Analysis of Slope Failure Behaviour Based on Real-Time Measurement Using the x–MR Method

Round 1

Reviewer 1 Report

This manuscript presents a field experiment to measure slope failure. I believe that the experimental design and results have the potential to add knowledge to the field and should be published, but the manuscript requires major improvements before publishing.

Line 67. Clarify which industry.

Lines 68-75. Yoo [27] and Yoo [28] are referenced as important justifications for the control chart methods for slope failure, but these two references are not readily available in the published literature. The authors may need to expand on and summarize the key findings of these papers.

Line 80-81. This is unclear as written, and it is probably a very important point to set up the main objective of the paper.

Lines 82-85. Split this into two sentences for clarity.

Line 90. Provide references to back up this statement.

Section 2 or 3. The authors need to clarify some critical details of the control charts and real input data. What are your input datasets to create PDFs for mu and sigma? Does it start with historical or spin-up data? Or does it start with an n of 1 and then grow? Or is it based on moving average of 10 or 30 samples or more? Are mu and sigma updated in real time as new data come in? I think Fig 10-14 may have data to answer some of these questions, but further interpretation and discussion are necessary. What happens if you change the analysis interval (K)? / moving window? How do you select K?

Line 136. Variables in the table need full explanations

Line 140. Need references and further elaboration of the free fall method.

Figs 10-14. A critical interpretation of Figs 10-14 is lacking. For example, what are the implications of sample sizes of 10 vs 30?

Section 3-4. A clearer description between “case” and “step” is needed.

Line 170. It is not clear why “failure” in case 1 is considered at 49.5 min, instead of 38 min. Similar comment for case 2.

Line 246 / Table 2. Should displacement be in mm over some unit of time?

Line 255. What is the basis for this range of evacuation times?

Line 272. How does K=3 compare to other similar studies?

Line 276. The discussion in this study does not focus on disaster prevention but on disaster preparedness and early warning.

Line 268. There is no case 4.

Line 324. Reference 15 is incomplete.

Author Response

Dear Reviewer, 

Thank you for your time in reviewing our manuscript.

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

The manuscript under review reads well. The x-MR seems a promising technique for slope failure signal detection and better for early warning. The authors provided a good demonstration example as of slope failure monitoring and data analysis. The manuscript may be improved in the following aspects:

Line 113, there is an error in defining the UCL of the control limit range.

Line 117, analysis interval – K value has not been clearly explained. Is it simply the number of data within moving range analysis? What is the effect of K to the results if any? As the later part involves different K scenarios, more details of this parameter are needed.

Line 164: a brief description of what are the three cases should be provided before Figure 8.

Figures 10 and afterwards, what is the intention to introduce inverse displacement, which is a bit confusing to this reviewer. What is the weakness and/or advantages? How does it linked to the other subfigures for data interpretation. Lines 198 to 201 are not sufficient for justification.

All figures indicate the results are quite sensitive to the K-value. The suggested K=3 is not convincing to this reviewer. The authors should provide in-depth discussion on how to select the model parameter, and in caution to give any practice recommendations.

Line 267-268, here you mentioned cases 1 to 4, not consistent with the previously mentioned three cases. Please check.

Author Response

Dear Reviewer, 

Thank you for your precious time in reviewing our manuscript.

 Please see the attachment. 

Sincerely yours, 

Author Response File: Author Response.pdf

Reviewer 3 Report

The work has important objectives, the method applied is rigorously illustrated and the application of a statistical method to tests carried out on a scale model represents an excellent experimental basis.
There are no substantive comments on the manuscript but the part of the more extensive conclusions on a protocol of predictive techniques to be adopted for the failure of slopes could be extended.
In fact, the conclusions foresee an increase in the experimentation in order to identify a collapse phase that can pre-alarm with a longer time the population possibly involved.

Very useful is the diffusion of the data set used, which can be compared by other research groups under similar conditions but in different contexts.

Round 2

Reviewer 2 Report

Can be accepted for publication.