CAPS: A New Method for the Identification of Different Surface Displacements in Landslide and Subsidence Environments through Correlation Analysis on Persistent Scatterers Time-Series from PSI

Round 1

Reviewer 1 Report

This study applies a method proposed by the authors, namely CAPS, to cluster PSInSAR deformation time-series over two landslide areas and two subsidence environments. The method was first presented by the authors in reference [15], using one of the landslide areas which is also discussed in the paper under review. Since the method is not new, the main interest of the paper lies in the application of the method to these new test-cases.

Although the topic is relevant and the approach is technically interesting, I find the scientific soundness of the experiments which are described to be low. The study lacks a hypothesis. In the landslide and/or subsidence scenarios what properties are the authours interested in mapping, and what would be their signatures in terms of radar LoS deformation time-series? Would the proposed analysis have the required sensitivity to distinguish these signatures (and what would the requirements be, given that the viewing geometry can have huge impact at least for landslide applications)? Can the proposed analysis be done in an objective way, without knowing the result beforehand?

The user-defined correlation threshold determines the number and size of the output clusters, which varies significantly depending on the user selection. The authors themselves state that the appropriate threshold cannot be set a priori. However this leads to two possible outcomes: either the user ends up selecting, by trial and error, a threshold which confirms prior knowledge, as in the Santo Stefano d’Aveto landslide case, or the user has no objective way to select the threshold and interpret the results, which is the case for the other 3 test-sites described in the paper. Concerning the Arzeno landslides, the authours comment that the results “highlight an interesting subdivision of the landslide, important, for instance, to plan future in situ analysis”. In the Venice case it is stated “an in-depth study could highlight the causes that led to such differential deformations”, and in the Rome case “This subdivision could help to better understand the dynamics of the area and, for instance, of the aquifer”.

If the authors could prove a relation between the PS clusters their method identifies and any geophysical property of interest, e.g. related to landslides or subsidence, this would be a very interesting scientific result. To state however that the somewhat arbitrarily identified PS clusters might be related to some property to be identified in the future is too weak a result for a scientific publication.

Author Response

Dear Reviewer,

Please find attached the pdf file with our answers to your comments and suggestions.

Best Ragards

Author Response File: Author Response.pdf

Reviewer 2 Report

Dear authors,

thank you for an interesting article. It regards a new method to detect the families of PSs based on the similarity of their time series.

The article is well written, clear, and I have only few minor notes:

- table 2: I would appreciate to see also units for each parameter

- line 460: should be 4.4 Rome instead of 4.3 Venice

- line 462: analysis -> analyses

- line 466: missing comma in the brackets

- line 550, 554: 90 -> 0.90

Author Response

Dear Reviewer,

Please find attached the pdf file with our answers to your comments and suggestions.

Best Ragards

Author Response File: Author Response.pdf

Reviewer 3 Report

Please consistence in using family or Family in the manuscript. For PS time series correlation analysis, how many data points are the minimal requirement? Since the temporal data reflects the total ground movements that is not just landslide or subsidence, what is the effect then? For various demo cases, please make a comparison table to reflect the similarity and differences. Some of your sentences are either too long or that I cannot understand is high lighted within attached pdf file.

Comments for author File: Comments.pdf

Author Response

Dear Reviewer,

Please find attached the pdf file with our answers to your comments and suggestions.

Best Ragards

Author Response File: Author Response.pdf

Reviewer 4 Report

The authors are trying to add value to Persistent Scatterer Interferometry implementation on landslides and subsidences by extending its use to different geological situations and at different scales in order to recognize both local and regional differential deformations, using a method called Correlation Analysis on Persistent Scatterers (CAPS).

A first comment reading carefully the article is that the paper does fit the scope of Remote Sensing, and it could be of interest to the readership of the journal. Furthermore, the data and the results are well presented and the authors do answer the questions they set out. However, there are some points that should be written more clearly. Therefore, I recommend minor revision. I explain my points in more detail below. Thus, to make this paper publishable, the authors need to respond to the following remarks.

Minor comments:

Line 127: What does “Ne” mean?

Line 180: The title of subunit “2.4 Rome” must be transferred to the next page.

Line 460: Please change 4.3 Venice title with the appropriate one (4.4. Rome)

Lines 500, 546, 561: There are longer spaces among the words

Author Response

Dear Reviewer,

Please find attached the pdf file with our answers to your comments and suggestions.

Best Ragards

Author Response File: Author Response.pdf