A Novel Algorithm for the Retrieval of Chlorophyll a in Marine Environments Using Deep Learning
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsDear Editor,
Water
This paper proposes a universal method for Chla inversion in coastal waters, which combines 1D CNN and other traditional machine learning algorithms to establish a relationship model between remote sensing reflectance (Rrs) and Chla concentration. The authors use the original Rrs as input features to predict Chla concentration and demonstrate the performance of the model. Through comparison with other algorithms, the authors verify the high accuracy of the model in coastal waters with different nutrient levels. The subject addressed is interesting and within the scope of the Water. Also, the novelty of the manuscript is acceptable. Nevertheless, some minor revisions have been found. After carefully solving the following issues, the manuscript could be a possible publication.
1- What is the importance of retrieval of رhlorophyll based on satellite images and how does it help to reduce its harmful effects. More explanations about this issue should be added in the introduction.
2- The caption of Figure 2 and Figure 3 are the same. Please correct caption of Figure 3.
3- Are in equations of R2, MAE and bias logarithm of data used? Please recheck it.
4- In Figure 5, are logaritimic data showed or orginal data?
5- Which software package is used in this study? Please state it in more details.
6- The quality of Figure 9 and 11 are low.
7- The main quantities results should be added in the conclusion.
8- How your model helps to reducing eutrophication or algal blooms? Please state it in the conclusion
Considering the mentioned points, this study in the current version needs major revisions.
With kind regards,
Author Response
Dear reviewer,
Thank you very much for your suggestions, which will greatly help to improve the quality of this article! For details, see the cover letter and revised draft.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for Authors1 In your introduction, please list and cite more other works on global chlorophyll with similar statistic and machine learning models. RFR, SVM, and one or two levels CNN are prevailing models.
2 In your result section, you displayed the error analysis for us which is applauded. However, the methodology of how you did this is not introduced in the methodology section. Please complete this gap.
3 In line 270, there might be a typo. At the beginning of this paragraph, I suppose the Figure 9 reflects the errors, not the figure 8. Please check...
4 Figure 7 is redundant. You can easily mark these two boxes on any other figures, like 1,2,6,8.
5 For the results shown in Figure 10, the current conclusion looks pretty arbitrary and subjective. Do you have any quantitative data to confirm the smoothness of the CNN/SVM method? Well, the danger to talk about the smoothness as a judge for a model's performance is that in nature, chlorophylls in some adjacent areas are indeed largely different. I don't mean your result is wrong. Just leave a comment here and it is up to you to remain unchanged or add more discussions.
6 How did you get the Figure 11B? Please have a full explanation as the figure legend. And add the methodology narration to the methodology section. This is a similar suggestion to my comment #2.
Author Response
Dear reviewer,
Thank you very much for your suggestions, which will greatly help to improve the quality of this article! For details, see the cover letter and revised draft.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsPlease see the attached PDF file.
Comments for author File: Comments.pdf
Author Response
Dear reviewer,
Thank you very much for your suggestions, which will greatly help to improve the quality of this article! For details, see the cover letter and revised draft.
Author Response File: Author Response.pdf
Round 2
Reviewer 3 Report
Comments and Suggestions for AuthorsThe questions are addressed.
