Satellite Multi/Hyper Spectral HR Sensors for Mapping the Posidonia oceanica in South Mediterranean Islands

Round 1

Reviewer 1 Report

While there are a number of relatively minor issues that could be addressed to improve the quality of the manuscript, the primary issue is that the formatting of tables, figures, and their captions is so poor that entire sections of text are unclear, illegible, or entirely hidden. The manuscript must be reformatted and resubmitted before a comprehensive evaluation can be made as to the quality of the work.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Thank you for giving me the opportunity to review this paper.

In my opinion this work is very interesting as it performs a very comprehensive study to map the state of oceanic Poseidonia in coastal habitats, using high resolution multi- and hyperspectral sensors by using classifiers some advanced machine learning algorithms with different supervised regression and classification models. In my opinion it is a great work.

However, the manuscript is not yet ready for publication but has a lot of promise after minor revisions:

• Line 228-289 The text refers to Figure 6, I understand this is an error because it seems to refer to Figure 3.
• Line 616-617 Figure 5 appears to be incorrectly referenced and Figure 7 referred to in the text does not exist.
• Line 643 figure 8 there is text superimposed at the bottom of the figure.
• Lines 649-650. This paragraph seems to be erroneous, it appears as figure 5 when it is under figure 10.
• Lines 655-656 The caption of figure 10 is not below the figure.
• Lines 665-679 Poorly laid out. The caption of figure 11 is superimposed on table 3.
• Lines 712-719 Layout problems in figure 12.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

The paper is well written and presents interesting and novel methods and results. Most of the shortcomings and my comments relate to the visual design of the paper. The authors should go back over the paper and fix all the minor flaws they have and think about using Latex next time.

1. All figure and table captions should be consistent; some have a hyphen, some do not; some have spaces around the hyphen, some do not.
2. Figure 3 - strings with geo. coord. system are not justified and slightly blurred.
3. Figure 4 - subfigures are not aligned and are not the same size.
4. Figure 5 - caption should be on the same page 
5. Referencing of figures and tables should be consistent; parts of the text refrence floats as "figure X" and others use the short reference "fig. X".
6. Figure 10 - the caption is not directly below the figure.
7. Figure 11 - same as figure 10, but now the caption is over the table.
8. The date format in the paper should be consistent. Sometimes authors use a slash and sometimes a hyphen; see lines 682 and 697 for examples.
9. There is text around the caption of Figure 12.
10. The caption of Table 4 is over the table and the text is interwoven with the table.
11. Table 5 - MCLA column should be moved to the left side of the results.
12. References 1, 51 and 52 are missing, reference 42 has a blue number.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 4 Report

Dear author,

I have reviewed your article. You can find some of my suggestions and corrections in the attached file. I wish you success. 

Best wishes

Comments for author File: Comments.pdf

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors describe an attempt to use EO data to map seagrasses surrounding recently installed energy production infrastructure. While some of the methods for satellite data processing are appropriate, much of the study misapplies statistical techniques to derive information. Figure 10 should have been a glaring sign that the data are incompatible – one set is categorical and the other continuous. Such data cannot be statistically evaluated as done and produce meaningful information. Most of the figures and some of the tables are very poorly done and are not suitable for publication even after the initial revisions. The authors should consider recruiting a graphic designer or someone experience with making geospatial figures to improve, and should review manuscripts with similar figures to know what elements to add or remove. There is some merit to the work, but the writing and display make it very difficult for the reader to extract meaningful information.

The writing could benefit from review and editing for English language grammar and syntax. Additionally, the writing style is fraught with run-on sentences that needlessly make it difficult to comprehend. This reviewer suggests revisions throughout for unnecessarily long sentences with multiple clauses.

Detailed comments:

Much of the Introduction section belongs in the Methods section. Some of the information is repeated in the Methods section and should be deleted to avoid unnecessary redundancy.

Figure 1 – Remove one of the two north arrows and add a graticule to indicate latitude and longitude of either the upper or lower map.

Figure 2 – The legend is illegible. Change the label name, explain the units, and reorient so that it is contained within the image borders or remove it altogether.

Figure 3- The legend needs revision: Only 3 bathymetry contours are shown, so delete the others, and remove the first line, which does not appear to correspond to a bathymetry number; the color of the sampling stations polygon does not match the numbered boxes in the image; and the background element is entirely unnecessary. The “x-axis” of the sampling schema appears out of place or a mistake. If the authors intend to indicate distance between plots, they need to rework this element of the figure. What does “d” mean?

Line 325 - The procedures mentioned must be clearly described. There is no description for one to follow.

Figure 4 – What are the colors in the lower right panel showing? A legend or caption description is needed.

Figure 5 – Suggest removing the PRISMA section from the legend as unnecessary.

Line 386 – Section is misnumbered. Revise and review all others for consistency.

Section 2.3.2 – This section is robustly described, but the details are relayed in an unclear manner that is occasionally hard to follow. Suggest adding one or more figures to diagram this section of the methods.

Figure 9 – PRISMA is not necessary in the legend.

Figure 10 – The vertical nature of the plots indicates that the authors are effectively plotting categorial data against continuous data, and any regressions resulting from such an analysis are meaningless.

Results – why are no p-values or other measurements of statistical significance reported? One would assume, based on the previous comment, that these results cannot be statistically significant given the misapplication of regressions to incomparable data types.

Comments for author File: Comments.pdf

Author Response

The manuscript was reviewed in depth following the suggestions and remarks provided, in particular, 

• the section numeration was corrected;
• the redundancies and inessential parts of the method section were removed;
• all the figure improvements and related caption modifications suggested have been implemented;
• a methodology scheme was added and described;
• a better explanation of sea truth collecting and processing was introduced;
• the manuscript was improved for English language grammar and syntax;

Author Response File: Author Response.pdf

Round 3

Reviewer 1 Report

While I appreciate the authors' effort to revise weaknesses in the manuscript, I still have serious concerns about the foundational application of remote sensing techniques and statistical analyses. From the results and conclusions provided, it seems clear that the authors do not have substantial experience working with remote sensing data or statistical analyses. While some superficial errors were corrected, their nature and that of remaining issues suggests that there is not enough fundamental expertise in the relevant fields to warrant publication of resulting research. For example, Figure 10/11 indicates that the data are effectively, if not exactly, inconsistent (i.e. continuous vs discrete) perhaps due to sampling precision inadequacies or otherwise, and thus the results are effectively meaningless. Additionally, the added statistics do not reveal the information necessary to determine if the results are indeed significant. It appears that the authors did not actually derive a p-value, but instead an F statistic and t-test, yet attempt to pass off the results as including p-value significance. This is the kind of egregious error that suggests the authors do not have the expertise to derive relevant information from the data. I suggest revising the entire research endeavor to include experts in statistics and remote sensing. Additionally, the figures and tables are still comically unprofessional with elements partially cutoff, unnecessary legend elements, and illegible table headings, among other issues. I'm sure this evaluation appears harsh, but that is not this reviewer's intention. I see value in the intention of the work and wish to help the authors identify the fundamental flaws in their analyses to produce a valuable contribution to the field.

Author Response

Even though the manuscript has been extensively modified with all the changes previously requested and following suggestion indicated in the previous round of the review, further generic remarks still refer to figures (even jet improved as indicated) and tables. In the perspective of further improvements suggested, the Figure 15 and 16 were enhanced with legend correction and resolution increase. Most of columns heading in tables are explained in the text and the description of the remaining was introduced in the captions.  In addition, the references were integrated following the reviewer indication.

The reviewer is reminded that the tables of Fig. 11 are obtained from the statistical analysis of the data shown on the graph and, as clearly stated at lines…., include the required p-values parameters with their alternative denomination of Prob <| t | (p-values) and Prob <F (P-value), derived respectively from t and F statistics.

I can partially agree on the remaining weakness due to the limited sea truth data (focused mainly on the ISWEC area) but, although the LAI-c measured at sampling stations look as discrete, they derive from plot’s average at station level and are characterized by decimal values typical of a continuous distribution. In any case the highlighted and reported p-values (< 0.05) confirm the confidence level ( > 95%) of the regression models  assessed.