Analysis of Depths Derived by Airborne Lidar and Satellite Imaging to Support Bathymetric Mapping Efforts with Varying Environmental Conditions: Lower Laguna Madre, Gulf of Mexico

Round 1

Reviewer 1 Report (New Reviewer)

Comments and Suggestions for Authors

The authors evaluated depths derived by airborne Lidar and Sentinel-2 satellite images in various environmental conditions.

However, the main scientific contribution or the novelty of this study in Remote Sensing science is not clear. This study can be classified as a scientific report for evaluating the airborne Lidar for bathymetry detection. Also, the introduction and discussion sections need to be improved and more related recent studies about evaluating the airborne Lidar and Sentinel-2 satellite images for bathymetry and water quality assessment should be cited and discussed. Finally, the evaluation of these systems for bathymetry and water quality assessment was performed before in many studies.

Thus, I suggest rejecting this manuscript.

Comments on the Quality of English Language

In general, there were some structure and English language errors scattered throughout the manuscript, for which the authors need to account for.  The manuscript should have English proofreading.

Author Response

Authors’ response:

Thank you for reading and critiquing our article. We appreciated your time and your attention, and your decision.

The article summarizes our comprehensive efforts to measure the depths of lower Laguna Madre using airborne lidar technology. The literature review supports the novelty that in locations with relatively transparent locations, it is common and straight-forward to measure the bottom of shallow water columns, and this is noted and referenced in our manuscript with citing previous studies. However, there is a knowledge gap in the airborne lidar community about lidar surveys conducted in challenging, very shallow, and variable environmental conditions. As stated very clearly in the introduction section, we had a principal and four supplementary study motivation topics. The principal motivation was to experiment and analyze the possibility of conducting a feasible airborne lidar survey in such challenging and not-so-transparent water column conditions.

As indicated in the results section, greater than 51% of the entire lidar survey had high probability of measuring the bottom of the lagoon. Supplementary surveys and analysis of Sentinel-2 imagery backed this finding and we presented the results with relevant statistical analysis.

With your suggestion, we referenced an additional three studies in the literature that investigated satellite bathymetry and airborne lidar on coastal waters. These studies investigate fused bathymetric datasets on transparent waters where both technologies were successful measuring the coastline at relatively shallow and transparent waters. These additions are now included in the second paragraph, and we highlighted them in the revised manuscript.

We also emphasized the importance of our study in the introduction section, and in the conclusion. These editions are also highlighted.

Reviewer 2 Report (New Reviewer)

Comments and Suggestions for Authors

see attachment

Comments for author File: Comments.pdf

Author Response

Thank you for reading and critiquing our article. Reviewers like yourself make this a positive and an efficient process. We appreciate your time and attention to the detail. With your feedback, our manuscript has improved and will serve better to the scientific community. Please find comments/response in the below text:

Comment: In this paper, various means such as Lidar, Sonar, and satellite imagery were used for bathymetry in a high salinity lagoon to compare the measurement accuracy and applicability, and the conclusions are detailed and reliable. The following recommendations are for reference.

• Comment: Overall, the style of the article is more engineering-oriented, with a slight lack of innovation in the methodology and theory. Although the first part clarifies the concern of the article and summarizes the response in the conclusion part. The authors could have summarized more innovative points in data processing, such as laser point classification (page 6), correlation analysis of multi-class data, etc.
  • Response: You are certainly correct. There is a slight twist of engineering-oriented knowledge and expertise in our study, which falls onto geomatics and hydrologic engineering. We believe that we lacked a detailed explanation of the study motivation; therefore, we edited the introduction section, better outlining the principal and the supplementary research motivation. Please see the highlighted sections.
  • Furthermore, the article “summarizes” the vector data analysis in Section 2.4 – Vector data analysis. This is not a in-deep explanation rather outlining our “innovative” data analysis process. We have removed a fictional data comparison figure to save space, but provided reference to a previous study that illustrated a similar methodology: Saylam et al 2018 and 2020.
  • We included brief explanation of lidar data classes that we used in our analysis in Section 2.5 – Airborne lidar bathymetry and system calibration. We believe that there is adequate information presented and any more information could be redundant with previous studies and may be unnecessary.
• Comment: Page 2, line 58, delete “Can we predict varying levels of turbidity based on satellite imaging”. It's a relatively mature topic for certain.
• Response: Absolutely correct! There are various studies in the literature that summarized those efforts, and we cited a few relevant ones in our manuscript in the introduction section. We also edited the research question to: “How can we predict varying levels of turbidity based on satellite imaging? What were the potential benefits of conducting quantitative pixel reflectance analysis in bathymetric lidar mapping?”.
• Comment: Page 6, line 190. Is the point cloud categorization a result of LLSS own, or is it a result of the author's processing? It is recommended that typical waveform diagrams for each category be added for illustration to facilitate the reader's understanding.
  • Response: These classes are output by LLSS, and the user has control over the backscatter thresholds. We edited the text to reflect this information.
  • We agreed that waveform plots illustrating each bathymetric lidar class would facilitate better understanding of classes. Therefore, we created an Appendix (B) and added to the manuscript with brief captions.
• Comment: Page 8, line 288-289, please provide more information about in-situ data, such as the number of points.
  • Response: This information is now available in the edited text.

 

 

 

Reviewer 3 Report (New Reviewer)

Comments and Suggestions for Authors

As noted in my marked up PDF, I would emphasize that some of the regression analysis and distribution analysis could be improved. In particular, the cubic regression seems to deviate/detract from a physical interpretation of the data (perhaps a normalized power law/logarithmic fit would better). Also, there were a couple histograms that were close to normal, but exhibited some skew, and that might be worth getting into a tiny bit. Finally (again, noted in the marked up PDF), for such shallow depths when/where turbidity impacts both sonar and lidar, maybe future independent data collections useing a mechanical measurement technology would be insightful (like a lead line/CTD cast).

Comments for author File: Comments.pdf

Author Response

Thank you for your time and friendly comments. Reviewers like yourself make this an effective process. Please find comments/response text below:  

Comment: As noted in my marked-up PDF, I would emphasize that some of the regression analysis and distribution analysis could be improved. In particular, the cubic regression seems to deviate/detract from a physical interpretation of the data (perhaps a normalized power law/logarithmic fit would better). Also, there were a couple histograms that were close to normal, but exhibited some skew, and that might be worth getting into a tiny bit. Finally (again, noted in the marked-up PDF), for such shallow depths when/where turbidity impacts both sonar and lidar, maybe future independent data collections using a mechanical measurement technology would be insightful (like a lead line/CTD cast).

• Comment: Because sonar and lidar are both impacted by turbidity, and because the depths are generally shallow, it might be beneficial to also measure the depth from the kayak using a more mechanical method, like a CTD cast/lead line. (Not for this paper, of course!).
  • Response: Yes, absolutely, and thank you for the suggestion. We have done this practice in the past, in much smaller environments. Laguna Madre survey was more than 800 km^2 in area size, and was only accessible in the south, therefore, measurements using a lead pipe/line would be exhaustive. However, as you suggested, it could provide an insight as an additional confirmation when needed. We will consider this again in the future and in similar studies.
• Comment: Note the weather/atmospheric conditions here - relatively uniform during the 1-2 weeks?
  • Response: Yes, and this information is now added to the article.
• Comment: Previously noted S~35.2%o?
  • Response: Great catch and thank you! We corrected the error.
• Comment: Figure 3: Explored other fits? normalized power or logarithmic?
  • Response: With your feedback, we explored other fits. Quickly, we realized that the relationship between maximum depth versus turbidity was linear using logarithmic conversion. This was a great suggestion and makes more sense looking at the graph. We revised the article and the figure accordingly.
  • We are no longer using a cubic relationship, and this also was suggested by the editor. We only kept linear relationships.
• Comment: Therefore, measurements greater than 0.4 m (> 51%) have a higher probability to be accurate and represent 376 the actual lagoon bottom (Figure 8). Can you quantify this probability?
  • Response: This is an assumption based on the lidar bathymetry results. We have no other means of quantifying it, and we need to trust lidar bathymetry results. However, as a matter of fact, computations reveal that Chiroptera cannot resolve depths shallower than ~36 cm because of the sampling rate and the frequency resolving capability of the digitizer. Therefore, our statement in the article describes this accurate.  
• Comment: Figures 9a and 10a: I disagree with the use of cubic regression here in Figures 9a and 10a (similar argument as above in Figure 3); while the correlation may be higher, it doesn't help with understanding of mechanism/reasoning for observed differences. The linear fits seem to better match any potential extrapolation of patterns in the data (in particular the left-side trends).
  • Response: Indeed, we agree with you! This was a suggestion of an earlier reviewer, hence why all regression figures were showing both linear and cubic regression results. Also, this correction was also suggested by the editor, therefore, we reverted all of them back to linear regression.
• Comment: Figures 12a and 13a: Surprising. From (a), I would expect more of a bi-modal distribution than you see in (b). Both histograms do indicate a measure of skewness.
  • Response: This is a good observation. We agree with you that both histograms indicated a measure of skewness, which we expected as well, observing the results. There is a slight bi-modal distribution in the survey 05/01 result, and we edited the figure captions accordingly.

Round 2

Reviewer 1 Report (New Reviewer)

Comments and Suggestions for Authors

The authors applied the suggested amendments and the manuscript is ready for publication.

Comments on the Quality of English Language

No comments for the English language.

Author Response

Thank you for your kind response and taking time to review our manuscript. With your efforts, we have improved the quality of our study.

Best regards,

Kutalmis Saylam

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

The paper deals with the comparison of different methods to determine water depths in a very shallow and difficult access lagoon. 

The article is very well written, the methods and experimental procedures are clearly described, and the results well demonstrated.

Figures are of good quality, but legends on figures 19, 14 and 15 are too small and difficult to read.

No need for further adjustments on the text.

Comments for author File: Comments.pdf

Author Response

Please see attachment. Thank you for your comments and suggestions.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

1.      Figures 1, 3, 4,  and 7 are unnecessary.  Figure 5 needs scale bars. The aerial photo-based maps (Figure 2, 6, and similar) are not legible at page scale. They can be read after zooming in on the PDF, but I don’t know if Remote Sensing accepts this. Figures 11 and 12 need 1:1 lines on them. Depths (Figure 11, 12, 13, 15) are not negative. Caption for Figure 15 needs to explain what white areas are. Caption for Figure 18 needs to explain what black and pink-gray areas are.

2.      Essential context is not presented, or is not presented clearly: saltwater inlets to lagoon are not identified, nor is noted that inlet channels are dredged.  Locations of Port Mansfield and Port Isabel tide gauges not plotted on any figures.

3.      Tidal corrections are not clearly addressed. 

a.      I found the discussion of tidal corrections (lines 139-147, Table 2, and elsewhere) nearly incomprehensible and difficult to reconcile with what I see on the NOAA web pages for the Port Isabel and Port Mansfield gauges.

b.      Lines 410-412 state that lidar (and sonar?) depths were adjusted with water heights as observed at the Port Isabel gauge. I could not find any similar statement on correction of SDB depths for short-term changes in water height.  Were SDB depths corrected for tides (and maybe wind surge)?

c.      Differences between water levels predicted (and observed) at the two gauges indicate that the tidal effects vary significantly along the length of the lagoon.  How was this dealt with? 

d.      It appears the Chiroptera lidar instrument can run NIR and green systems simultaneously.  What did NIR measurement of water-surface elevation have to say about tidal variations over the extent of the lagoon, and at a given place over the duration of the survey?

e.      Figure 14 is tantalizing, but what difference is being depicted?  ALB swath to swath differences? ALB to SDB differences?  The caption suggests this is a map of tidal range (“surface heights difference due to tidal influences”), but this is clearly wrong.

f.       What datum (NAVD88, MSL, MLLW, local water level at time of survey, …) are depths in Appendix A and B referenced to? Was the sonar system calibrated for weight of the watercraft propulsion unit? Was watercraft propelled with no significant side-to-side rocking? What is relation between “WP SNR depth (m)” and Avg SNR depth (m)” I Appendix B?

4.      The explanation offered for the 3.8 cm height difference between green and NIR lidar heights (lines 370-373) makes no sense to me.  Perhaps this is differential range walk (pulse detection is brightness dependent) showing up where the target albedos (in green and NIR) are different than the target albedos the system was calibrated with?  

5.      2nd point of the conclusions (“In-situ campaigns are essential…”; lines 590-594) _may_ be correct, but this point is not demonstrated by the preceding manuscript

Comments on the Quality of English Language

I found much extraneous material (e.g., most of 1^st two paragraphs), the prose is far too wordy, and there are many mildly to moderately inappropriate word choices (e.g., tidal variation used interchangeably with tidal variance, when in both cases tidal range would be better). Some statements, e.g. "We excluded depths that registered greater than 0.01 m..." (line 445 and following), appear to be flat-out wrong.

Author Response

Please see attachment. Thank you for your review and your comments.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

Some qustions and suggestions are listed below. Also,  the quality of description needs to be improved. Many confusing sentenses distract the reader from the main points of the study. Below I provide some examples of parts needing revision, but the authors should revise the whole manuscrip.

1.      13-15/1: -- “due to size (1600 km2 ) and the dynamic environmental conditions that influenced the depth and water quality ” This statement is confusing. Please explain the reason of “the nature of this data acquisition campaign was unique” clearer

2.      15-16/1: the pharse “ and completed in-situ measurements” is redundant, rephrase the sentense

3.      18-19/1: “lidar measurements were compared to GPS elevations, sonar, and satellite derived bathymetry using least-squares algorithm” confusing sentense

4.      19-21/1: “Because Airborne Lidar Bathymetry (ALB) technology is detailed compared to satellite imaging, results produced skewed distribution for satellite derived bathymetry with an average depth disparity of 6-25 cm where water depths were shallower than 1.5 m” confusing sentense

5.      24-25/1: “however, varying depth, turbidity, bottom properties and tidal influences impact depth 24 accuracy and wholeness of data sets.” Confusing sentense

6.      Give prominence to the main results and conclusions in abstract

7.      29/1: “enables information retrieval of objects” what’s mean

8.      29-31/1: “Remote sensing … ” confusing sentense

9.      32/1: “where”? and?

10.   33/1: “EM energy reflected” should be “reflected EM energy”

11.   33-35/1: “Active systems …” confusing sentense. Rephrase

12.   The language is bad, the sentenses are too confusing to understand

13.   116,159/5: the table had better to design according to the style of the three-line table. the following three horizontal lines: One under the title, above the column headings; One between the column headings and the body of the table; One at the bottom of the table. 

14.   142: “and the mean height was 0.19 m at 1mean sea level (MSL).” What does this sentense mean, the mean height of what, more specifical,. Figure 9 uses a median height bias, it’s best to unify the error parameters.

15.   145: the statement “was 0.12 m MSL” is confusing

16.   Figures with sub-fiugre missing the number

17.   Figure 5 missing the Area-2

18.   237: in equation (2), what’s the n1, n2, n3, …..n9

19.   state the process of system calibration in detail

20.   269: “is created by (Surface 1, Figure 7 b) by defining” confusing

21.   In the results, taking ” the mean height” as the typical parameter to compare the error is not reasonable for coastal wave levels. You should calculate the significant wave height

22.   465-468,515: The error parameters of the Snar depth (RMSE) and ALB depth (Standard deviation) are inconsistent in this paper. the measurement accuracy of them cannot be compared directly.

23.   Figure 11, describe it clearer and more detailed.

24.   The contents of the results are not divided clearly. the titles and contents of the results section can correspond one-to-one with the research questions

Comments on the Quality of English Language

the quality of description needs to be improved. Many confusing sentenses distract the reader from the main points of the study.  I provide some examples of parts needing revision below, but the authors should revise the whole manuscrip. 

Author Response

Please see attachment. Thank you for your review and your comments.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript is unacceptable.

*  1)    In this lagoonal setting, correction of measured water depths for tide stage is essential. How, and in some cases whether, tidal corrections were made is largely unclear.

Green lidar measures depths by the time lag between returns from the water surface and the bottom.  Depths vary with time (tides!), so to compare lidar depths with lidar depths from an adjacent swath, depths from SDB, or depths from sonar, they must be tide-corrected:

tide stage – depth = bottom height

I cannot find a clear statement in the manuscript that this was done for the green lidar data.   Or, perhaps even better, one can take advantage of the absolute water-surface height obtained by the lidar survey and subtract lidar depth from that instantaneous, at-that-point height: 

lidar-determined water surface elevation – depth = bottom height

In general, in a tidal setting, comparison of survey results (sonar to ALB  to SDB) should be done in the height domain, not the depth domain.  One needs to talk about bottom elevation, not water depth.

22)     Inexplicable differences in data tables from version 1 of this manuscript (Table 2) to this version (Table 1) raise questions about the reliability of all reported data.

33)     Figures are out of sequence, misnumbered (two figures 2, two figures 5, no figures 10, 11, or 12), in some cases not referenced when they should be, and in at least one case (Figure 11) missing.  Color ramps for figures 5 and 6 are poorly chosen and do not convey essential information. Figures 2 (the second one) and 3 (p. 17, 18) need 1:1 lines drawn on them so that readers can better evaluate significance of the observed polynomial fit.

44)   The manuscript is poorly organized and poorly written.

Below are incomplete comments on parts of the manuscript.

==============================================

 

1^st 3 paragraphs are essentially unnecessary boilerplate.  What, beyond the references, is there that a reader could not get from a remote sensing textbook or Wikipedia?

“Airborne lidar bathymetry technology”  delete technology

P 2, bulleted list at end: 

·        What is locational turbidity?

·        “conducting reflectance analysis” might be more clearly stated as “quantitative analysis of color in satellite images” or something similar

·        Last paragraph of page should probably be part of list

P 3, sec 2.1.  1^st sentence should reference Fig 2. (Which should be 1^st figure). 

P. 3, sec 2.1, last sentence of 1^st paragraph: “lagoon water differed…”  Differed from what?

P. 3, sec 2.1, last sentence of 2^nd paragraph: “In deeper, smaller, or choppier water”  (delete smaller)

P. 3:  Discussion of sonar methods should be in a separate paragraph, perhaps even a separate subsection.  

·        State that sonar unit was mounted in a kayak (cf version 1)

·        Was position of sonar sensor relative to water surface measured with the operator in the kayak?

·        Was sonar unit calibrated, or were measurements adjusted, for slightly faster hypersaline water? Or for water temperature?

·        I’m dubious about the lack of “rotational adjustment”.  I can’t paddle a kayak without rolling it.  10 degrees of roll = circa 2% increase in apparent depth. Or was the kayak stationary when all measurements were made?

·        Were depths from sonar (table 5, appendix B, 2^nd Figure 2 on page 17, Figure 3 on page 18)  tide-corrected? How? Please clarify this!

 

Section 2.2. Needs discussion of

·        How lidar depths were corrected for tide stage

·        Comparison of 3 different measures of bottom elevation:

o   NIR elevations where measured at low tide

o   tide-corrected green laser depths

o   green-laser depths hung from lidar-measured water surface elevation

or a clear statement that one or more of these data sets was ignored

 

p. 3, last paragraph.  What is the datum for TCEQ SWQM station depths?

P. 3. End of 1^st paragraph.  “Lagoon water differed in quality and depth.”  Differed from what? By how much?

P. 3, 2^nd paragraph, 3^rd line.  Are water transparency and turbidity different things?  Unless I am missing something, you measured turbidity and it is functionally equivalent to transparency

p. 4, Table 1.  Observed tidal heights differ from those reported in version 1 Table 2. Differences are incompatible with a simple datum shift (e.g., MLLW – MSL). Which table is wrong, and why?

P. 6, near bottom. Cite source for typical salinity and temperature.  How much variation is there over the course of the year?  Are the cited values appropriate for May 2017?

P. 12, sec 4.1. ALB system calibration:  please define “slant range” explicitly, or do not use the term.  “Detection capability differences” is not range walk. Figure 5 (referenced at end of paragraph) is mis-referenced—probably the reference should be to Figure 9

p. 9. Section 2.4 belongs with ALB discussion (e.g., after section 2.2), not after “2.3. Satellite imagery”

p. 10, Section 2.6,  “The satellite-derived bathymetry concept includes active and passive spaceborne sensors to measure the bottom of shallow and transparent water column.”   What active sensors?  Discussion here is entirely about passive measurements.

P. 16, sec 4.3, 1^st paragraph.  “Prior to comparing lidar depths (dL) to sonar measurements (dS), we adjusted the depthssurface elevations with observed tides using…”     What “surface elevations”?  Were tidal corrections made for both lidar and sonar measurements? 

P. 16, 2^nd paragraph: “The findings produced a polynomial fit that corresponded to the least-squares method minimizing the variance of the unbiased estimators of the coefficients, under the conditions of the Gauss-Markov theorem [42].”   What findings? 

p. 17, top: “wavelength) revealed surface height differences of 0.84 m (lowest = -0.28 m MSL, highest = 0.56 m MSL). The mean surface height was 0.04 m MSL, and standard deviation was 0.18 m MSL throughout the data acquisition campaign. Surface heights variation findings were was in line with the 0.9 m tide range that was observed at the Port Isabel gauge, in May 2017”

p. 17, 1^st complete paragraph: 1^st sentence is NOT a topic sentence for remainder of paragraph. And figure 11, referenced in the last sentence, seems not to exist.

P. 18, Figure  4.  What is meaning of “Return count” vertical axis?

Figure 2 (p. 17) needs a 1:1 line plotted on it for reference. Or commensurate axis ticks so the reader can plot their own 1:1 line.

P. 18, Figure 3 also needs a 1:1 line for reference.  And I found no text reference to this figure, nor any reference to figure 4.

P. 19, Figure 5.  Please use a color ramp in which the maximum and minimum values can be easily distinguished.  And invert the legend, so that higher elevations are up and lower elevations are down. 

P. 19, Figure 6.  Colors on image and legend do not appear to correspond. What depths do drab green and off-white in the image indicate? Why are there no depths in the lagoon greater than 0.75 m?

Comments on the Quality of English Language

See brief comments above

Reviewer 3 Report

Comments and Suggestions for Authors

The authors have made sufficient modifications according to the previous comments, and I recommend the publication of this manuscript.