Shallow Sea Topography Detection from Multi-Source SAR Satellites: A Case Study of Dazhou Island in China
Round 1
Reviewer 1 Report
Comments and Suggestions for Authors
The paper focuses on recovering shallow bathymetry using satellite images, which is SAR data's actual and prospective application. The authors extracted the same depth field from four SAR images using the known method presented in the paper [2] for optical images. They analyze errors of depth recovery via comparison with a true bathymetry chart. The paper may be the subject to consider for publication in Remote Sensing. But the presented version of the manuscript includes doubtful claims, unclearness, inappropriate terminology, and underdeveloped figures. I recommend a major revision.
General comments:
- Authors claim that they propose a new method MSSTD to evaluate water depth. But Section 2 (Methods) includes only a description of the method presented in the paper [2]. What is MSSTD? I recommend presenting of the method MSSTD in a separate subsection in Section 2 if the method MSSTD was developed.
- What is the novelty of this work? Is it in the application of the method for optical images to radar ones? I recommend clarifying it (say, in the Introduction or Discussion).
- In this paper, the error analysis use only “valid depths” selected from the known bathymetry chart. What must we do in case when bathymetry is not known? How can we isolate “valid” depth estimations? If authors cannot answer, then presented error analysis is senseless. In this case, they must analyze the errors for the entire retrieved depth fields. This point requires serious clarification. It needs for assessing the future application and performance of the method.
- In my opinion, the errors in retrieved depths are the most interesting issue of this research. Why they are shown in so an unclear manner? I recommend presenting figure 6 in plain view, with the sizes of each subfigure (a, b, and c) similar to the size of figure 7.
- Subsection 2.1 describes linear wave theory and the dispersion relationship for shallow water waves with references to recent remote sensing articles. As these issues are classical in general oceanology, I recommend changing references to more appropriate, say, to “Upper ocean dynamics” of O. Phillips or “Waves in the Ocean” of LeBlond and Mysak.
Specific comments:
97-99 “Then part of the wave in shallow water is covered by the wave in deeper water, which causes a change in the direction of wave propagation …” – This is unclear. I recommend reformulating or deleting it.
103-106 “The relationship between waves and water depth can be described by the linear dispersion relationship, but it does not apply to all sea areas. When the wave is in shallow water (?<?/2), the relationship between wave and water depth is as follows [19,20]” – It is not true. Equation (1) describes any gravity wave, but at ?>?/2 the depth rather weekly affects the waves.
Equation (8) and explanation text – Why wave direction is introduced if further there is no referring to wave direction?
Figure 4 b and c – These versions of the figures provide no information for the reader. I recommend redrawing them in plain view with bigger sizes or delete.
264-266 “It was not difficult to notice that the wave patterns of ALOS PALSAR and ENVISAT ASAR were not as distinct as those of GF-3 and Sentinel-1.” – A reader cannot compare these wave patterns. I recommend rewriting the sentence.
359-361 “However, the resolution and wavelength scale are relative. For a specific wavelength, the extraction ability may not continue to increase when the resolution reaches a certain level, which needs to be further verified.” These are unclear statements. They require clarifying.
394-396 “The size of FFT is crucial for wave extraction because the wavelength retrieved by FFT is the length of main wave rather than a single wave” – What is the “main wave”? What is the difference between main and single waves? This is using of non-traditional terminology. It is unclear and requires rewriting. In the second, it is unclear why the size of FFT is relevant to it.
411-413 “Figure 8 (a) revealed that the area with 0-10 m depth mainly located in the nearshore and west area of Dazhou Island.” - It is not true. Is it an error in Figure number?
415-416 “waves will diffract and refract when passing through Dazhou Island, causing weak wave patterns on the west area of island” – Why say about refraction and diffraction? If waves come from the sea then the island shades the area behind it. That is, the wave pattern is not weak, simply there are no sea waves behind the island.
417-419 “The large topographic gradient and weak frictional effect between waves and bottom topography resulted in an unsatisfactory bathymetry error when the water depth exceeded 50 m.” – The discussion of “large gradient” and “frictional effect” are groundless. The fact, that when the water depth exceeds one-half of wavelength the waves are insensible to the water depth, follows from the dispersion relation (1). This fact is known for a long time. Your research again confirms this fact.
Comments for author File: 
Comments.pdf
Author Response
Dear Reviewer,
Thank you very much for your positive and constructive comments and suggestions on our manuscript. We have studied the comments carefully and revised our manuscript. The response by point to point is in the attachment.
Once again, thank you very much for your comments and suggestions. And we hope that the corrections will meet with approval.
Thank you and best regards.
Author Response File: 
Author Response.docx
Reviewer 2 Report
Authors presented shallow sea topographical observations using multi-source SAR satellites around Dazhou Island in China. Authors presented very interesting topographical results combining GF-3, Sentinel-1, ALOS PALSAR and ENVISAT ASAR datasets. Manuscript is structured as follows: introduction, theory input, study area, results, discussion and conclusions. However some parts need clarifications and reordering.
Critical comments:
Fig. 1. – The caption is not informative. What data was used to make the figure? Please clarify.
Fig. 3. – The caption is not clear. What schematic chart do we see here? Please clarify.
Lines 209-212 – Please move the introduction to the other section. Please reorder manuscript.
Lines, 215, 226 – Information on subregions selected should be placed in the section of the study area. Please reorder manuscript.
Lines 217-219 – This information should be placed in a caption of the figure. Please reorder manuscript.
Lines 257 and 261 – Description of the figure 2 cannot be placed after description of the figure 4. Please reorder manuscript.
Fig.6. – Presented material is not clear. Please improve visualization of the results.
Lines 292–293 – Should be moved to introduction.
Lines 305–306 – This information is not needed in this place. Please reorder manuscript.
Lines 308–312 – Either this part should be moved to subsection 4.3 or figure 6 should be reorganized. Please improve.
Line 323 – Please note that figure 3 is placed in the other section. Please reorder manuscript.
Fig. 7. – Caption is not informative. What topography was visualized here, please clarify.
Line 327 – This is not clear subsection’s title. Please clarify.
Line 329 – Figure 6 was described in the section 4.3. Why authors citeed particularly the figure 6B in the section 5.2? This is not clear, please clarify.
Lines 333–337 – These are not authors’ results. Please move these lines to the other section.
Lines 370–371 – Why authors mentioned here about polarization if it had no influence on topography detection? This is not clear, please clarify.
Lines 399 and 416 – This is not clear why authors citeed (second time) figures from the previous sections (pages). It looks as order of the manuscript is not correct. Please reorder manuscript.
Author Response
Dear Reviewer,
Thank you very much for your positive and constructive comments and suggestions on our manuscript. We have studied the comments carefully and revised our manuscript. The response by point to point is in the attachment.
Once again, thank you very much for your comments and suggestions. And we hope that the corrections will meet with approval.
Thank you and best regards.
Author Response File: Author Response.docx
Reviewer 3 Report
Inshore bathymetry can be obtained from synthetic aperture radar (SAR) images of ocean waves using the dispersion relation. This method is well established, but generally only SAR images from a single satellite have been used. In this manuscript the authors propose using SAR imagery from several different satellites to improve the accuracy of the resulting bathymetry.
More accurate results would be expected from using multiple SAR images, because random errors will be reduced with more data. This has been found in previous work using multiple images from the same satellite, for example Pereira et al. (2019, reference 10 in the manuscript). The necessity or desirability of using images from different satellites is not explained in the manuscript. Why is it better to use images from several different satellites instead of using multiple images from the same satellite? The single image from the GF-3 satellite gave the best results out of the four individual satellites. Using additional images from GF-3 could give better results than including the images from the other satellites.
In the first paragraph of Section 3.2, and again in Section 6.4, it is claimed that waves travelling in the azimuth direction in SAR images cannot be used to derive accurate bathymetry using fast Fourier transforms (FFT). This is not correct, because Brusch et al. (2011, reference 12 in the manuscript) obtained accurate bathymetry from azimuth travelling waves in TerraSAR-X images using FFTs. It is true that SAR will only image azimuth travelling waves with wavelengths longer than the cut-off, so it may be more difficult to find suitable images, but the non-linearity discussed by Collard et al. (2005, reference 23 in the manuscript) and Wan et al. (2020, reference 25 in the manuscript) is mainly a problem for estimation of the entire wave spectrum, whereas only the peak wavelength is required for bathymetry.
I don't think the depth estimation method used in this work makes the best use of the available data. For each SAR image, the wave period is estimated using reference water depths in four sub-regions, with the average of the four estimates used. A maximum wavelength L_max is then calculated from the estimated wave period, and all wavelength estimates greater than L_max are discarded. If the estimated wave period is too short, then L_max will be too low, resulting in both incorrect depth estimates and too many data points being discarded. The method used by Wiehle et al. (2019, reference 30 in the manuscript) seems much better. They estimated the wave period iteratively, by minimising the root mean square error in the depth estimates across the entire region of interest. Rather than setting a maximum wavelength, Wiehle et al. used the longest estimated wavelength to set a minimum wave period for the iteration.
Figure 8 shows that there is a systematic error in the method, with depths greater than about 20 m being consistently underestimated for all four individual satellites, as well as the combined case. This should be pointed out in the discussion in Section 6.3. Clearly the method becomes less accurate as the water depth increases and the dispersion relation approaches the deep water limit, but it is not clear why the depth is almost always underestimated. Correction of this systematic error could improve the accuracy of the method more than using images from multiple satellites, at least for water depths of 20 - 50 m. Possibly the error is a result of the wave period being underestimated, as discussed above. If the authors have any ideas on this, it would be worth discussing them in the manuscript.
Corrections to the manuscript:
Throughout the manuscript: The standard expression in physics is "dispersion relation", not "dispersion relationship".
The abbreviations MAE, MRE and MAR need to be defined on first use. (I suspect that MAR, used in Figure 8 and line 363, is a misprint.)
Abstract, last sentence: "hot issue" is colloquial English, inappropriate for a journal publication.
Lines 67-69: "The accuracy of topography detection with a single SAR image was usually low, due to unfavourable sea conditions and sensor noise."
Section 2.2: Change the heading to "Wave retrieval by FFT"
Section 3.3, line 202: "acquired", not "inquired"; "points", not "point"
Section 4.2: Use "wavelength estimation", not "inversion". Change the first sentence to "Wavelength estimation is ...", and replace "inversion" with "estimation" in the last paragraph of this section (3 instances).
Figure 6: This figure is not presented well. The details are almost impossible to read, due to the way that the plots A, B and C are shown projected at an angle to the page. It would be better to display them conventionally like Figure 5.
Section 4.3, last sentence: "The sensitivity of the bathymetry model to different water depths will be discussed in detail in Section 6.3."
Section 4.4, first sentence: "The detection accuracy will be low if ..." (rather than "disappointing")
Section 5.1, line 323 should read: "... depths increasing rapidly from 20 to 90 m."
Section 5.2, line 336: "... obtained water depths ...", not "inversed"
Figure 7: The depths on the scale should be positive, not negative. It would help the reader to compare Figure 7 with Figure 3 if these figures were combined, or shown together on the same page. It would also help to use the same colour scale in both figures - red indicates deep water in Figure 3, and shallow water in Figure 7.
Figure 8: The label on graph (a) should be "MSSTD"
Section 6.4, line 426: "However, this method has some limitations ..." (not "exists some restrictions")
Author Response
Dear Reviewer,
Thank you very much for your positive and constructive comments and suggestions on our manuscript. We have studied the comments carefully and revised our manuscript. The response by point to point is in the attachment.
Once again, thank you very much for your comments and suggestions. And we hope that the corrections will meet with approval.
Thank you and best regards.
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
Comments and Suggestions for Authors
The authors adequately addressed most of my previous comments. A minor revision will improve the paper further.
Comments:
L72: The term MSSTD requires full definition: based on multi-source SAR topography detection (MSSTD)
L302-305: The term “valid depth” is defined in L317, but it used early in the discussion of Figure 8. I recommend moving the definition before L302.
Figures 4 (b) and (c) were not revised (see comment 9 in Review 1): A drawback of these figures is their small size. Therefore a reader cannot get any information from these figures. I recommend enlarging or deleting them.
L372 (see comment 11 in Review 1): I recommend deleting of the sentence “However, the scale of image spatial resolution and wavelength are relative.”
L428-429 (see comment 14 in Review 1): The basic idea of this article is the persistence of the wave frequency along the wave rays. Therefore, the bathymetry can be retrieved only along the wave rays from “initial points” up to the island. If you see “wind-wave or different direction waves” in the shadowed area, then the frequencies of these waves are unrelated to the wave frequency at the initial point. Thus you cannot say about bathymetry retrieval for shadowed areas. I recommend replacing
“waves will diffract and refract when passing through Dazhou Island, causing weak wave patterns on the west area of island”
with
Commonly the area to the west of Dazhou island is shadowed and cannot be used in bathymetry retrieval.
Comments for author File: Comments.pdf
Author Response
Dear reviewer,
Thank you for your decision and further constructive comments on my manuscript. We have carefully considered the suggestions and made some changes. We have tried our best to improve and made some changes in the manuscript. Please see the attachment.
Author Response File: Author Response.docx
Reviewer 3 Report
The authors have addressed the reviewers’ comments in the revised manuscript. It is acceptable for publication with the minor corrections listed below.
While the authors have demonstrated that using multiple SAR images results in better estimates of the bathymetry, their work does not prove that it is better to use multiple images from different satellites rather than from the same satellite. It would be worth considering for future work how this hypothesis might be properly tested.
Robust and accurate estimation of the wave period and wavelength from short data records is difficult. A method that the authors may find useful in future work is described by P. S. Bell and J. C. Osler, “Mapping bathymetry using X-band marine radar data recorded from a moving vessel”, Ocean Dynamics, vol. 61, pp. 2141-2156 (2011). See Section 4.5. Images from a real aperture radar are used in this paper, but the method would also be applicable to SAR images.
Some minor corrections to the revised text:
Section 2.1, line 104: Replace “Peculiarly” with “In particular”
Section 4.3, line 282: “these were”, not “there were”
line 289: “average values” not “average value”
line 302: “Figure 8 shows …”
line 304: “… SAR is significantly …”
Section 4.3, lines 304, 305: Don’t state the aim of the paper near the end! Delete this sentence or move it to the Introduction.
Section 6.3, line 430: “(exceeding 50 m)”
Section 6.3, lines 438-440: Replace with “… if the wave period is overestimated, so it is more likely to be due to underestimation of the wavelength.” (I don’t understand what “interpolation” (line 439) refers to.)
Section 6.4, line 456: “… waves with wavelengths smaller than the cut-off …”
line 458: “ … considered for waves propagating along the SAR azimuth direction.”
Author Response
Dear reviewer,
Thank you for your decision and further constructive comments on my manuscript. We have carefully considered the suggestions and made some changes. We have tried our best to improve and made some changes in the manuscript. Please see the attachment.
Author Response File: Author Response.docx
