Multi-Scale Ground Deformation Analysis and Investigation of Driver Factors Based on Remote Sensing Data: A Case Study of Zhuhai City

Round 1

Reviewer 1 Report

The paper analyzes the impact of multiple parameters on ground subsidence in the Zhuhai region using the MGWR method and determines its underlying drivers and effects, which has certain research significance. However, there are several issues in the paper. Please consider the following points in the attachment.

Comments for author File: Comments.pdf

Please modify and improve the language and expressions throughout the paper to ensure smoother readability.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors propose a methodology that seeks to correlate observed InSAR ground deformation in Zhuhai (China) with external parameters (data). For this purpose, the authors use regression techniques and compare their residuals to assess which one performs better. Finally, statistical indicators were used to highlight which driving factors of ground deformation are most likely to be responsible for the observed phenomenon. 

If so, I think a novelty is being proposed by the authors. But it is really hard to follow and understand the manuscript. The method turns out to be explained not clearly and very confusing. 

I suggest to work well on the explanation of the method and especially to improve the results, among other things, the figures turn out to be almost all grainy and cannot be analyzed properly.

Then it is not clear what this statement in the abstract "but the intrinsic multiscale mechanisms of the mechanisms are not clear" means, again in the abstract and in the text "mm/a" is used but who is "a"? 

Also, the title of the article turns out to be a bit misleading, as multiscale leads one to think of multiple scales of resolution.

1. What is the main question addressed by the research?
The paper attempts to address the dependence of the InSAR mean ground displacement velocities of Zhuhai (China) with other remotely sensed data by adopting regression strategies, specifically OLS, GWS and MGWS. The degree of correlation between InSAR products and driving factors (externally collected data) is assessed by statistical metrics that essentially evaluate the fitting between observations and estimates.

2. Do you consider the topic original or relevant in the field? Does it address a specific gap in the field?
The topic is promising from a risk prevention perspective. But unfortunately it is not clearly explained, and in the next points I try to be as clear as possible.

3. What does it add to the subject area compared with other published material?
Unlike previous studies, here an attempt is made to address a plethora of external causes/processes by additionally using a regression technique that is non-stationary, i.e., spatially adaptive, whereby different regression estimates are obtained depending on the geographical context in which the technique is applied. This is certainly valuable, but it is also disadvantageous because there is a risk of model overfitting.

4. What specific improvements should the authors consider regarding the methodology? What further controls should be considered?
Regarding methodology, it is not quite clear how the authors discard OLS over GWS and MGWS. It is my understanding that they apply local regressions, considering all external driving forces, and then go on to evaluate the residuals and their autocorrelation by Moran's I test. By virtue of that outcome they determine that OLS is not preferred. If so, this part needs to be better specified. Also, after the most appropriate regression technique is identified, do the authors proceed by applying it for each individual external driving factor and then evaluate the degree of fitting? All these issues should explained better.

5. Are the conclusions consistent with the evidence and arguments presented and do they address the main question posed?
The discussion and conclusions should spend a few more words on the comments I made in the previous points, and I think they should also say a little more about the limitations of methodology (if any). Also, I don't think it's entirely correct to treat NDVI as an external physical factor; after all, it is an index.

6. Are the references appropriate?
Yes.

7. Please include any additional comments on the tables and figures.
In practice, all the figures are almost grainy and and make it really difficult to read the data in them; especially for figures ranging from 4 to 10.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Dear Authors,

This paper was interesting to read, especially for the section related with the use of statistical methods driven by remote sensing datasets.

Nevertheless, the study relevant lacks in the analysis of displacement data.

Since I’m not an expert in the geostatistical field, the following comments apply to the SAR monitoring section and the outcomes of this analysis.

Please find them below:

 

-  I suggest adding some relevant references regarding the use geomatics techniques for evaluating ground deformations:

Pepe, A., Calò, F., 2017. A review of interferometric synthetic aperture RADAR (InSAR) multi-track approaches for the retrieval of Earth’s Surface displacements. Appl. Sci. (Switzerland) 7 (12). https://doi.org/10.3390/app7121264.

Alberico Sonnessa, Annamaria di Lernia, Davide Oscar Nitti, Raffaele Nutricato, Eufemia Tarantino, Federica Cotecchia, “Integration of multi-sensor MTInSAR and ground-based geomatic data for the analysis of non-linear displacements affecting the urban area of Chieuti, Italy,” International Journal of Applied Earth Observation and Geoinformation, Volume 117, 2023, 103194, ISSN 1569-8432, https://doi.org/10.1016/j.jag.2023.103194.

David A. Schmidt1and Roland Bu rgmann.  Time-dependent land uplift and subsidencein the Santa Clara valley, California, from a large interferometric synthetic aperture radar dataset. JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 108, NO. B9, 2416, doi:10.1029/2002JB002267, 20031Now at Department of Geological Sciences, University of Oregon,Eugene, Oregon, USA.

Improve the image resolution. Values (coordinates, displacement rates, NDVI and so on) are not clearly readable.

Length units are not unique (e.g. miles in fig.4 and km in fig. 5).

L64: InSAR is a well consolidated ground observation method. Please modify this statement.

Par. 2.2.1. Why did you consider only one orbit (i.e. ascending) in your study? And what about the time span? Sentinel SAR data are available since 2015, so it would be possible to strengthen the analysis by increasing the considered period. The authors should better motivate their choices.

L157: Use range instead of distance direction.

Table 1. Use timespan instead of image time period and make it more readable (e.g. 8 Jan 2020 to…).

Par4.:

-  - What do you mean by mm/a. Does it stand for mm/year? It would be better to use mm/y.

-  - Since SAR can’t detect displacement/velocities under some mm, too much decimal digits are not significant. Please use one digit after the dot (e.g. 1.5 mm/y).

-- L312-314 In my opinion, this statement could lead to misunderstanding. A displacement rate between -30 mm/y and 30 mm/y can’t be considered indicative of a stable area. What the Authors can infer is that a uniform lowering rate (on the 97% of monitored points) could not be dangerous for structures because no differential displacements are detected, but this should be supported by field evidence (e.g absence of cracks on edifices).

-   Fig.4. The legend must be improved. The Authors should divide the velocity range into equal intervals (e.g. 5 mm/y steps) and maintain the choice in the whole paper.

More in general, the analysis on deformations (and consequently the following conclusions) should be strongly improved. The use of a single track could lead to wrong interpretations because the Authors are considering just one component of the real movement along the vertical direction. That’s why the detected uplift could be fictitious, and the subsidence rates not properly evaluated.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Please replace all the figures with clearer ones.

Reviewer 3 Report

The suggestions indicated in the previous round have been implemented.

The paper can be published in the present form