Fusion of GNSS and Satellite Radar Interferometry: Determination of 3D Fine-Scale Map of Present-Day Surface Displacements in Italy as Expressions of Geodynamic Processes

Round 1

Reviewer 1 Report

Thanks for your replies. However, I think there are a few more places which need to be fixed:

Figure 2, top shows the GPS velocities for the whole area, while the interpolated components in the bottom only covers part of that area. The same question for the figures 3, 5, 7 and 8 in which they only show part of the velocities. I didn’t find anywhere in the text how they have selected or filtered the velocities and I think it is important to mention this in the text.

Figure 10 and 12: What is the thick black line? Missing in the legend or caption.

In Figure 11,GNSS velocities are shown , why not the combined velocities of GNSS and PS?

Author Response

See attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors have accepted my suggestions.

I'm still thinking that the part of discussion about the geophysical interpretation of the results is too long for this type of Journal, but this concerns the Editor's decision.

Author Response

Any comments.

Reviewer 3 Report

Authors have correctly accepted and corrected most of the comments from the previous review procedure. However, for the procedure used in the calculations it would be good to at least roughly estimate the mistake that was ignored with respect to the comment:

1)    In Chapter 6, lines 203-206, it is stated: „Displacements occurring along the north–south direction are almost parallel to the satellite orbit and therefore cannot be measured accurately, as their projections along the LOS are negligible for both ascending and descending orbits“. How much can this statement be considered true and whether it brings some error with it?

Author Response

See attachment.

Author Response File: Author Response.pdf

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

This version was improved but I still see many problems which needs to be corrected. This time, I give a chance and suggest major revision.

Here are my comments which might help for the re-submitted/revised version:  

Main one:

Figure 2: Again, I am concerned with this Velocity field: vectors in Italy generally show a consistent pattern whereas close by velocity vectors to the right side of the figure (to the east or north east of Italy) show different directions.  That is not only one station, many close by stations seem suffer from either the incorrect combination of velocity fields or incorrect processing or having different reference frames. I think because this velocity field is the main base for this study must to be correct for further combination with InSAR results.  

And more minor ones

-Although English of the paper is better in this version but I see  there are still problems in new edits. For example lines 13-17. Or in line 20, velocity rate is not correct, velocity is already a rate of displacement. Or in line 19, SAR products is not clear, better to write InSAR derived deformation maps. Or in line 37: better to change program to satellites. Line 56: has -> have. Line 127: like > similar to ...and more.

Line 20: We don’t know what is correct velocities. Always velocity estimations have uncertainties, and these are the uncertainties which decrease if authors mean that.

Figure 1: I don’t think a map with moho depth information is a good choice for this paper.  Authors just need to present a map of tectonic settings with tectonic structures, faults, ….No need to have moho depth which is not the focus in this paper.

Line 174: I haven’t seen the term roto-translation before. Do you mean Rotation and translations?

Equation 7: j is not clear if this is the coordinates or something else?

Figure 6:  This interpolation of velocities in north and east components are interesting. However, that can be done before Fusion  only for GPS velocities and then can be compared with the one after fusion (GPS+Insar). In this way, we can see how much the velocity field is improved and the novelty of the presented method in this paper.

Author Response

See document in attachment  

Author Response File: Author Response.docx

Reviewer 2 Report

Dear authors,

thank you for the answers and the corrections made to the previous version of your manuscript. I have also appreciate the improvement in the title. 

However, there are some questions that are not solved and require a further major revision of the manuscript.

concerning the of the two Italian GPS velocity solutions, the authors write that a roto-translation transformation with six parameters has been computed. Why a 7-parameters transformation is not applied? Why don’t you consider the scale factor? The Bernese and the GAMIT/GLOBK softwares deal the definition of the reference frame in a different way, so a scale factor could affect the results. 

How many common stations have been use in this roti-translation transformation?

I really think that the authors should present the statistics of the residuals of the roto-translation 

Why do you decide to use weekly solutions instead of daily solutions?

the description of the processing of the PSI is very poor. How the deformations along the LOS have been obtained? How the LOS time series have been processed?

Chapter 7, it’s too short! 5 text lines with respect a figure caption of 8 lines with a formula seem me not equilibrate. I think that the text of the chapter could be improved moving part of the information from the caption to the text. In this chapter it is also important explain why the total decorrelation between dataset is expected, as you have answer me in the previous review. 

Chapter 9. In Figure 5 the horizontal velocity field obtained by the combination of SAR and GNSS datasets are presented, but it’s not explained how it has been obtained. In fact, in the previous chapters the authors describe just the calibration process to remove the indetermination due to the SAR differential method. In chapter 6, the formula to obtain the horizontal velocities by those along the LOSs are presented, but the resulting components are the velocity in the east and in the vertical direction. So, at the end for me it’s not clear how the authors combine the horizontal GPS velocity vectors with the SAR east velocities. This part should be represent a key aim of the manuscript.

A map of the only-SAR velocity filed for the Italy is missing. 

In the abstract, I think there is a bit confusion in some definitions (line15-17), because at first the ITRS89 is a system and not a frame and the Line of Sight (LOS) is not a frame. I think that should be more correct to write that the “The above techniques provide displacements with respects different components of the ground point position” 

In chapters 9.1.1 and 9.1.2 are you discussing the only-SAR solutions? By the figure it looks like this. 

The GPS velocity solutions are framed in the ITRF2008 (by the cited papers), but in the manuscript you report the velocity with respect the ETRF89. I suppose that a reference system transformation has been performed. 

in the manuscript, the geological interpretation is too long for this type of Journal.

Other remarks: 

Line 15: I suggest “:” instead of “;”

Line 28: “local ground” —> ground 

Line 145: “elevation” —> “Elevation”

Line 167: check the data “784 continuos GPS stations”, because in the reference [36] I have read 379 cGPS + 80 eGPS stations

Line 178: put attention in using the terms frame and system: “European reference frame(ETRS89)” —> European reference frame (ETRF89)

Line 179: “at a95%”, check the space

Line 194: “by the SAR satellite” —> by the SAR technique 

Line 215: in the answers to Reviewer #3, the authors affirm that the calibration is performed by using all the PS around a circle of 1000 m from the GNSS site locations. I think that it’s important to report this  important information also in the text.

Author Response

See document in attachment

Author Response File: Author Response.docx

Reviewer 3 Report

Basic comments and objections:

Here are some suggestions for clarification and improvement:

1)    In Chapter 6, lines 203-206, it is stated: „Displacements occurring along the north–south direction are almost parallel to the satellite orbit and therefore cannot be measured accurately, as their projections along the LOS are negligible for both ascending and descending orbits“. How much can this statement be considered true and whether it brings some error with it?

2)    In the description of Figure 3 there is also a formula for Pearson's correlation. Would it be better to write this formula in the normal text of chapter 7 and number it as equation (7), and rename the present formula (7) to (8)?

3)    Chapter 10, at the end of the sentence written in lines 489-490, consider whether the sign":" is better than "." ?

4)    Looking at Figure 5 it can be seen that the authors had at their disposal a relatively small number of speed vectors in Slovenia, Croatia and Bosnia and Herzegovina. In addition, some of these vectors are quite strange, especially in the direction. Do you have any explanation for that?

Author Response

See document in attachment

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Thanks for your replies to my comments. I am afraid, I am still not happy with this version and some of your answers. Here are my comments:

-The title to me has problem: it consists of two separate sentences and doesn't look a proper title for this MS. I think the 1st part should be enough and representative for this MS.

-There is a similar work by the 1st author (ref. no 45, published in IEEE, Sept. 2018) which seems to me have used the same method for integration of GNSS and PS velocities and is for north of Italy. Good to mention that work in Introduction and say if this fusion method has already been used in that paper and   that method is applied in this new paper, or if they develop the method further in this paper. Also good to compare the new results in this paper with that work (ref. 45) in details using some new maps showing the comparison of vertical velocities estimated on the common area in both papers.

-Figure 8 is not well illustrated. The figures are not aligned and the font size for color scale is too small.

-I just noticed that Conclusions is written in a very general form. The conclusions should be related to the results and discussions of this work. For example, they have the rates of deformation, a new map of deformation, some features that  have been tectonically interesting and more that can be concluded. I would suggest rewriting this section.

-And still I don't think the Moho depth is important to show in that map (Figure 1). However, if authors insist on keeping that, I accept it.

Author Response

Thanks for your replies to my comments. I am afraid, I am still not happy with this version and some of your answers. Here are my comments:

-The title to me has problem: it consists of two separate sentences and doesn't look a proper title for this MS. I think the 1st part should be enough and representative for this MS.

Corrected.

-There is a similar work by the 1st author (ref. no 45, published in IEEE, Sept. 2018) which seems to me have used the same method for integration of GNSS and PS velocities and is for north of Italy. Good to mention that work in Introduction and say if this fusion method has already been used in that paper and   that method is applied in this new paper, or if they develop the method further in this paper. Also good to compare the new results in this paper with that work (ref. 45) in details using some new maps showing the comparison of vertical velocities estimated on the common area in both papers.

The cited work regards the vertical  surface movements in Po Plain of Northern Italy, focussing on the subsidence of UNESCO cities such as Venice and its lagoon, Ravenna, Ferrara and the historical cities of Bologna and Modena.
The scope of the present work instead is to identify detailed boundaries of velocity patterns and then to provide new information about the complex geodynamics involved in on the Italian peninsula.

Currently, the present study is one of the kind that contemplates a wide area such as the Italian peninsula studied by the combination of the GNSS and DInSAR for two decades of satellite observations.

-Figure 8 is not well illustrated. The figures are not aligned and the font size for color scale is too small.

Corrected. We increased the dimensions of the maps.

-I just noticed that Conclusions is written in a very general form. The conclusions should be related to the results and discussions of this work. For example, they have the rates of deformation, a new map of deformation, some features that  have been tectonically interesting and more that can be concluded. I would suggest rewriting this section.

We provided a wide discussion of tectonical implications in chapter 8 involving some area of interest. In the chapter Conclusions we provide a synthesis of our studies, its beneficts and the future applications.

-And still I don't think the Moho depth is important to show in that map (Figure 1). However, if authors insist on keeping that, I accept it.

We prefer to show this map.

Reviewer 2 Report

Dear authors,

thank you for the answer to my questions. I think that most of that should be reported in the text of the manuscript, to clarify doubts also to the readers. 

You didn't answer to one question about the combination method of the two solutions (GPS and GNSS).  In chapter 7, you have described how the PS velocity calibration has been computed, but how the GNSS and the Calibrated InSAR solutions have been combined together to obtain a an integrated solution (figure 8) is not described. Did you compute an average between the two solutions? Did you use any weight?

Check the numeration of the chapters: from chapter 7 you jump to chapter 9 (line 271), without any Chapter 8 (also in the previous version of the manuscript).

 Line 237: I really suggest, again, the authors to reduce the caption of figure 4  and put the formula and the explanation in the the normal text of the chapter.

The authors answered to my question "How many common stations have been use in this roto-transformation" reporting the number of common stations (112). I think that this information should be reported also in the text. 

Another important requirement is reporting in the manuscript the statistic of the transformation.  If the residuals are of the same values of uncertainties probably the transformation is not statistically significative and it could be avoid, is'n it? 

Line 30: Lithosphere -> crust

Line 287 and Line 301: frame —> system

Line 301: and Line  “;” —> “:”

Line 305: “Right - Vertical component.” —> “Right - Vertical component:”

Line 229: 1000 m or 100 m, as reported at Line 233?

I think that these questions have to be fixed before publishing the paper.

Author Response

Dear authors,

thank you for the answer to my questions. I think that most of that should be reported in the text of the manuscript, to clarify doubts also to the readers. 

You didn't answer to one question about the combination method of the two solutions (GPS and GNSS).  In chapter 7, you have described how the PS velocity calibration has been computed, but how the GNSS and the Calibrated InSAR solutions have been combined together to obtain a an integrated solution (figure 8) is not described. Did you compute an average between the two solutions? Did you use any weight?

Inserted. Lines 268-270

Check the numeration of the chapters: from chapter 7 you jump to chapter 9 (line 271), without any Chapter 8 (also in the previous version of the manuscript).

Corrected.

Line 237: I really suggest, again, the authors to reduce the caption of figure 4  and put the formula and the explanation in the normal text of the chapter.

Corrected.

The authors answered to my question "How many common stations have been use in this roto-transformation" reporting the number of common stations (112). I think that this information should be reported also in the text. 

Corrected.

Another important requirement is reporting in the manuscript the statistic of the transformation.  If the residuals are of the same values of uncertainties probably the transformation is not statistically significative and it could be avoid, is'n it? 

Absolutely right. Transformation in this case should be avoid. The reason is that we followed the the recommendations of the EPN (European Permanent Network) http://www.epncb.oma.be/_documentation/guidelines/guidelines_analysis_centres.pdf

Line 30:

Lithosphere -> crust

Corrected.

Line 287 and Line 301: frame —> system

Corrected.

Line 301: and Line  “;” —> “:”

Corrected.

Line 305: “Right - Vertical component.” —> “Right - Vertical component:”

Corrected.

Line 229: 1000 m or 100 m, as reported at Line 233?

1000 m. Corrected..

I think that these questions have to be fixed before publishing the paper.