Data Quality Evaluation of Sentinel-1 and GF-3 SAR for Wind Field Inversion

Round 1

Reviewer 1 Report

I do not believe that the paper “Data quality evaluation of Sentinel-1 and GF-3 SAR for wind field
inversion”, by Yong Wan, Sheng Guo, Ligang Li, Xiaojun Qu, Yongshou Daican should be published on Remote Sensing.

 

My main objection is that the Authors do not really “invert” SAR data to  retrieve wind speed, but – from what I understand - they also make use of the wind direction from model data ( ERA5, line 215). There is a whole body of literature on how to extract or estimate such direction from SAR images, but they do not mention it, nor they explain why they decided to neglect it

A second critical aspect is how the actual inversion is carried out:  CMOD models can compute a backscattering coefficient Sigma0 once wind speed , direction, incidence angle are provided. The opposite (computing wind speed from the other parameters, hence the word “inversion”) is not a trivial procedure, but the Authors do not supply any information on how it is carried out.

The paper is also lacking from various aspects: 

-there is no description of the IW and SLC SAR mode data the Author  are using, and no clear refence is provided

-Sigma0vv values are obviously averaged in in blocks (line 213) but no information is give on the size of such blocks

-the references are inadequate for such a complex and ambitious job

A few more remarks are reported in the attached pdf file

 

Since the  Authors have access to a large amount of SAR data and they  obviously  have the competence the experience to handle them, they could be perhaps encouraged to produce a new paper, with a less ambitious aim, as for instance validating ERA data with SAR. i.e they could  compare CMOD computed   Sigma0 with experimental SAR data- that would not require “inverting” CMOD

Comments for author File: Comments.pdf

Author Response

Thank you for giving constructive comments for our manuscript. We have given serious consideration to Reviewer’s comments and suggestions and have revised the manuscript carefully.

General Comments: I do not believe that the paper “Data quality evaluation of Sentinel-1 and GF-3 SAR for wind field inversion”, by Yong Wan, Sheng Guo, Ligang Li, Xiaojun Qu, Yongshou Daican should be published on Remote Sensing.

Revision detail:

Thank you for your careful review. We know that our manuscript has many shortcomings, but we will try our best to modify and improve our manuscript according to your comments.

Specific Comments:

1. My main objection is that the Authors do not really “invert” SAR data to retrieve wind speed, but – from what I understand - they also make use of the wind direction from model data ( ERA5, line 215). There is a whole body of literature on how to extract or estimate such direction from SAR images, but they do not mention it, nor they explain why they decided to neglect it

Revision detail:

Thank you very much. The use of CMOD requires wind direction information. The sources of wind direction information can be divided into wind direction extracted based on SAR image wind stripe information and external auxiliary wind field data (such as ERA5). The wind direction information used in this paper is the external auxiliary wind field data.

2. A second critical aspect is how the actual inversion is carried out: CMOD models can compute a backscattering coefficient Sigma0 once wind speed, direction, incidence angle are provided. The opposite (computing wind speed from the other parameters, hence the word “inversion”) is not a trivial procedure, but the Authors do not supply any information on how it is carried out.

Revision detail:

Thank you for your careful review. We have added the process of wind speed inversion.

3. there is no description of the IW and SLC SAR mode data the Author are using, and no clear refence is provided.

Revision detail:

Thank you for your careful review. We have added a simple description of the IW and SLC SAR mode data.

4. Sigma0vv values are obviously averaged in in blocks (line 213) but no information is give on the size of such blocks

Revision detail:

Thank you for the helpful comments. In this study, the backscatter coefficient of each scene is divided into 25×25 blocks.

5. The references are inadequate for such a complex and ambitious job

Revision detail:

Thank you for your careful review. We know the references are inadequate, but we have tried our best to add some references at the end of our manuscript.

6. Line 8: Please check this statement through an appropriate reference search.

Revision detail:

Thank you very much. Maybe our search is not comprehensive enough, we searched found no information about the evaluation of Sentinel-1 and GF-3 SAR data quality applied to wind field inversion. It is because there is no relevant reference to prove that no one has evaluated and compared these two data, so this paper evaluates and compares these two data.

7. please clarify what you mean by "inversion" , i.e extracting wind speed from the backscattering coefficients. Mention, at least once that you are considering "wind sea speed retrieval"

Revision detail:

Thank you for your careful review. “Inversion” means extracting the wind speed at 10m above the sea surface from SAR data.

8. Line 22: “forefront”? what do you mean?

Revision detail:

As the result given in the manuscript, “The RMSE of wind field inverted from Sentinel-1 SAR data and GF-3 SAR data is less than 2m/s, which is in line with the recognized standard” and “the data quality of Sentinel-1 SAR is slightly better than the quality of GF-3 SAR”, indicating that the data quality of GF-3 is very close to that of Sentinel-1. Sentinel-1 data is world-class, so the data quality of GF-3 has also  achieved the world leading level.

9. Line 37: This sentence is incomprehensible

Revision detail:

Thank you for the helpful comments. We have rewritten this sentence.The sentence means: these methods can only obtain offshore wind field information.

10. Line 39: You probably mean: the intervals between measurements from ships at sea are scarce and irregular.

Revision detail:

Thank you for your careful review. This sentence means: Ship observation takes a long time, and the data obtained are only distributed in the navigation trajectory.

11. Line 45: You forgot to mention altimeters.

Revision detail:

Thank you very much. We have added the altimeter.

12. If I understand well, you make use of a VV CMOD, and your data are actually VV; so, why do you need eq 1 ?

Revision detail:

Thank you for your careful review. In introduction section, we introduce the current situation of wind field inversion, including VV, HH, and cross-polarization. HH polarization SAR data wind field inversion generally requires polarization ratio model, so there is an eq 1.

13. Line 87: appear?

Revision detail:

Thank you for your careful review. We have revised the word. We changed “appear” to “show”.

14. “The two-inversion wind speed”?

Revision detail:

Thank you for the helpful comments. We have revised it. We changed “the two-inversion wind speed” to “the inverted wind speed of the two satellites”.

15. It would be useful if you could provide a simple description of the IW mode which you are using.

Revision detail:

Thank you for your careful review. We have added a simple description of the IW mode.

16. It would be useful if you could provide a simple description of the SLC mode which you are using.

Revision detail:

Thank you for the helpful comments. We have added a simple description of the SLC products.

17. This is the most questionable part of your paper. CMODs are not wind retrieval fucntions, quite the opposite, they provide the backscattering Sigma0 coefficient sas a function of wind speed, direction and incidence angle - as you make clear in your eq 4 . Retrieving wind speed is in any case a complicate procedure - which you do not clarify; besides it needs an a priori knowledge of the wind direction, which you do not have- unless you use ERA data, as you state later in the paper (line 215) You are not really retrieving or "inverting" anything, just perhaps validating Sigma0

Revision detail:

Thank you very much. The use of CMOD requires wind direction information. The sources of wind direction information can be divided into wind direction extracted based on SAR image wind stripe information and external auxiliary wind field data (such as ERA5). The wind direction information used in this paper is the external auxiliary wind field data. We have added the process of wind speed inversion.

18. Please clarify which Sigma0vv do you use: you have a large number of data , given the high resolution of SAR images. Do you average them in blocks? (line 213); and how?

Revision detail:

Thank you for your careful review. We have added these in our manuscript. The backscatter coefficient of each scene is divided into 25×25 blocks, and each block represents a sub image. The backscatter coefficient of each sub image takes the average value as the backscattering coefficient of the sub image.

19. How are the blocks? How are they chosen?

Revision detail:

Thank you very much. We have added these in our manuscript. Each scene is divided into 25 × 25 blocks, 25 × 25 wind speeds are retrieved, and then invalid data are removed.

Reviewer 2 Report

The paper investigates the accuracy of SAR date delivered by Sentinel-1 and GF-3 satellites for Hormuz Strait, Malacca Strait, and US East and West coasts. The obtained inverted wind speed is analyses in relation to the wind speed provided by ERA5 and by buoys, respectively. The paper concluded, as expected, that “The quality of SAR data from both satellites is high” and “The RMSE of wind field inverted from Sentinel-1 SAR data and GF-3 SAR data is less than 2m/s, which is in line with the recognized standard”. Hence, my concern is about the usefulness of the paper as it reached obvious expected conclusions!

 

The following issues are recommended to improve the paper:

1. Define/introduce the acronyms at their first use, even they are well known in the topic literature, e.g. CMOD5.N, ERA5, HH, VV, etc.
2. Abstract: please justify the last (exaggerate!) sentence “The SAR data quality of GF-3 has reached the forefront of the world.”
3. Typically, the Introduction section states explicitly the challenging problem of the paper in relation with the knowledge limits in the field, the original aspects and novelty of the paper results, along with the paper organisation.
4. Line 141: “…and swath. The width can reach 650km,…” – unclear statement.
5. Line 144: “GF-3 provides four standard products of L0, L1, L2 and L3 and L4” please clarify four or five standard products, as there 5 items are enumerated (L0 to L4)!
6. Recommendation to improve the English style in several (few) cases!
7. Equations (3) and (4): please indicate appropriate references.

Author Response

Thank you for giving constructive comments for our manuscript. We have given serious consideration to Reviewer’s comments and suggestions and have revised the manuscript carefully.

General Comments: The paper investigates the accuracy of SAR date delivered by Sentinel-1 and GF-3 satellites for Hormuz Strait, Malacca Strait, and US East and West coasts. The obtained inverted wind speed is analyses in relation to the wind speed provided by ERA5 and by buoys, respectively. The paper concluded, as expected, that “The quality of SAR data from both satellites is high” and “The RMSE of wind field inverted from Sentinel-1 SAR data and GF-3 SAR data is less than 2m/s, which is in line with the recognized standard”. Hence, my concern is about the usefulness of the paper as it reached obvious expected conclusions!

Revision detail:

Thank you for your careful review. The purpose of our manuscript is to show the application effect of the two satellites in wind field inversion and compare the quality of Sentinel-1 and GF-3 SAR data applied to wind field inversion.

Specific Comments:

1. Define/introduce the acronyms at their first use, even they are well known in the topic literature, e.g. CMOD5.N, ERA5, HH, VV, etc.

Revision detail:

Thank you very much for your valuable suggestions. The acronyms have been defined/introduced.

2. Abstract: please justify the last (exaggerate!) sentence “The SAR data quality of GF-3 has reached the forefront of the world.”

Revision detail:

As the result given in the manuscript, “The RMSE of wind field inverted from Sentinel-1 SAR data and GF-3 SAR data is less than 2m/s, which is in line with the recognized standard” and “the data quality of Sentinel-1 SAR is slightly better than the quality of GF-3 SAR”, indicating that the data quality of GF-3 is very close to that of Sentinel-1. Sentinel-1 data is world-class, so the data quality of GF-3 has also achieved the world leading level.

3. Typically, the Introduction section states explicitly the challenging problem of the paper in relation with the knowledge limits in the field, the original aspects and novelty of the paper results, along with the paper organisation.

Revision detail:

Thank you very much. We add the paper organization. The purpose of our manuscript is to show the application effect of the two satellites in wind field inversion and compare the quality of Sentinel-1 and GF-3 SAR data applied to wind field inversion, so in introduction section, we introduce the purpose and significance of wind field inversion, the current situation of wind field inversion and research content.

4. Line 141: “…and swath. The width can reach 650km,…” – unclear statement.

Revision detail:

Thank you for your careful review. We have revised the statement: The swath width can reach 650 km.

5. Line 144: “GF-3 provides four standard products of L0, L1, L2 and L3 and L4” please clarify four or five standard products, as there 5 items are enumerated (L0 to L4)!

Revision detail:

Thank you very much. There are four standard products of L0, L1, L2, L3 and one industry application products of L4.

6. Recommendation to improve the English style in several (few) cases!

Revision detail:

Thank you very much. We will improve the English style.

7. Equations (3) and (4): please indicate appropriate references.

Revision detail:

Thank you very much. We have indicated appropriate references.

Round 2

Reviewer 1 Report

 The paper has been improved. I believe it can be published in the present form, with minor changes in the Englis style.

 

Reviewer 2 Report

No any additional recommendations.