Chlorophyll-a Estimation in 149 Tropical Semi-Arid Reservoirs Using Remote Sensing Data and Six Machine Learning Methods
, Iran Eduardo Lima Neto 2, Francisco de Assis de Souza Filho 2, Paulo Alexandre Costa Rocha 3, Jesse Van Griensven Thé 4 and Bahram Gharabaghi 1,*Round 1
Reviewer 1 Report (Previous Reviewer 3)
Comments and Suggestions for AuthorsThe estimation of chlorophyll a in 149 tropical semi-arid reservoirs based on remote sensing and machine learning is a very meaningful and interesting work. However, although the manuscript has undergone many revisions, there are still many problems in it. I suggest that the whole manuscript be revised and improved further before being resubmitted.
1. It is suggested that the introduction should be further condensed and summarized. So far, there are too many paragraphs in the introduction.
2. It is suggested to further condense and summarize the section 2.1. Study site location, which is too redundant.
3. For Figure 1, either compass or latitude and longitude information is sufficient.
4. The language of the entire manuscript needs to be revised and polished by a native English speaker.
5. In the part of research methods, if the methods used are common methods, it is suggested that there is no need to introduce too much, because the current research methods are too redundant.
6. In the section of research methods, it is suggested to use the past tense for expression.
7. In this manuscript, the names of some Figures are too long and need to be further condensed. For example, Figure 4.
8. In this manuscript, some of the figures need further embellishment, such as Figures 5 and 6.
9. In this manuscript, some sampling times are inconsistent with the satellite images used, such as 2015. How is this taken into account?
10. In this manuscript, the tense of the result section is also suggested to use the past tense.
11. In this manuscript, some references have been added to the results. How is this considered? For example, line 516.
12. In this manuscript, the discussion part needs to focus on the research results and further comb out and highlight the innovation points.
13. In this manuscript, the conclusion part needs to be further condensed and summarized, rather than simply repeating the result part.
Comments on the Quality of English LanguageThe language of the entire manuscript needs to be revised and polished by a native English speaker.
Author Response
Please see the attachment.
Author Response File: 
Author Response.docx
Reviewer 2 Report (New Reviewer)
Comments and Suggestions for AuthorsThis study examined various input parameters for modeling Chlorophyll-a (Chla) across 149 freshwater reservoirs in the semi-arid tropical region of Brazil's Ceará state. We employed satellite remote sensing data and ground-truth Chla measurements to study temporal and spatial distributions, particularly influenced by interannual rainfall variability. We explored multiple machine learning approaches to assess performance. Overall, the paper needs improvement in the following areas:
1. The title mentioning machine learning is too broad. The authors should specify the type of machine learning method used.
2. The related work section should be separate, organized by keywords like Machine learning, Freshwater Reservoirs, Eutrophication, to emphasize research motivation.
3. Clarify the rationale for selecting the XGBoost model in Figure 4 and compare it with other models.
4. Table 2 compares with only classical models; more novel hybrid models should be included for a comprehensive comparative analysis and to identify the state-of-the-art (SOTA) method.
5. Incorporate statistical analyses like Friedman and subsequent Nemenyi tests from Figure 2 to highlight methodological uniqueness.
6. Improve image clarity, and reduce excessive references, especially in the experimental analysis section, to avoid unnecessary additions.
Comments on the Quality of English LanguageMinor editing of English language required.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report (Previous Reviewer 3)
Comments and Suggestions for AuthorsThanks to the author's careful revision of the manuscript, the quality of the present manuscript has been greatly improved, but there are still some language problems in the entire manuscript that need further revision and improvement.
Comments on the Quality of English LanguageMinor editing of English language required
Reviewer 2 Report (New Reviewer)
Comments and Suggestions for AuthorsAccept in present form.
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 AuthorsThis paper develops a machine model based on data from the Sentinel-2 satellite to estimate Chla concentrations in 149 freshwater reservoirs. Several machine learning models are trained and tested respectively, and the forward step parameter selection is realized, and the optimal subset of input parameters is determined step by step. The innovation point is that the optimal attribute selection of input parameters is carried out. Here are some suggestions for the article:
1. The study compares different machine learning methods. Is any method improved in the process? What is the basis for selecting these methods?
2. In the subset selection of bands and indices, what is the basis for their combination? Are the results different if you choose another model as the benchmark method?
3. The smaller the value of RMSE, MAE and MAPE, the smaller the model error. Although GMDH has achieved the best results, the index values of the selected machine learning methods are high, can it further reduce the error?
4. The title of this paper is "Chlorophyll-a estimation in 149 tropical semi-arid reservoirsusing remote sensing and machine learning". If this is the title, the inversion results of chlorophyll a can be further analyzed in this paper; If you want to focus on the method, can you make the topic more specific?
Author Response
Please, see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsReview attached
Comments for author File: Comments.pdf
Author Response
Please, see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsGeneral comment:
It is interesting to estimate chlorophyll a in reservoirs based on remote sensing and machine learning. In this study, the data from the Sentinel-2 satellite and the machine learning method were used to estimate the Chla concentration on 149 freshwater reservoirs. However, in my opinion, there are many problems in the current manuscript that need to be revised and explained, so I suggest re-submission after revision.
Specific comments:
1. Abstract should focus, rather than generalize, should give the research methods and results used, as well as innovative points.
2. The introduction of this study needs to be further condensed in order to highlight the theme and logic of the study.
3. The sampling time per year is recommended to be listed.
4. The language of the entire manuscript needs to be polished by a native English speaker.
5. Optical satellite effects are susceptible to cloud effects. How is this aspect taken into account in this study? How is the remote sensing data collected in the study area handled?
6. How are mixed pixels treated in this study?
7. There are many methods and data involved in this manuscript, and it is recommended to draw a technical road map to clearly show the logical relationship.
8. When using machine learning for analysis, is the modeling data set used sufficient to support the research results, and how to assess the uncertainty caused by the data set?
9. In the discussion section, it is suggested to divide the discussion into different headings in order to make the discussion more logical and hierarchical. The current discussion logic is very confusing.
10. In the conclusion part, there is language repetition, which needs to be further modified and perfected.
11. In terms of the paper as a whole, it feels less like a research paper and more like a report, suggesting that various parts of the manuscript be reconstructed.
Comments on the Quality of English Language
The language of the entire manuscript needs to be polished by a native English speaker.
Author Response
Please, see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 3 Report
Comments and Suggestions for AuthorsThe estimation of chlorophyll a in 149 tropical semi-arid reservoirs based on remote sensing and machine learning is a very meaningful and interesting work. However, although the manuscript has gone through the first round of revision, in my opinion, there are still many places that have not been well revised and improved. So I recommend rejection.
1. The present abstract still fails to highlight the innovation of this study.
2. The introduction part of this study is still a little rough, and it does not well summarize the existing research and highlight the innovation of this study.
3. The Study site location part of this study is too redundant, so it is suggested to further simplify it.
4. In Figure 1, since the warp and weft network already exists, a compass is not needed.
5. The language of this study still needs to be further polished.
6. If the method used in this study is an existing method, it is suggested that there is no need to introduce the method too much.
7. The language tense aspect of this study has been proposed in the first round of revision suggestions, but the author ignored it.
8. The title of Figure 4 is too long, it is recommended to simplify.
9. Both figures 5 and 6 need further embellishment.
10. The language tense in the discussion section is suggested to be used in the past tense.
11. In the conclusion, it has been suggested in the first round of revision that further refinement and improvement should be made, but the author has not revised it.
Comments on the Quality of English Language
The language of this study still needs to be further polished.
Author Response
see attachment
Author Response File: Author Response.pdf
