Mapping Tropical Forest Cover and Deforestation with Planet NICFI Satellite Images and Deep Learning in Mato Grosso State (Brazil) from 2015 to 2021
, Griffin Carter 1,3, Alison L. Ritz 3,4, Mayumi C. M. Hirye 5, Jean P. H. B. Ometto 6 and Sassan Saatchi 1,2,3Round 1
Reviewer 1 Report
Review Report
The manuscript entitled “Mapping Tropical Forest Cover and Deforestation with Planet NICFI Satellite Images and Deep Learning in Mato Grosso State (Brazil) from 2015 to 2021” presents an approach to estimate the Spatio-temporal variation of forest cover using remotely sensed data and advanced geospatial computation applications. A detailed review of the manuscript is described below.
The abstract need to be modified, there is no flow in the abstract.
The introduction required more technical information with respect to the recent technology, and data that can be used for forest monitoring.
What is the significance and novelty of the study? Which is not clearly presented.
Figures should be modified; the quality should be improved.
Methodology: The methodology should be modified and substantially improved, classification process, methods for forest extraction site selection for sampling, etc are not clearly discussed
The result is not well discussed properly, there must be qualitative and quantitative information about the findings.
Discussion must be improved with the following contents: Why these models can be used for a similar study, and how the method is significant from other methods. How the regional configuration controlled deforestation.
Why reference in the conclusion? The conclusion should be your view and summary of the work,
There is a certain confusion in the manuscript flow.
I recommend a substantial modification in the manuscript for possible publication.
Author Response
Thank you very much for your review. Please find my response in the attached pdf.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
In the study, the tropical tree cover and deforestation were mapped using U-net model and Planet NICFI satellite images. In addition, the advantages of the methods used in the study and the differences between the results and other products were discussed. However, the language expression of the paper is not very well, and some format is WRONG, which largely affects the readability of the paper. The specific suggestions and issues are as follows:
1. The abstract should be rewritten, which the main contents and results of the research are not well condensed.
2. The images used were from 2105-12-01 to 2022-06-01, but the results had not involved any information about 2022. WHY?
3. In the ‘Introduction’ section, it is stated that the Planet dataset from December 2015 to March 2022, a total of 85056 images. However, in the ‘Planet satellite images of Mato Grosso - Brazil’ section, it is stated the Planet images were from 2105-12-01 to 2022-06-01, also 85056 images. It should be checked.
4. The reason why the U-net deep learning model was chosen, and not other models should be further elaborated.
5. The content of the '2. Materials and Methods' should be shorten largely.
6. An overall flowchart is suggested to be added. Now, only the architecture of U-net model was illustrated which was unable to fully demonstrate the research technology process.
7. Sections 3.1 and 3.2 suggested to be merged.
8. The caption of all tables should be placed before the table rather than after it.
9. The section of the ‘Conclusion’ should be rewritten, and the reference should be deleted.
10. The language must be further improved so that so that the readability of the paper is enhanced. And, the overall content of the article needs to be further condensed to highlight the theme.
Author Response
Thank you very much for your review. Please find my response in the attached pdf.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Revised Manuscript entitled “Mapping Tropical Forest Cover and Deforestation with Planet NICFI Satellite Images and Deep Learning in Mato Grosso State (Brazil) from 2015 to 2021” presents an approach to estimate the Spatio-temporal variation of forest cover using remotely sensed data and advanced geospatial computation applications. The manuscript has substantially improved the quality and I recommend accepting it in its present form.
Author Response
Dear Reviewer 1,
Thank you very much for accepting our manuscript for publication and for the time you took to consider our work.
Best regards,
Fabien Wagner
Reviewer 2 Report
The research has scientific significance, and the quality of the paper has been improved after modification. However, there are still redundant information in the methods section. And some unclear title expression presents in the results and discussion sections. It is suggested to refine each title and method sections. Moreover, the paper has written errors, need to be carefully checked and revised.
Author Response
Please see the attached pdf
Author Response File: Author Response.pdf
