State fusion entropy for continuous and site-specific analysis of landslide stability changing regularities

Liu, Yong; Qin, Zhimeng; Hu, Baodan; Feng, Shuai

doi:https://doi.org/10.5194/nhess-18-1187-2018

Articles | Volume 18, issue 4

https://doi.org/10.5194/nhess-18-1187-2018

© Author(s) 2018. This work is distributed under
the Creative Commons Attribution 4.0 License.

Special issue:

Landslide early warning systems: monitoring systems, rainfall...

https://doi.org/10.5194/nhess-18-1187-2018

© Author(s) 2018. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 18, issue 4

Research article

|

19 Apr 2018

Research article |

| 19 Apr 2018

State fusion entropy for continuous and site-specific analysis of landslide stability changing regularities

Yong Liu, Zhimeng Qin, Baodan Hu, and Shuai Feng

Download

Final revised paper (published on 19 Apr 2018)
Preprint (discussion started on 01 Nov 2017)

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC1: 'Reviewer', Anonymous Referee #1, 23 Nov 2017
- AC1: 'responseTo RC1', Zhimeng Qin, 12 Dec 2017
- AC4: 'Revised paper', Zhimeng Qin, 12 Dec 2017
RC2: 'Referee comment', Anonymous Referee #2, 28 Nov 2017
- AC2: 'responseTo RC2', Zhimeng Qin, 12 Dec 2017
- AC5: 'Revised paper', Zhimeng Qin, 12 Dec 2017
RC3: 'State fusion entropy for real-time and site-specific analysis of landslide stability changing regularities', Anonymous Referee #3, 29 Nov 2017
- AC3: 'responseTo RC3', Zhimeng Qin, 12 Dec 2017
- AC6: 'Revised paper', Zhimeng Qin, 12 Dec 2017

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

ED: Reconsider after major revisions (further review by editor and referees) (03 Jan 2018) by Stefano Luigi Gariano

AR by Zhimeng Qin on behalf of the Authors (30 Jan 2018) Author's response Manuscript

ED: Referee Nomination & Report Request started (31 Jan 2018) by Stefano Luigi Gariano

RR by Anonymous Referee #1 (04 Feb 2018)

RR by Anonymous Referee #2 (06 Feb 2018)

Suggestions for revision or reasons for rejection

The reviewer appreciates the effort put by the authors in replying to the remarks and working on some of them. However, many of the doubts and issues raised by the reviewer remain unsolved. Therefore, the reviewer suggests that the manuscript should not be accepted for publication in the present form, but encourages resubmission after all the following points being solved or, at least, discussed appropriately.

1) A comparison of several methods to provide evidence that the proposed model does give a practical advantage over the existing methods for the “continuous and site-specific analysis of landslide stability changing regularities”. The reviewer expects to see a comparison with the methods mentioned by the authors (Saito’s, FEM, LEM) in at least one case study.

2) Application to early warning: the authors state in the abstract that “state fusion entropy may serve as a novel index for assessing landslide stability and landslide early warning”. The reviewer expects to see, in the revised manuscript, an example of application of the proposed method to the early warning of a landslide (a past case study is fine), showing a successful (and usable) prediction of failure.

3) Integration of multiple monitoring sources. The reviewer expects to see an example of performance of the model by taking displacement data from different locations as input (or at least a good discussion about this). Multiple monitoring, to the reader’s opinion, is already a diffused standard in landslide monitoring, and must be taken into account when proposing a “novel method”. The reviewer also expects a deeper discussion on the incorporation of different monitorable quantities (e.g. pore pressures, moisture, deep displacements, etc.).

4) Arbitrary choices. The reviewer expects to see in the revised manuscript an analysis of the effect of arbitrary choices on the model’s performance. For instance, how does the result changes if a different number of clusters is used?

5) Real-time analysis. The reviewer appreciates that the authors changed the title from “real-time” to “continuous”. However, by using monthly data, early warning in acceleration phase seems unlikely. Hence, the authors are suggested to eliminate all references to early warning. Alternatively, they should use daily or, better, hourly / minutely input to run the model. During the acceleration phase of a movement, the input frequency should be increased if a real-time (or a continuous) monitoring for early warning purposes is desired. Moreover, landslides moving at mm/day or mm/week rates are certainly monitorable with daily frequency without significant error, if proper monitoring is used (e.g. high-quality GPS measurements, inclinometer measurements; examples in the literature are available), so the reviewer does not understand why the authors insist on using monthly measurements. The authors could easily apply the model to a different case study, where better monitoring data series are available, to prove the good performance of their model.

6) Smoothed CSFE. I have severe doubts on the use of a smoothed CSFE in the work, as it seems to me that the authors erroneously used a 5-period average obtained from two past data, the current value and two future data. The future data are obviously unavailable during real-time monitoring, so the SCSFE they obtained simply makes no sense! I suggest the authors avoid any smoothing, as the period is chosen arbitrarily. Moreover, if monthly data are already very far from being a “real-time” monitoring, a 5-months average value (of the past data) is even farther. Without the smoothing, it becomes also clear from the figure the authors present that the CSFE is unable to predict the failure: in the case of the Xintan landslide (fig. 7), failure occurred at a value of CSFE which is lower than the historical maximum. If a warning threshold was set to correspond to the CSFE overcoming its past maximum, you would have failed to warn the authorities/population before failure occurred. Talking about Fig. 7, you did not include the displacement trend of the landslide, despite my explicit request. I believe this is fundamental (as done for fig. 8), to let the reader compare your statistics with the actual monitoring data.

Hide

RR by Anonymous Referee #4 (09 Feb 2018)

ED: Reconsider after major revisions (further review by editor and referees) (12 Feb 2018) by Stefano Luigi Gariano

AR by Zhimeng Qin on behalf of the Authors (21 Mar 2018) Author's response Manuscript

ED: Referee Nomination & Report Request started (21 Mar 2018) by Stefano Luigi Gariano

RR by Anonymous Referee #2 (22 Mar 2018)

ED: Publish subject to technical corrections (23 Mar 2018) by Stefano Luigi Gariano

Short summary

Stability analysis is very important for landslide early warning, especially the continuous stability, but there are some limitations in obtaining continuous stability by using traditional stability analysis methods. In view of this, we have further explored the displacement monitoring data in order to find new indicators and methods that can show landslide stability changing regularities.