Towards Operational Dam Monitoring with PS-InSAR and Electronic Corner Reflectors^†

Round 1

Reviewer 1 Report (Previous Reviewer 1)

Comments and Suggestions for Authors

Dear Authors,

The corrections made to the text were sufficient, except for the figures that require coordinates for better localization. Moreover, the "Mann-Kendall" statistical test was conducted, but no reference or explanation was included in the text about how the test is performed and how it should be analyzed. These two points need to be reviewed.

Sincerely,

Reviewer.

Author Response

Comments 1:

Dear Authors,

The corrections made to the text were sufficient, except for the figures that require coordinates for better localization. Moreover, the "Mann-Kendall" statistical test was conducted, but no reference or explanation was included in the text about how the test is performed and how it should be analyzed. These two points need to be reviewed.

Sincerely,

Reviewer.

Answer 1: Thank you for your comment. We added coordinates to the figures. The Mann-Kendall test was performed and some information regarding this issue were added in lines 599-608 to make it clearer.

 

Reviewer 2 Report (Previous Reviewer 3)

Comments and Suggestions for Authors

The revised manuscript shows significant improvement compared to the previous version. However, I still have the following concerns:

1. Lines 738-740: Neither Figure 9 nor the accompanying text specifies the exact timeline for ECR installation. Additionally, Figure 9 remains unchanged from the previous version, making it unclear whether it includes surface scattering characteristics before ECR installation. This uncertainty challenges the conclusion in Lines 738-740.

2. The abstract mentions that the ADI of ECR is between 0.1-0.4, comparable to a regular CR. However, the main text lacks data on CR to support this claim.

Author Response

Comments 1: Lines 738-740: Neither Figure 9 nor the accompanying text specifies the exact timeline for ECR installation. Additionally, Figure 9 remains unchanged from the previous version, making it unclear whether it includes surface scattering characteristics before ECR installation. This uncertainty challenges the conclusion in Lines 738-740.

Answer 1: Thank you for your comment. We have added some additional information after line 510-516 to make the timeline clearer and explain Figure 9 better.

 

Comments 2: The abstract mentions that the ADI of ECR is between 0.1-0.4, comparable to a regular CR. However, the main text lacks data on CR to support this claim.

Answer 2: Thank you for your comment. We have some references in the Discussion section. I made some little changes to the text in line 70/704.

Reviewer 3 Report (Previous Reviewer 4)

Comments and Suggestions for Authors

• Terminology. We start with Differential InSAR (DInSAR) to quantify movement. Then, Persistent Scatterer InSAR (PSInSAR) was developed as an advanced form of DInSAR. It is characterized by pixels with high signal to noise ratio, referred to as PS. This is my understanding.
• Is DInSAR-PSI term, used in your title and on line 20, commonly used in other scientific publications? Please verify and modify if needed.
• Figure 1 - the bottom box indicates "Signal Stability Assessment" as one of your tasks. Can you formulate some conclusion(s) about it? In the past there has been concern about active transponders not being phase stable over long time periods. Your findings in this area would be valuable.
• Figure 9 - what do these findings translate into when it comes to estimated accuracy of measurement? Can you put a number on it, for example +/-5mm?
• Is it proposed to use this setup for alarm purposes? If so, how would it be implemented to prevent false alarm and provide timely detection?
• How long can the transponder operate on battery power alone?
• If more than one transponders is used on site, what is the minimum separation distance required between them (if any)?

Author Response

Comments 1: Terminology. We start with Differential InSAR (DInSAR) to quantify movement. Then, Persistent Scatterer InSAR (PSInSAR) was developed as an advanced form of DInSAR. It is characterized by pixels with high signal to noise ratio, referred to as PS. This is my understanding.

Answer 1: Thank you for the comment. We have unified the terminology, especially in the introduction in line 60/61.

 

Comments 2: Is DInSAR-PSI term, used in your title and on line 20, commonly used in other scientific publications? Please verify and modify if needed.

Answer 2: Thank you for the comment. We have adapted it to PSInSAR, since this is more commonly used.

Comments 3: Figure 1 - the bottom box indicates "Signal Stability Assessment" as one of your tasks. Can you formulate some conclusion(s) about it? In the past there has been concern about active transponders not being phase stable over long time periods. Your findings in this area would be valuable.

Answer 3: Thank you for the comment. We have done this in the Discussion section. We have added some additional information to clarify the conclusion (lines 737-739).

Comments 4: Figure 9 - what do these findings translate into when it comes to estimated accuracy of measurement? Can you put a number on it, for example +/-5mm?

Answer 4: Thank you for your comment. No, unfortunately they don’t. These are certainly only backscatter values to estimate a first phase stability.

 

Comments 5: Is it proposed to use this setup for alarm purposes? If so, how would it be implemented to prevent false alarm and provide timely detection?

Answer 5: No, it is not intended to be used as a alarm system. If your question refers to the stability analysis, It was carried out in order to assess the possibility of a PSI analysis and to provide information about stability. If anomalies are recognizable despite functioning devices, this is of course an indication and to a certain extent an “alarm” that a PSI analysis is not feasible.

 

Comments 6: How long can the transponder operate on battery power alone?

Answer 6: Depending on the outside temperature for up to 2/3 weeks. We added this information in the text in lines 91-93.

Comments 7: If more than one transponders is used on site, what is the minimum separation distance required between them (if any)?

Answer 7: Thank you for your question. There is no recommended minimum distance. They are to be treated like CR in this respect, but the occurrence of side lobes must be taken into account. Therefore, 100 - 200 meters are recommended. We have added this information in line 746-748.

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

Comments and Suggestions for Authors

Dear Authors,

The problem to be studied is well defined in the article, which highlights the viability of using ECRs to improve monitoring capabilities in dam infrastructures using the PSInSAR technique. However, some points need to be improved, as per the comments presented below:

1 – The article needs to review the numbering of the document sections, since some items are numbered out of order. For example, item “4.1. CR Index” comes right after item “3. Results”.

2 – In Figure 2, the maps presented should show coordinates and a graphic scale, so that readers can more easily locate what is being represented. In addition, red dots could be inserted on map (a) to highlight the position of each reservoir.

3 – In figures 5, 6, 7 and 8, coordinates should be inserted to facilitate the location of each represented area. In addition, the captions for figures 5, 6 and 7 are very small and this makes them difficult to read.

4 – In Figure 9, the graphs should be presented vertically and not horizontally, as is shown, so that they can be enlarged and thus make it easier to read what is being presented in each graph.

5 – Which “standard deviation” values ​​should be considered as low, good and excellent? The authors present results based on Table 2 that lack an explanation of how the “standard deviation” should be evaluated so that such classification can be made objectively. Was any significance test performed for the results achieved? If so, shouldn’t the test results be presented?

6 – In item “4. Discussion”, was any statistical analysis performed to confirm what is presented about the validation of the “CR Index”? Shouldn’t a trend test be performed to confirm what is stated in lines 570, 571 and 572: “This trend indicates that areas with optimal geometric and land-use characteristics, as captured by the CR Index, are associated with a high density of PS points.”?

7 – Still in item “4. Discussion” it is said that “The observed correlation between high CR Index values ​​and increased PS densities validates the underlying methodology, which combines geometric factors (captured in the R Index) with land-use data (represented by the Land-Use Index, LUI)”. Was the correlation between the variables analyzed calculated in such a way as to prove what was observed? If not, could the authors make a statistical evaluation and include the result in the article?

Author Response

Comments 1: The article needs to review the numbering of the document sections, since some items are numbered out of order. For example, item “4.1. CR Index” comes right after item “3. Results”.
Response 1: Thank you for pointing this out. The numbering is corrected. This change can be found at all corresponding headlines (line 427).

Comments 2: In Figure 2, the maps presented should show coordinates and a graphic scale, so that readers can more easily locate what is being represented. In addition, red dots could be inserted on map (a) to highlight the position of each reservoir.
Response 2: We highlighted the positions of the reservoirs with red arrows, but did not add coordinates to the jurisdiction of the Ruhrverband, as this is an official illustration of the Ruhrverband. We think the small map of Germany is enough for orientation. This change can be found in Figure 2 and below (line 225/226).

Comments 3: In figures 5, 6, 7 and 8, coordinates should be inserted to facilitate the location of each represented area. In addition, the captions for figures 5, 6 and 7 are very small and this makes them difficult to read.
Response 3: Thank you for pointing this out. We have enlarged the captions for the figures. These changes can be found in figures 5, 6, 7 and 8 on pages 6, 7, 8 and 9 respectively. However, we have decided not to add any extra coordinates, as Figure 2 provides a sufficient overview, and additional coordinates would make the figures smaller.

Comments 4: In Figure 9, the graphs should be presented vertically and not horizontally, as is shown, so that they can be enlarged and thus make it easier to read what is being presented in each graph.
Response 4: This is a good comment, it has been changed accordingly. This change can be found in Figure 9 (between line 527 and 528).

Comments 5: Which “standard deviation” values ​​should be considered as low, good and excellent? The authors present results based on Table 2 that lack an explanation of how the “standard deviation” should be evaluated so that such classification can be made objectively. Was any significance test performed for the results achieved? If so, shouldn’t the test results be presented?
Response 5: Thank you, we have added a few details for better explanation. These changes can be found on in Section 4.2 (lines 676-689).

Comments 6: In item “4. Discussion”, was any statistical analysis performed to confirm what is presented about the validation of the “CR Index”? Shouldn’t a trend test be performed to confirm what is stated in lines 570, 571 and 572: “This trend indicates that areas with optimal geometric and land-use characteristics, as captured by the CR Index, are associated with a high density of PS points.”?
Response 6: Thank you for the commend. We have added some further details on this matter. These changes can be found in lines 591-598 of section 4. Discussion.

Comments 7: Still in item “4. Discussion” it is said that “The observed correlation between high CR Index values ​​and increased PS densities validates the underlying methodology, which combines geometric factors (captured in the R Index) with land-use data (represented by the Land-Use Index, LUI)”. Was the correlation between the variables analyzed calculated in such a way as to prove what was observed? If not, could the authors make a statistical evaluation and include the result in the article?
Response 7: Thank you for the comment. We think that the wording here is somewhat misleading. It refers to the validation described above. We have therefore slightly changed the wording here. The changes can be found in section 4. Discussion (lines 637/639).

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript presents a novel approach to increasing PS points, and the authors have conducted extensive work to demonstrate its effectiveness.

 

However, despite these advancements, the paper identifies several areas that necessitate further refinement.

 

1、The original phrase “due to their small size a practical” seems to be grammatically incomplete and fails to clearly convey the author's intended meaning.

 

2、Section 2.1 provides an overly detailed background on the PS-InSAR technique. While it includes valuable references, this technique is a common method for monitoring dam deformation and may not require such extensive coverage.

It is recommended that the author condense this section, focusing instead on the specific applications of PS-InSAR in dam monitoring and its advantages when combined with ECRs, to avoid lengthy descriptions overshadowing the main points of the article.

 

3、The article does not mention the specific technical parameters of the ECRs, which may lead to insufficient understanding of their design and performance by the readers. It is suggested that the author include details such as the size, weight, backscatter intensity (e.g., Sigma-0 values), power supply, and operational lifespan of the ECRs.

 

4、Although the article discusses the application of the CR Index, it does not provide a detailed explanation of its calculation method. This omission may hinder readers' understanding of the index's composition and applicability. It is recommended that the author add the formula for the CR Index or describe its calculation process.

 

5、The current article does not mention the deployment density of ECRs, which may prevent readers from evaluating the practical deployment requirements and costs of the method.

 

It is suggested that the author provide information on the recommended ECR density, such as the number of units per square kilometer, and discuss how the density might vary based on terrain or monitoring accuracy requirements.

 

Author Response

Comments 1: The original phrase “due to their small size a practical” seems to be grammatically incomplete and fails to clearly convey the author's intended meaning.
Response 1: Thank you for pointing this out. This has been changed and corrected on page 1, line 29/30 in the abstract.

Comments 2: Section 2.1 provides an overly detailed background on the PS-InSAR technique. While it includes valuable references, this technique is a common method for monitoring dam deformation and may not require such extensive coverage.

It is recommended that the author condense this section, focusing instead on the specific applications of PS-InSAR in dam monitoring and its advantages when combined with ECRs, to avoid lengthy descriptions overshadowing the main points of the article.
Response 2: Thank you, we have straightened this chapter a little bit. Changes can be found in Section 2.1 between lines 109 and 127.

Comments 3: The article does not mention the specific technical parameters of the ECRs, which may lead to insufficient understanding of their design and performance by the readers. It is suggested that the author include details such as the size, weight, backscatter intensity (e.g., Sigma-0 values), power supply, and operational lifespan of the ECRs.
Response 3: Some relevant technical details have been added in section 2.2.1 Background and State of the art (lines 142-160)

Comments 4: Although the article discusses the application of the CR Index, it does not provide a detailed explanation of its calculation method. This omission may hinder readers' understanding of the index's composition and applicability. It is recommended that the author add the formula for the CR Index or describe its calculation process.
Response 4: Some details have been added to this and also to its origin in section 2.2.2 CR Index calculation and modification (line 360/361).

Comments 5: The current article does not mention the deployment density of ECRs, which may prevent readers from evaluating the practical deployment requirements and costs of the method. It is suggested that the author provide information on the recommended ECR density, such as the number of units per square kilometer, and discuss how the density might vary based on terrain or monitoring accuracy requirements.
Response 5: Thank you for the comment. We added a short assessment of this in the 4. Discussion section (lines 740-748). This is indeed one of the limitations of this approach, as ECRs are costly.

 

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript presents the rationale for deploying Electronic Corner Reflectors (ECRs) in dam areas based on CR Index, and analyzes ECR performance through continuous monitoring experiments conducted across multiple dams. While this research provides a valuable foundation for ECR site selection and validates several key performance aspects through experimental verification, several areas require further attention and improvement:

 

(1)    Section 3.2 lacks detailed illustrations of ECR installation locations for the Lister and Sorpe dams. Additionally, CR index data for Bigge1, Bigge2, Lister, and Sorpe should be incorporated into Table 2.

(2)    The Bigge dam area exhibits notably high CR Index (>50) in both ascending and descending orbits, indicating inherently favorable satellite observation conditions. This raises questions about the necessity and potential advantages of ECR installation in such well-monitored locations.

 

(3)    Despite ECR installations at Lister and Sorpe dams, signal fluctuations and high ADI phenomena persist. The manuscript does not adequately explore the underlying causes of these observations, making it difficult to justify ECR deployment in high-ADI regions.

 

(4)    To better evaluate ECR effectiveness, comparative experiments documenting conditions before and after installation are needed to clearly distinguish between ECR targets and background reflection characteristics, as well as to identify potential positive or negative observational impacts.

 

(5)    In Section 4.1's discussion, PS and CR Index analyses should focus specifically on dam areas, excluding unrelated targets such as urban zones to ensure analytical accuracy and relevance. In addition, it is recommended to provide the spatial distribution image of PS.

 

(6)    The reflection signal intensity observed in the amplitude map (Fig. 8) does not fully correspond to the CR distribution, particularly in dam areas with high CR Index showing relatively low reflection signals. This observation, combined with the question (5), necessitates a reassessment of the correlation between dam area CR Index and PS quantities.

Author Response

Comments 1: Section 3.2 lacks detailed illustrations of ECR installation locations for the Lister and Sorpe dams. Additionally, CR index data for Bigge1, Bigge2, Lister, and Sorpe should be incorporated into Table 2.
Response 1: Thank you for your comment. The procedure at these dams was similar to the examples shown, which is why Lister and Sorpe are not shown separately. However, this was added to the text (line 505/506). The CR index is a preliminary step in the backscatter analysis and would not provide any additional information in this section.

Comments 2 The Bigge dam area exhibits notably high CR Index (>50) in both ascending and descending orbits, indicating inherently favorable satellite observation conditions. This raises questions about the necessity and potential advantages of ECR installation in such well-monitored locations.
Response 2: Thank you for your notice. However, this only affects a thin strip that represents the asphalted slope of the Bigge. This is disturbed by a hedge, which is why an ECR installation is advantageous here. Such small areas with high CR values are often not sufficient. An additional note on this was made in the text (line 491-494).

Comments 3: Despite ECR installations at Lister and Sorpe dams, signal fluctuations and high ADI phenomena persist. The manuscript does not adequately explore the underlying causes of these observations, making it difficult to justify ECR deployment in high-ADI regions.
Response 3: This is an important point. We ourselves do not yet know exactly why there are such strong fluctuations. Possible explanations are double bounce in connection with the water level or additional influences of the dam. However, it is important to note that the high ADI values persist despite the use of ECRs. The deployment of ECRs is intended to lower the index to enable a PSI analysis (line 721/722).

Comments 4: To better evaluate ECR effectiveness, comparative experiments documenting conditions before and after installation are needed to clearly distinguish between ECR targets and background reflection characteristics, as well as to identify potential positive or negative observational impacts.
Response 4: We have some dates where the ECRs were down due to power failure or bad weather. Backscatter values are significantly lower here. We were also able to obtain PS points by using the ECR (line 737-739).

Comments 5: In Section 4.1's discussion, PS and CR Index analyses should focus specifically on dam areas, excluding unrelated targets such as urban zones to ensure analytical accuracy and relevance. In addition, it is recommended to provide the spatial distribution image of PS.
Response 5: The main focus here is of course on insulation, but the validation of the CR Index includes the entirety. On the other hand, the CR Index would be obsolete, as it is primarily concerned with differences between different land use classes and geometry.

Comments 6: The reflection signal intensity observed in the amplitude map (Fig. 8) does not fully correspond to the CR distribution, particularly in dam areas with high CR Index showing relatively low reflection signals. This observation, combined with the question (5), necessitates a reassessment of the correlation between dam area CR Index and PS quantities.
Response 6: Yes, it does. It is exactly the location of the installed ECR at the Bigge and Möhne dams.

 

Reviewer 4 Report

Comments and Suggestions for Authors

The subject matter is important, and some interesting observations are presented but overall, it is hard to determine what new information and what novel ideas are advanced in this manuscript. 

My comments are as follows:

- The title needs adjustment - the word 'operationalising' is awkward in this context.

- Compact Active Transponders (or ECR) have been used previously. What is new in this application? It would be useful to describe the type of equipment (make/model) used and the observed performance. Are all currently available transponders operating only in C-Band or are there any units also operating in X-Band? What are the maintenance requirements for these units? Do they require lightning protection?

- If the PSI alone is insufficient then why not use a normal corner reflector (passive type) before considering active transponders?

- What are you trying to detect at each dam? What are the expected horizontal and vertical movements depending on the reservoir level? Is the proposed equipment capable of detecting displacement that may be considered a warning signal of impending failure?

- Can you present some displacement monitoring data from ECR? What results did you collect?

- The strategic placement of reflectors should not depend on the visibility potential. Sensor location should be selected by dam engineers (in this particular case). If the ECR can provide the required data then it may be used, otherwise some other sensor should be placed at that critical location.

- The advantage of a reflector is that it provides measurement at a particular location. In just about any type of infrastructure monitoring, the location of a potential defect is important. Corner reflectors can theoretically serve that role. 

- In some structural monitoring applications, the size of a reflector is important. You certainly wouldn't want to install a huge reflector on a bridge. It would be interesting to learn more about what is currently available commercially.

- There may be a way to present this paper as a case history of dam monitoring, but significantly more detail is required about each dam, the purpose for such monitoring, and what can be potentially gained. 

- The alternative approach would be to present this paper as a review of available ECR technology and potential applications to infrastructure monitoring.

- The conclusions should be relatively short and to the point. Currently this section reads like a discussion.

 

Comments on the Quality of English Language

Some minor English edits would be helpful. Current title should be modified.

Author Response

Comments 1: The title needs adjustment - the word 'operationalizing' is awkward in this context.
Response 1: The title was adjusted (line 2/3).

Comments 2: Compact Active Transponders (or ECR) have been used previously. What is new in this application? It would be useful to describe the type of equipment (make/model) used and the observed performance. Are all currently available transponders operating only in C-Band or are there any units also operating in X-Band? What are the maintenance requirements for these units? Do they require lightning protection?
Response 2: Thank you for your comment. The new approach was the use of ECR in combination with the CR-Index. There has been added some more detailed information about the ECR from MetaSensing (line 88 and 91/92). These devices are only available in C-Band.

Comments 3: If the PSI alone is insufficient then why not use a normal corner reflector (passive type) before considering active transponders?
Response 3: Because of the size of normal corner reflectors (too big for many dams), which is mentioned in the paper. Also, the issue of Monument Protection is important. Due to this it is not allowed to install bigger devices, such as passive corner reflectors, on dams. Some additional information has been added to the text (line 86-88).

Comments 4: What are you trying to detect at each dam? What are the expected horizontal and vertical movements depending on the reservoir level? Is the proposed equipment capable of detecting displacement that may be considered a warning signal of impending failure?
Response 4: Thank you for the question. We want to try to receive additional PS for dams that don’t have enough natural PS for better displacement monitoring. This is explained in the text though.

Comments 5: Can you present some displacement monitoring data from ECR? What results did you collect?
Response 5: This will be the contents of following investigations with sufficiently long time series data. However, it was possible to detect PS using ECR, so displacement monitoring would be possible. Some further information was added in 5. Conclusions (line 750-780).

Comments 6: The strategic placement of reflectors should not depend on the visibility potential. Sensor location should be selected by dam engineers (in this particular case). If the ECR can provide the required data then it may be used, otherwise some other sensor should be placed at that critical location.
Response 6: Yes, the visibility is crucial. If it is not visible from the satellite, there are no results. Sensor location was selected in close cooperation with the engineers from the Ruhrverband. An additional supplement in the text was made (line 493-494).

Comments 7: The advantage of a reflector is that it provides measurement at a particular location. In just about any type of infrastructure monitoring, the location of a potential defect is important. Corner reflectors can theoretically serve that role.
Response 7: We agree with this point. That is why we used the CR-Index on the one hand and on the other hand worked closely with the dam engineers.

Comments 8: In some structural monitoring applications, the size of a reflector is important. You certainly wouldn't want to install a huge reflector on a bridge. It would be interesting to learn more about what is currently available commercially.
Response 8: ECR from MetaSensing are the only commercially available devices (added in line 84/85 and line 731-734). As mentioned, the size is important. That is why “normal” corner reflectors can’t be installed on many dams. This is also mentioned in the paper.

Comments 9: There may be a way to present this paper as a case history of dam monitoring, but significantly more detail is required about each dam, the purpose for such monitoring, and what can be potentially gained. 
Response 9: The focus of this study lies more on the potential of ECR for bigger infrastructures like dams, for which passive corner reflectors cannot be used.

Comments 10: The alternative approach would be to present this paper as a review of available ECR technology and potential applications to infrastructure monitoring.
Response 10: Thank you for your suggestion, but we are sticking to our project.

Comments 11: The conclusions should be relatively short and to the point. Currently this section reads like a discussion.
Response 11: Thank you for the notice. We have straightened the conclusions a little bit (line 750-783).