There are some terms (e.g., conventional, hydrological and mechanical methods) that are not really well explained but continuously being used throughout the manuscript. I am sure the authors are clear about what they are referring to but unfortunately the same is not valid for the readership. Also, the structure of the manuscript is a bit “out of ordinary”. There is no flow like introduction, study area, data, method, results, discussion and conclusions. This also makes the text difficult to follow.
2- I am quite surprised to see that there is no landslide inventory or a specific earthquake the authors examined but still the manuscript is evolving around earthquake-induced landslides. I still do not understand how this could be possible. I might have missed something because of the reasons I mentioned above. However, taking this possibility aside, I do not understand the rationale of the manuscript. If there are no landslides triggered by earthquakes, the whole promise of the manuscript is just hanging in the air.
I was planning to provide comments in detail for the entire manuscript but after realizing these issues mentioned above, I cut it short because, with all due respect to the authors’ labor on the manuscript, I have to say those small revisions would not be adequate to make it publishable. Still, below I've included some of my line-by-line suggestions.
Lines 17-18: “Landslides occur on slopes and have been the rationale behind making earthquake-induced landslides and seismic engineering a scientific and national demand”. It is not clear what you mean here.
Line 18: “its evaluation provides general estimates of future earthquake-induced landslides” Do you mean evolution of earthquake-induced landslides? What do you mean? Or are you referring to the evolution of techniques that we use to assess landslide hazards or something? This line is not clear. Please revise and express it in a clearer way.
Line 21: You can still examine historical landslide catalogs and address some research questions even today, right? What do you mean? Do you think it is not an old fashion approach?
Lines 22-23: “current scientific and engineering stability analysis models” What are you referring to? Could you please be more specific?
Line 24: “The statistical method” Are you referring to a statistically-based method developed to assess landslide susceptibility? What sort of method are you referring to? Please be more specific. Also, from lines 24 to 28, you do not say “such as” each and every time. You can just say, for instance:
multivariate statistics (i.e., Logistic Regression, LR; Atkinson & Massari, 1998)
Btw, I also noticed that you do not refer to, for instance, the logistic regression later on in the text. If this is the case, you do not need to indicate the abbreviation either.
Line 29: “The robustness of the statistical method is, however, suspect”. Of course, there is no perfect model. But you can not say that “The robustness of the statistical method is, however, suspect”. Based on what? You can be critical for sure and mention some uncertainties, but not like this. Please revise the line.
Line 29: “statistical method generates landslide maps” What do you mean? Do you mean “landslide susceptibility maps”? You are resisting saying what we are really talking about. Susceptibility? Hazard? Or something else?
Lines 29-31: “Because the statistical method generates landslide maps by using a combination of maps generated by different control points that are assumed to be conditionally independent of each other, thus questioning its accuracy”. What does this mean now? I have a hard time following the logic behind your argument. You are saying they are mostly using predictive variables which suffer from multicollinearity and this is why outputs of the statistically-based method have some accuracy issues? Then please say it clearly. Btw, I would say this is quite a minor issue among many others regarding the statistically-based method and actually, there are ways to deal with this issue in the literature. If you would like to be really critical, you should come up with better arguments/stronger. Also, are you sure that these papers support your argument (e.g., Youssef & Pourghasemi, 2021). Or these are the examples that you think that they are clearly representing the problem you mention.
Lines 31-32: “Recent engineering earthquake induced landslides and displacement analysis…(Rathje et al., 1998; Jibson & Keefer, 1993; Saygili, 2008)” these do not sound “recent” to me. Btw, please rewrite the line, there is no such thing called “engineering earthquake induced landslides”. Also, “analysis” should be “analyses”.
Line 34: “Newark Rigid Dynamic Block Model” should be “Newmark's sliding block method”. Please do the same corrections through the text.
Lines 34-36: “The precision of the Newark Rigid Dynamic Block Model cannot be misconstrued, as it produces a stronger correlation between the estimated sliding block displacement and the mapping location of the earthquake-triggered landslide”. Please do not say “cannot be misconstrued” out of blue and please support your argument by citing the literature. This is quite a subjective statement I would say. As there is no perfect machine learning technique to assess the spatial distribution of landslides, the same is also valid for their physically-based counterparts. If you would like to list the pros and cons of both approaches, you have to do it in an objective manner. Btw, this may not be even required because I still do not understand where you want to go from here. Do we really need to list all these here? Are these relevant for this paper? There is a large literature associated with both approaches and no need to destroy or glorify one of them compared to the other. However, if you would like to do this then do it properly. For instance, how accurately can you identify geotechnical parameters to run a regional-scale landslide susceptibility/hazard analyses?
Line 38: Why is that the final stage?
Line 39: “achievable through slope Mapping units”. Please first tell us what the slope unit is and why you prefer working with it. Also, please prefer using either “Slope units” or “mapping units” not both.
Line 40: You started using susceptibility (e.g., in line 39) and hazard (e.g., in line 40) terms and which is ok but do not use them as if you can use them interchangeably. They are not the same thing, right?
Line 42: “slope unit model” there is no model, just slope units. Also, if you are referring to slope units, it is more appropriate to cite papers that proposed slope units, not the ones that only used them based on available sources.
Line 44: “(Xie et al., 2003)” please remove the parentheses
Line 48: “the best mapping unit for earthquake-induced landslide and displacement analysis” Do you mean it is the best mapping unit specifically for earthquake-induced landslides? Why is that? How about rainfall-triggered landslides?
Line 48: “The slope unit method” Slope unit is not the method but the output of some landscape partitioning methods which you haven’t mentioned yet.
Lines 51-54: Please separately cite the corresponding paper of each method you mention.
Lines 62-63: “The conventional watershed method for slope unit extraction” Please be specific and cite corresponding papers.
Line 68: “The application of the framework is validated in Ghana.” Which earthquake is that?
Line 70: “The impact of cohesion c is negligible, therefore neglected” In which context? You did not say anything about the landslides you examined. You can keep this statement for your method section and better explain it there.
Lines 71-72: I did not understand what you mean here.Citation: https://doi.org/10.5194/nhess-2022-43-RC2 - AC2: 'Reply on RC2', Yingbin Zhang, 21 Jul 2022
- AC3: 'Reply on RC2', Yingbin Zhang, 21 Jul 2022
- AC5: 'Comment on nhess-2022-43', Yingbin Zhang, 25 Jul 2022
Status: closed
-
RC1: 'Review for paper NHESS‐2022‐43', Anonymous Referee #1, 14 Apr 2022
Review for paper NHESSâ2022â43: “Hazard Assessment of EarthquakeâInduced Landslides Based on a
Mechanical Slope Unit Extraction Method, A Case in Ghana” by Antwi Buah et alii
Dear Authors,
I have evaluated the manuscript “Hazard Assessment of EarthquakeâInduced Landslides Based on a
Mechanical Slope Unit Extraction Method, A Case in Ghana”.
The authors analyse as a case study the whole territory of Ghana proposing a consolidated workflow that
goes through a definition of safety factors for each analysis unit from which a coseismic displacement is
computed by means of the Newmark rigid block method through the quantification critical acceleration K. The
authors systematically analyse the critical aspect in the definitions of coseismic displacement, introducing
factors that account for nonâinfinite slope failure or nonârigid block displacement.
Nevertheless, several critical aspects were found. Given the title of the work, I would have expected a
detailed analysis of the proposed slope unit extraction method, also in relation to the other methods listed
in the manuscript. The authors stated that “Ghana (West Africa) is selected to test the proposed slope unit
extraction method”, however, most of the manuscript deals with an overview of coseismic displacement
analysis methods. The mechanical slope unit extraction method is poorly explained, and the strength and
weaknesses of the method are not explored. I missed what the attribute “mechanical” is referred to in the
proposed method, and which of the considered mechanical parameters are included in the slope unit’s
definition.
In addition, in the summarised approach, it is not clear if slope units or grid cells were adopted as analysis
units. In my opinion, this study seems to have been conducted on grid cell analysis units and not on SU. In the
workflow in Fig. 2 the analysis adopted the slope eight H and α and β angles of Fig.1, however, the images
reported in Fig. 7 and the following variables (e.g., slope angle) seem to contradict it.
According to the Flowchart reported in Fig. 2, the proposed framework aims to provide a "Displacement
prediction for slope units.", however, what is understood from the method is that the slope units are used
only as an area in which calculating the Prediction rate rather than a unit of analysis for evaluating the
coseismic displacement.
If so, how the improvement of SU delineation implemented in the proposed approach can improve the
prediction rate should be better discussed. Are the different slope unit extraction methods more important
than the adopted SU area?
A quantitative analysis of the differences introduced by the three approaches is missing, and there is a
limited quantitative analysis except for the assessment of prediction rates. Only a general comment on the
number and extent of the resulting slope units (lines 306â312) is reported. The parameters considered for
the calculation of the slope units are not clear (e.g., threshold of the accumulation surface of the initial
flow, minimum and maximum surface for the slope units).
Noticeably the manuscript poorly discussed the validation of the computed displacement with
respect to the landslide catalogues. What about the existing or available inventories? I suppose that “scars”
reported in the prediction rate table refers to some landslide catalogue? Are the ones reported in Fig. 10 from
an eventâbased inventory or generic landslides catalogue? The authors report a generic comparison with
the susceptibility map of Ghana which seems to express no more than the spatial relationship between
steep slopes and landslides. If these catalogues are available, are the prediction rate and failure rates
compared to seismically induced landslides?
Regarding the adopted seismic input, how the seismic loading in terms of PGA was selected to 0.13g? Is it
representative of Ghana's seismicity? Are they referred to as a seismic hazard map? Is it comparable to
some of the earthquakes available in the record? Is the seismic input assumed in the analysis the same in
the whole Ghana territory? The authors state that Eq. (20) is “used to determine the extent of earthquake
vibration that can trigger slope displacement in Ghana”: Does the analysis consider a specific seismic
event? Please, clarify. Map of seismic distribution of historical earthquakes, as well as the inventory of ground failure, would help the reader understand.
The manuscript style is not adequately organized in styles and formatting. Figure and not sequentially cited
in the text and often placed in the wrong chapters. Many of them are not useful or simplistic (e.g., Fig. 4 or 5).
Reference to figures and tables in the text are not consecutive, captions not informative and not self-standing.
Tables 2, 3 and 5 can be combined or eliminated. Some of the reported equations are pleonastic
(e.g., Eq. 1, Eq. 7, Eq. 19) and can be substituted by reference.
Despite the adopted techniques being appropriate, the knowledge gap is not clarified or mentioned, and the
expected scientific progress is not clarified by analyses and conclusions. I do not understand the improvement
posed by the proposed SU method for the definition of scenarios or analysis of seismic induction. Regardless
of the specific interest in the case study here analyzed, the work conducts a systematic analysis of the
sensitivity of one or more models and of different mechanical parameters, which has already been the
subject of numerous studies in the literature, therefore, this does not add relevant scientific elements for the
engineeringâgeological community.
I think the subject of the article is within the scope of the journal, but the main objective stated in the
abstract and introduction was not deeply addressed in the manuscript. Major limitations regarding the SU
delineation and the consequent coseismic displacement analysis were found. The mechanical slope unit
extraction method is poorly detailed, as well is not clear the use of SU for displacement analysis.
Given these aspects, the paper would benefit from a deep systematic revision of the presentation of results and
proposed approach with respect to the claimed aim before consideration for publication in Natural Hazard
and Earth System Science. Please note the specific comments included. I hope that you will find the
comments to be of use to you.
Specific Comments
1) The Introduction chapter is unbalanced on the literature empirical relations available for the calculation
of the coseismic displacement and of the different solutions for landslides stability analysis, with respect to
slope units’ delineation approaches.
2) The authors stated from line 80 that the main aspect of the model used and proposed has three distinct
features compared to the others and include: i) the SU delineation, ii) the consideration of poreâwater
distribution and iii) the GIS computation of Fs and the ky to avoid iterative errors:
The first important aspect is the SU definition approach that should solve slope heterogeneity defects is not
adequately presented and is not clear and easy to understand by the general audience. Secondarily, the role
and the areal constraints of pore pressure (hydrostatic?) lacks in the whole manuscript. Is it considered
parametrically? Does the SF of Fig. 7 account for the ru pore pressure ratio reported in the methods
section? Finally, the areal GISâbased quantification of SF and Ky appears to be a commonly used approach in
the scientific community.
3) The rigid and flexible block effect is considered important and treated analytically, however, the
contribution of cohesion in earth shallow landslides is considered irrelevant, which are mainly governed by
the effect of apparent and mechanical cohesion of unsaturated media. A critical comment about this topic
would improve the manuscript. In addition, eq. (8) by Saygili & Rathje, 2009 is dependent on cohesion.
Which values have been adopted?
4) Are the landslides plotted in Figures seismically induced? In some figures are reported locations of
“failure areas and catalogue” that are not explained and/or not fully considered in the validation of results.
5) Regarding the Prediction rate I’d spend more effort in discussing the general concept and the meaning of
the two curves for the validation of the results. With a few landslides’ observations, I supposed the Pr is
overbalanced by a large number of negatives. I’ll express the success rate as the ratio between true positive
rate and false-positive rate. Would the success rate for true positive be more informative? What about the
effect of strength parameters on the lone True Positives (S1)?
6) Concepts expressed in methods are often repeated in the results section. Technical language is often less
precise (see Detailed comments). The introduction of the chapter “Seismic activity of Ghana” is mainly
focused on African landslides, however, a detailed analysis of the available landslide catalogue lacks. I would
suggest changing the title or rephrasing the introduction to the chapter.
7) The manuscript is not adequately cared for in styles and formatting. Figure and not sequentially cited in the text
and often placed in the wrong sections. Many of them are not useful or simplistic. Reference to figures and
table are not consecutive, captions not informative and not selfâstanding. Please consider a detailed review
according to Journal standards.
Detailed Comments
Line 20: Please change the typos sentence that is supposed to be “in order to”
Line 36: Newmark instead of Newark
Line 60: Please deeply explain the limitation in reflecting morphological features
Line 101: Do the authors referred to river thalweg with the term “crevasses”?
Line 116: It is not tensile. It corresponds to the shear stress component parallel to the failure surface
Line 121: “m” should be formatted in Italic
Line 139: References are cited twice
Line 143: Please rephrase the definition of Ky(g)
Line 190: The list of coordinates can be omitted in the text since are included in the figures.
Lines 192â194: Please use the International System of Units
Line 148: Reference of DCF is missing. The factor is reported sometimes in subscript, please uniform it.
Line 205: The Romanche Transform fault and earthquake epicentres should be located on Map.
Line 217: Reference to figure missing
Line 227: Unit weight cannot be considered a strength parameter
Line 258: Annum
Line 323: “Ninetyânine per cent correlation is obtained”. I missed the presentation on this correlation
analysis.
Line 337: Ts/Tm areal or lithological distribution is not discussed.
Table 2: In the percentage column 1/3 has been changed to 3/Jan by the corrector
Comment on References and Figures
âMost of the figures are not informative (Es. Fig. 4 or Fig. 5) and can be delated.
âFig. 7a the redâgreen colour bar looks counterintuitive to express the SF of slopes.
âFig. 14 caption is wrong, looks like a repetition of fig 13.
âTable 2 can be merged with Table 3 since it is not informative. Areal percentage distribution of the two
major complexes is not indicative. I’d report it for the lithological unit in Table 3.
âFig. 5 can be merged with fig. 6.
âFig. 6 can be improved including the relief map reported in Appendix Fig.A1.
âFig. 9 It is difficult to appreciate the difference between the three adopted models.
âFig. 12 Change "Hydrolical"
âBu et alii 2019 is not listed in the referenceCitation: https://doi.org/10.5194/nhess-2022-43-RC1 - AC1: 'Reply on RC1', Yingbin Zhang, 21 Jul 2022
-
CC1: 'Comment on nhess-2022-43', Dongliang Huang, 23 Jun 2022
I have evaluated the manuscript "Hazard Assessment of Earthquake-Induced Landslides Based on a Mechanical Slope Unit Extraction Method, A Case in Ghana." The authors methodically dissect the fundamental viewpoint of the meanings of coseismic dislodging, presenting factors that record an infinite slope failure or flexible sliding block failure, as well as proposing a mechanical slope unit extraction method. However, perspectives are found that need verification or correction.
- The manuscript is not adequately organized in style and formatting.
- The introduction must adequately present the proposed slope unit method and the gap in the displacement method used.
- Eliminating the boundary and heterogeneity effect of the hydrological slope unit extraction method are the main innovations of the newly proposed slope unit extraction method in the manuscript; however, they have not been adequately explained in the manuscript. An explanation of the heterogeneity and boundary effect will sufficiently improve the manuscript.
- Fig. 4 can be deleted because it is not serving any purpose.
- Line 148, The Depth Correction Factor (DCF), should be cited and well written.
- Fig.5 should be modified to contain the geological details of the study area (Ghana).
- The heading "seismic activities of Ghana ", seems to be more focused on seismic activities in Africa rather than Ghana.
- Table 2, Table.3, and Table.4 can be merged and summarized because they seem to be repetitive and not very informative.
- Fig.6 and Fig .7 should be modified to reflect the actual safety factor and critical acceleration of Ghana.
- The authors should also check the typing errors in the manuscript, for example, Line 34 and 36, Newmark and not Newark. The authors need to check some of the grammatical errors and correct them.
- Fig.16 isn't informative and can be deleted because the highest elevation of the study area is already at line 191 of the manuscript.
Apart from these few comments, I think the subject and principle of the manuscript are within the scope of the journal. Minor revisions are needed to improve the quality of this manuscript.
Citation: https://doi.org/10.5194/nhess-2022-43-CC1 - AC4: 'Reply on CC1', Yingbin Zhang, 21 Jul 2022
-
RC2: 'Comment on nhess-2022-43', Anonymous Referee #2, 23 Jun 2022
In this paper, the authors claim that available methods used to generate slope units have some significant drawbacks. To address those issues, they propose a technique called mechanical slope unit extraction. Specifically, they focus on an area in Ghana where they claim that earthquake-triggered landslides could provide a good case study to prove the validity of their approach.
Overall, slope units are widely used in landslide susceptibility and hazard assessment as an alternative landscape partition to grid-cells. In this regard, the manuscript has a reasonably good motivation. However, there are some fairly large points that I believe need further consideration. Below, I outline two main points to address before giving some detailed line-by-line comments.
Main comments:
1- Unfortunately, the manuscript was not written well and this effect my revision as a whole because I had a hard time following the flow and messages the authors are trying to convey. There are some terms (e.g., conventional, hydrological and mechanical methods) that are not really well explained but continuously being used throughout the manuscript. I am sure the authors are clear about what they are referring to but unfortunately the same is not valid for the readership. Also, the structure of the manuscript is a bit “out of ordinary”. There is no flow like introduction, study area, data, method, results, discussion and conclusions. This also makes the text difficult to follow.
2- I am quite surprised to see that there is no landslide inventory or a specific earthquake the authors examined but still the manuscript is evolving around earthquake-induced landslides. I still do not understand how this could be possible. I might have missed something because of the reasons I mentioned above. However, taking this possibility aside, I do not understand the rationale of the manuscript. If there are no landslides triggered by earthquakes, the whole promise of the manuscript is just hanging in the air.
I was planning to provide comments in detail for the entire manuscript but after realizing these issues mentioned above, I cut it short because, with all due respect to the authors’ labor on the manuscript, I have to say those small revisions would not be adequate to make it publishable. Still, below I've included some of my line-by-line suggestions.
Lines 17-18: “Landslides occur on slopes and have been the rationale behind making earthquake-induced landslides and seismic engineering a scientific and national demand”. It is not clear what you mean here.
Line 18: “its evaluation provides general estimates of future earthquake-induced landslides” Do you mean evolution of earthquake-induced landslides? What do you mean? Or are you referring to the evolution of techniques that we use to assess landslide hazards or something? This line is not clear. Please revise and express it in a clearer way.
Line 21: You can still examine historical landslide catalogs and address some research questions even today, right? What do you mean? Do you think it is not an old fashion approach?
Lines 22-23: “current scientific and engineering stability analysis models” What are you referring to? Could you please be more specific?
Line 24: “The statistical method” Are you referring to a statistically-based method developed to assess landslide susceptibility? What sort of method are you referring to? Please be more specific. Also, from lines 24 to 28, you do not say “such as” each and every time. You can just say, for instance:
multivariate statistics (i.e., Logistic Regression, LR; Atkinson & Massari, 1998)
Btw, I also noticed that you do not refer to, for instance, the logistic regression later on in the text. If this is the case, you do not need to indicate the abbreviation either.
Line 29: “The robustness of the statistical method is, however, suspect”. Of course, there is no perfect model. But you can not say that “The robustness of the statistical method is, however, suspect”. Based on what? You can be critical for sure and mention some uncertainties, but not like this. Please revise the line.
Line 29: “statistical method generates landslide maps” What do you mean? Do you mean “landslide susceptibility maps”? You are resisting saying what we are really talking about. Susceptibility? Hazard? Or something else?
Lines 29-31: “Because the statistical method generates landslide maps by using a combination of maps generated by different control points that are assumed to be conditionally independent of each other, thus questioning its accuracy”. What does this mean now? I have a hard time following the logic behind your argument. You are saying they are mostly using predictive variables which suffer from multicollinearity and this is why outputs of the statistically-based method have some accuracy issues? Then please say it clearly. Btw, I would say this is quite a minor issue among many others regarding the statistically-based method and actually, there are ways to deal with this issue in the literature. If you would like to be really critical, you should come up with better arguments/stronger. Also, are you sure that these papers support your argument (e.g., Youssef & Pourghasemi, 2021). Or these are the examples that you think that they are clearly representing the problem you mention.
Lines 31-32: “Recent engineering earthquake induced landslides and displacement analysis…(Rathje et al., 1998; Jibson & Keefer, 1993; Saygili, 2008)” these do not sound “recent” to me. Btw, please rewrite the line, there is no such thing called “engineering earthquake induced landslides”. Also, “analysis” should be “analyses”.
Line 34: “Newark Rigid Dynamic Block Model” should be “Newmark's sliding block method”. Please do the same corrections through the text.
Lines 34-36: “The precision of the Newark Rigid Dynamic Block Model cannot be misconstrued, as it produces a stronger correlation between the estimated sliding block displacement and the mapping location of the earthquake-triggered landslide”. Please do not say “cannot be misconstrued” out of blue and please support your argument by citing the literature. This is quite a subjective statement I would say. As there is no perfect machine learning technique to assess the spatial distribution of landslides, the same is also valid for their physically-based counterparts. If you would like to list the pros and cons of both approaches, you have to do it in an objective manner. Btw, this may not be even required because I still do not understand where you want to go from here. Do we really need to list all these here? Are these relevant for this paper? There is a large literature associated with both approaches and no need to destroy or glorify one of them compared to the other. However, if you would like to do this then do it properly. For instance, how accurately can you identify geotechnical parameters to run a regional-scale landslide susceptibility/hazard analyses?
Line 38: Why is that the final stage?
Line 39: “achievable through slope Mapping units”. Please first tell us what the slope unit is and why you prefer working with it. Also, please prefer using either “Slope units” or “mapping units” not both.
Line 40: You started using susceptibility (e.g., in line 39) and hazard (e.g., in line 40) terms and which is ok but do not use them as if you can use them interchangeably. They are not the same thing, right?
Line 42: “slope unit model” there is no model, just slope units. Also, if you are referring to slope units, it is more appropriate to cite papers that proposed slope units, not the ones that only used them based on available sources.
Line 44: “(Xie et al., 2003)” please remove the parentheses
Line 48: “the best mapping unit for earthquake-induced landslide and displacement analysis” Do you mean it is the best mapping unit specifically for earthquake-induced landslides? Why is that? How about rainfall-triggered landslides?
Line 48: “The slope unit method” Slope unit is not the method but the output of some landscape partitioning methods which you haven’t mentioned yet.
Lines 51-54: Please separately cite the corresponding paper of each method you mention.
Lines 62-63: “The conventional watershed method for slope unit extraction” Please be specific and cite corresponding papers.
Line 68: “The application of the framework is validated in Ghana.” Which earthquake is that?
Line 70: “The impact of cohesion c is negligible, therefore neglected” In which context? You did not say anything about the landslides you examined. You can keep this statement for your method section and better explain it there.
Lines 71-72: I did not understand what you mean here.Citation: https://doi.org/10.5194/nhess-2022-43-RC2 - AC2: 'Reply on RC2', Yingbin Zhang, 21 Jul 2022
- AC3: 'Reply on RC2', Yingbin Zhang, 21 Jul 2022
- AC5: 'Comment on nhess-2022-43', Yingbin Zhang, 25 Jul 2022
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 668 | 327 | 39 | 1,034 | 21 | 21 |
- HTML: 668
- PDF: 327
- XML: 39
- Total: 1,034
- BibTeX: 21
- EndNote: 21
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1