An Improved Method for the Evaluation and Local Multi-Scale Optimization of the Automatic Extraction of Slope Units in Complex Terrains
Round 1
Reviewer 1 Report
It is ok.
Best regards
Author Response
Dear Reviewers:
We would like to resubmit our revised manuscript entitled “An improved method for the evaluation and local multi-scale optimization of the automatic extraction of slope units in com-plex terrains”. We would like to thank the editor and the reviewers for thoroughly reviewing our manuscript and making many thoughtful comments and precious suggestions. We are very pleased to see that you recognized the novelty and potential significance of our work. We have added necessary data and references and revised the manuscript to address reviewers’ comments.
Yours sincerely
Dr. Wei Jinbing
Sichuan University
Author Response File: 
Author Response.docx
Reviewer 2 Report
The paper entitled “A new method for the assessment and local multi-scale optimization of the automatic extraction of slope units in complex terrains” presents a method for slope units extraction exploiting the r.slopeunit software developed by Alvioli et al. in 2016 with some modifications to improve the validity and typicality of the SU subdivision, with an application in the Yuqu River Basin in China. In general, the manuscript is well written and well organized, though innovative element remains unclear. In particular, for the methodology proposed by the authors that exploits mainly previous methods, developed and presented by other authors, specifically the r.slopeunit software developed by Alvioli et al. (2016) for the Slope Unit creation (section 3.1), the Object Consistency Error (OCE) proposed by Polak et al. (2009) and the Moran’s I measure for the determination of the optimal subdivision scale (section 3.2), the Local Average Similarity (LS) method for the identification of OSSUs and USSUs (section 3.3). The use of a series of previous methods derived from other authors make it difficult to recognize the novelty in type of application and methodology used in this work. Furthermore, this aspect makes it difficult to speak about “new method” as highlighted in the title of the paper, but rather about an improvement of an existing method.
For this reason, the manuscript cannot be accepted in its present form and a major revision is recommended.
Other minor aspects:
Abstract, Line 18 “This method produces improved SUs…” this sentence is unclear. What the improvement refers to?
Page 14 Line 447 – The authors cite the Otsu method, please add a proper reference.
Author Response
Dear Reviewers:
Re: Manuscript reference No. remotesensing-1789297：An improved method for the evaluation and local multi-scale optimization of the automatic extraction of slope units in complex terrains
Thank you for your valuable time and effort to review this article. I read your comments carefully, and they are valuable for highlighting the innovations and unique contributions of the article. In order to clarify the innovative elements of the proposed method, the introduction is rewritten and the defects in the existing work are pointed out. Accordingly, necessary new descriptions are added to the methods and results to highlight the novelty of the improved methods. The response letter is divided into two parts, the first part briefly describes the innovations and contributions of the article, and the second part details of the revisions. The intrinsic difficulties of determining input parameters and correcting for unreasonable SUs have hindered their wide application. The optimal subdivision scale was obtained by a collaborative evaluation approach capable of simultaneously measuring objective minimum discrepancies and seeking a global optimum. Moreover, a new effective optimization mechanism containing the re-subdivision of USSUs and merging of OSSUs was put into effect. We thus obtained SUs composed of terrain subdivisions with multiple scales. This is infrequently available methods for fusing different scales currently. Benefiting from the sufficient integrating approach of diverse features for each object, it is a significant advantage that the processing object can be transferred from general entirety to each precise individual. This is the main innovation of this article, because it breaks through the difficulty of distinguishing unreasonable SUs and combining them at multiple scales.
In addition, this manuscript has been thoroughly edited by a native English speaker from an editing company. For more detailed revisions, please refer to the attachments uploaded.
Yours sincerely
Dr. Wei Jinbing
Sichuan University
Author Response File: Author Response.docx
This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.
Round 1
Reviewer 1 Report
The paper proposes a new method for the assessment and local multi-scale optimization of the automatic extraction of slope units (SU) in complex terrains
The proposed method produces improved SUs through a stepwise approach from a global optimum to a local scale in order to avoid/minimize the presence of OSSUs and USSUs.
The object-level consistency error (OCE) is used as key parameter for the selection of the optimal scale.
I understand that one of the main novelties introduced by this work is that reference data is used to evaluate the result of the algorithm (the generated SUs).
I consider that it is a well-organized work and that it demonstrates the authors' control over the matter.
The introduction is adequate.
The work area is adequately described.
The preparation of the reference data requires more detail. It is not indicated how many technicians have outlined the SUs. It is important that they are experts and that this circumstance is supported by evidence. If the reference has been generated by a single expert it can be criticized. In addition, when working on establishing reference data sets to carry out quality assessments, guidelines should be available that are known to experts, so that common criteria are agreed upon and personal biases are avoided. None of this is indicated in this section.
It is also not indicated if a sampling is carried out or of what type and on what statistical basis the sample size is justified.
From my point of view, this is the most critical aspect of the paper.
In general, I consider that the methodology section is correct, but it has some problems:
1st It is too long,
2nd some formulas are wrong (eg the subscripts of 4).
3rd There are different styles for subscripts and superscripts, this introduces confusion.
4th In general, many formulas have poor quality and are difficult to read.
The aspect variable is circular, the authors must clarify that they have worked with it as such, because otherwise, the calculation of an average value gives inappropriate results.
Figure 14 should be figure 11.
Figure 10.a should clearly indicate (highlight) the SUs displayed in Figure 10.b
Author Response
Dear Reviewers:
Re: Manuscript reference No. remotesensing-1711787
Thank you for your comments concerning our manuscript. I am very surprised that you read my article so carefully. Those comments are all valuable and very helpful for revising and improving our paper, as well as the important guiding significance to our research. We have studied comments carefully and have made correction which we hope to meet with approval.
Please find attached a revised version of our manuscript” A new method for the assessment and local multi-scale optimization of the automatic extraction of slope units in complex terrains’’, which we would like to resubmit for publication as an article. Revised portions are marked in red in the paper. The main corrections in the paper and the responds to the reviewer’s comment are as flowing:
1 The preparation of the reference data requires more detail. It is not indicated how many technicians have outlined the SUs. It is important that they are experts and that this circumstance is supported by evidence. If the reference has been generated by a single expert it can be criticized. In addition, when working on establishing reference data sets to carry out quality assessments, guidelines should be available that are known to experts, so that common criteria are agreed upon and personal biases are avoided. None of this is indicated in this section. It is also not indicated if a sampling is carried out or of what type and on what statistical basis the sample size is justified. From my point of view, this is the most critical aspect of the paper.
Response: One of the limitations inherent to the wide application of slope units are the controversy in the definition of slope unit, leading to that the methods used for extracting slope units do produce not similar slope units and even with different extraction performance. SUs can be delineated automatically using a digital elevation model (DEM) or image segmentation approaches. For both cases, the results of SUs are geomorphological subdivision of the terrain into mapping units bounded by drainage and divide lines, which is consensus of geomorphologist or a hydrologist (Carrara et al. 1991; Guzzetti et al., 2006; Evans et al,2008; Wang et al,2018; Huang et al,2021). The SUs were recognized as a single slope, a combination of adjacent slopes, or a small catchment, which is easily delineated in the field and in the three-dimensional terrain representation (Guzzetti et al., 2006; Evans et al,2008). To further clarity the guidelines of reference SUs, we follow the suggestion of Guzzetti et al., (2006), Evans et al., (2012) and Alvioli, et al. (2016,2018); Firstly, the reference SUs are delineated as the catchment of drainage lines and divide lines generated by GIS-based hydrological analysis. Secondly, maximizing the homogeneity within slope units and the external heterogeneity between slope units by considering the slope aspect as a measure of homogeneity. The criterion for being manually delineated as reference SUs through visual interpretation. Only those complete mountain triangle surface with clear uniform orientation were selected. To make the make the samples representative, the strategy of collecting those SUs of straight and well-defined aspect is traversing the whole research region rather than random. We have added a complete subsection to introduce the guidelines and process of obtaining reference SUs (Line132-147):
“2.3 Reference SUs
In the literature, the SUs subdivisions are obtained considering hydro–geomorphological properties of the landscape by selecting areas delimited by drainage and divide lines [1-2,4,6], corresponding to the single slope, combination of adjacent slopes, or small catchment in nature [6-8]. The maximizing continuity of aspect within the SU is taken as boundary that is distinguished from its surroundings SUs [19,25,28]. To make the samples representative, the strategy of reference SUs collection is to traverse the whole research region rather than random. There is no restriction on the size of the reference SUs, but there should have straight and well-defined aspect and show clear discontinuities with adjacent units [28]. The criterion for being manually delineated as reference SUs through visual interpretation was shown in Figure 2b. Only those complete mountain triangle surface with clear uniform orientation were selected. The amendment and elimination were executed by using an unmanned air vehicle (DJI phantom 4pro, flight altitude 500 m) in field investigations (Figure 2c). A total of 1256 undisputed samples were obtained as the reference SUs for the accuracy evaluation of automatic SUs in the later stage.”
Relevant references are as follows:
(1) Carrara, A., Cardinali, M., Detti, R., Guzzetti, F., Pasqui, V. and Reichenbach, P., 1991. GIS techniques and statistical models in evaluating landslide hazard. Earth Surface Processes and Landforms, 16(5), pp.427-445. https://doi.org/10.1002/esp.3290160505
(2) Guzzetti, F., Reichenbach, P., Ardizzone, F., Cardinali, M. and Galli, M., 2006. Estimating the quality of landslide sus-ceptibility models. Geomorphology, 81(1-2), pp.166-184. https://10.1016/j.geomorph.2006.04.007
(3) Minár, J. and Evans, I., 2008. Elementary forms for land surface segmentation: The theoretical basis of terrain analysis and geomorphological mapping. Geomorphology, 95(3-4), pp.236-259. https://doi.org/10.1016/j.geomorph.2007.06.003
(4) Alvioli, M., Marchesini, I., Reichenbach, P., Rossi, M., Ardizzone, F., Fiorucci, F. and Guzzetti, F., 2016. Automatic de-lineation of geomorphological slope units with r.slopeunits v1.0 and their optimization for landslide susceptibility modeling. Geoscientific Model Development, 9(11), pp.3975-3991. https://doi.org/10.5194/gmd-9-3975-2016
(5) Alvioli, M., Guzzetti, F. and Marchesini, I., 2020. Parameter-free delineation of slope units and terrain subdivision of Italy. Geomorphology, 358, p.107124. https://doi.org/10.1016/j.geomorph.2020.107124
(6) Wang, K., Zhang, S., DelgadoTéllez, R. and Wei, F., 2018. A new slope unit extraction method for regional landslide analysis based on morphological image analysis. Bulletin of Engineering Geology and the Environment, 78(6), pp.4139-4151. https://doi.org/10.1007/s10064-018-1389-0
(7) Huang, F., Tao, S., Chang, Z., Huang, J., Fan, X., Jiang, S. and Li, W., 2021. Efficient and automatic extraction of slope units based on multi-scale segmentation method for landslide assessments. Landslides, 18(11), pp.3715-3731. https://doi.org/10.1007/s10346-021-01756-9
2 In general, I consider that the methodology section is correct, but it has some problems:
1st It is too long,
2nd some formulas are wrong (eg the subscripts of 4).
3rd There are different styles for subscripts and superscripts, this introduces confusion.
4th In general, many formulas have poor quality and are difficult to read.
The aspect variable is circular, the authors must clarify that they have worked with it as such, because otherwise, the calculation of an average value gives inappropriate results.
Response: Given the challenges in the determination of the optimal scale and refining of unrealistic SUs, this study proposes an iterative process to achieve SUs more consistent with the terrain. The indexes of OCE, GS, LS and LSC were proposed in sequence. To make new method clear and the formula more concise, the unnecessary intermediate steps have been removed, and the formula were relabeled in pithier manner. Each formula was carefully checked to avoid ambiguity and errors. The styles for subscripts and superscripts of the new formula were unified. The superscripts represented the categories of the variable, while the subscripts denoted sequence of the variable. The main changes are concentrated in 3.3 section shown below (Line242-268):
The aspect map contains values in degrees, and the average values cannot be taken straightforwardly, otherwise, the calculation of an average value gives inappropriate results. Care must be taken in expressing angles and arcs consistently in degrees or radians. The following definitions from Alvioli, et al. (2016) were considered in this paper for calculating the average aspect.
To avoid lengthy non-core steps, the process is simplified as follows:
Line222-224:
“Note that the angle is needed to convert to radians. The average values and the difference should be intended vectorially following the equation in Alvioli et al. [25].”
3 Figure 14 should be figure 11.
Response: The sequence number of all figures have been rechecked, and Figure 14 has been corrected to Figure 11. Line561-563:
4 Figure 11. Statistical result of spatial distribution characteristics from the SUs in I, II, III and IV topographic patterns.
Figure 10.a should clearly indicate (highlight) the SUs displayed in Figure 10.b
Response: The SUs displayed in Figure 10.b have been clearly indicated with same yellow line in Figure 10.a. Line482-484
All of the comments have been carefully considered, examined and have been revised one by one. Special thanks to you for your good comments.
Yours sincerely
Dr. Wei Jinbing
Sichuan University
Author Response File: Author Response.docx
Reviewer 2 Report
This manuscript tries to establish a new method for the assessment and local multi-scale optimization of the automatic extraction of slope units in complex terrains. This manuscript had good results in data analysis and description. I just have some points to raise a suggestion.
- The information or scientific point presented could be more sufficient. Related applications for others could be also illustrated. After all, most of the issues related to slope units are used for landslide or geomorphic analysis, and the readers who are interested in this topic are limited. The manuscripts appear to be difficult to relate to new discoveries or methods. Also, in Introduction, the authors might emphasize the scientific points and contributions of this article so that readers could understand.
- Some abbreviated nouns like “MSSus” are used in the manuscript, and authors are advised to add more explanations.
- Some professional terms are a bit confusing to use, and it is recommended to integrate and use most people's usage. Example: “landslide sensitivity modeling (LSM)”, “landslide stability analysis”, “landslide susceptibility models”, “landslide susceptibility assessments”, “landslide assessment”.
- Some figures are not appropriate. Part (C) in Figure 2 shows the same region, it is recommended to mark which part of the same with (A) or (B). The black box line on the left side of Figure 7 represents OSSUs, but the description on the right side is a green line segment. It is recommended to unify or correct it.
Author Response
Dear Reviewers:
Re: Manuscript reference No. remotesensing-1711787
Thank you for your letter and for the reviewers’ comments concerning our manuscript. Those comments are all valuable and very helpful for revising and improving our paper, as well as the important guiding significance to our research. We have studied each comment carefully and have made all correction which we hope to meet with approval.
Please find attached a revised version of our manuscript” A new method for the assessment and local multi-scale optimization of the automatic extraction of slope units in complex terrains’’, which we would like to resubmit for publication as an article. Revised portions are marked in red in the paper. Replies were sent to each peer reviewer regarding their comments and suggestion. The main corrections in the paper and the responds are as flowing:
1 The information or scientific point presented could be more sufficient. Related applications for others could be also illustrated. After all, most of the issues related to slope units are used for landslide or geomorphic analysis, and the readers who are interested in this topic are limited. The manuscripts appear to be difficult to relate to new discoveries or methods. Also, in Introduction, the authors might emphasize the scientific points and contributions of this article so that readers could understand.
Response:
we added more information on emphasizing the scientific points and contributions of this article in Introduction (Line99-103):
“This paper aims at demonstrating the optimal scale criterion for delineating SUs, and effective integration of multiple spatial features into the process of the identification and refinements of undesirable SUs. The new multi-scale SUs complementary evaluation and optimization framework can provide a practical demonstration for similar post-processing of terrain classification.”
We expected the obtained SUs of complex terrains to have several potential applications. For landslide studies, the primary application is the production of statistically based LSM and related terrain zonation. We have added 3 related articles to enrich the introduction of related applications of slope units for others. After reading some recent studies on this topic, the main potential applications were summarized as followed to help improve the article (Line47-50):
“The terrain subdivisions of SUs also have other potential applications. It can be used as mapping units for hydrological and erosion modelling, and geo-environmental, ecological, forestry, agriculture studies that require the identification of homogeneous terrain domains [13-15].”
2 Some abbreviated nouns like “MSSus” are used in the manuscript, and authors are advised to add more explanations.
Response:
Comparing these undesirable SUs of OSSUs and USSUs, the MSSUs represent the correct and moderate-subdivided SUs. Specifically, the shapes of MSSUs are consistent with continuity and discontinuity of topographic relief, and the property of external heterogeneity and internal homogeneity both are moderate level rather than extreme value. Consequently, A brief explanation is given on the abbreviated nouns of MSSUs. The additions are as follows (Line 283-286):
“Through the dual inspection of OCE and GS value, the selected optimal scale can make the majority of the SUs consistent with the subdivision of the terrain and they are defined as moderate-subdivided SUs (MSSUs).”
3 Some professional terms are a bit confusing to use, and it is recommended to integrate and use most people's usage. Example: “landslide sensitivity modeling (LSM)”, “landslide stability analysis”, “landslide susceptibility models”, “landslide susceptibility assessments”, “landslide assessment”.
Response:
I am very surprised that you read my article so carefully. The comments are all valuable and very helpful for revising and improving our paper. The professional terms: “landslide sensitivity modeling (LSM)”, “landslide stability analysis”, “landslide susceptibility models”, “landslide susceptibility assessments”, “landslide assessment”, which have been used uniformly as “landslide sensitivity modeling” and its abbreviation “LSM”. The additions are as follows:
(Line 45-47) “SUs have the advantages of reflecting the physical relationships between landslides and the environmental information of homogeneous terrain [10], and thus are especially suitable for landslide sensitivity modeling (LSM) [11-12].”
(Line 54-55) “Despite the advantages of SUs over other map units, only a small part (5.1 %) of LSM used SUs in the past 30 years of literature retrieval [18-19].”
(Line 490-493) “Compared to previous research [20,21,25], the OCE value replaced the LSM prediction accuracy, which is another evaluation criterion for the optimal scale. The LSM depends on the availability of detailed landslide data and the appropriateness of machine learning models, which are often uncertain.”
(Line 566-567) “A set of appropriate SUs that parting a complex landscape into reasonable terrain subdivisions is crucial for LSM.”
(Line 578-579) “With improved organization and distribution of landslides data, refining SUs has great application potential in LSM of complex basins.”
Moreover, other use of professional terms also been carefully examined. The word about “the Tibetan Plateau”, “the Qinghai-Tibet Plateau” have been uniformly replaced by the most people's usage of “the Tibetan Plateau”. Similarly, the word of “slope units”, “units”, and “SUs” have been uniformly replaced as “SUs”.
4 Some figures are not appropriate. Part (C) in Figure 2 shows the same region, it is recommended to mark which part of the same with (A) or (B). The black box line on the left side of Figure 7 represents OSSUs, but the description on the right side is a green line segment. It is recommended to unify or correct it.
Response:
The Figure 2(C) has been polygon-cropped strictly according to the region shown in Figure 2(A) and (B) and the redundant area has been eliminated. In addition, the real shapes of reference SUs were depicted in the photo. The Figure 2 are as follows (Line147).
The color of the line represents OSSUs, USSUs and MSSUs in the black box and right side of Figure 7 has been unified. The details are shown in the figure below (Line420).
All of the comments have been carefully considered, examined and have been revised one by one. Special thanks to you for your good comments.
Yours sincerely
Dr. Wei Jinbing
Sichuan University
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
In general, the document has improved with the revision.
However, since reference data is used, it should be clarified:
1st) how many people have participated and their experience profiles on this topic. An explanatory text on this aspect must be included in the paper .
2nd) a guide and criteria must have been used to establish the "undisputed" cases. This information must be included, either in the form of a reference to a external document, or by including a text, table or diagram in the document.
