Unsupervised Machine Learning for GNSS Reflectometry Inland Water Body Detection

Round 1

Reviewer 1 Report

The manuscript utilizes unsupervised machine learning and GNSS-R to detect changes in the range and location of inland water bodies, which is a meaningful and interesting research objective. Space-based GNSS-R has high spatial and temporal resolution, which can improve the resolution of inland water monitoring. Three clustering algorithms, K-Means, Agglomerative and DBSCAN, were compared in the manuscript, and the results were given. But there are still some questions that the author needs to answer:

1. Reference error in Section 2.2, i.e., “Error! Reference source not found.!”. Furthermore, Such citation errors still occur in many places of this manuscript.

2.In machine learning method, sample selection has great influence on result accuracy. In this manuscript, the preprocessing of SNR, such as the small-size of SNR(SNR<0) and the high transmitting power of GPS Block IIF satellite data, lead to poor observation data quality. What does the authors think about it?

3. In this manuscript, because CYGNSS has a certain spatial resolution, what is the scale of inland water detection? What width is that? The article did not mention it.

4. Machine learning has strong ultra-nonlinear fitting ability, and the input sample of this manuscript is relatively single. What are the considerations of the selection of sample types?

5. Figure. 13 and 14 are not presented in the manuscript.

6. Figure 7 is two graphs, please add the subgraph numbers (a) and (b); and Figure 3 is the same

7. In table 2, Why only use Cyg2 as an example, and the other subsatellites?

8. The reference format in the reference should refer to the official format. The authors should be listed in full.

Author Response

Dear Reviewer,

We would like to thank you for your effort to detect and note these issues, and we really express our apologies for this inconvenience. Also, thank you very much for the comments that helped in improving the quality of our work. We made all the appropriate changes and to the best of our knowledge similar issues do not appear in the revised document. Apologies again for this inconvenience.

Kind regards,

Stylianos Kossieris

Author Response File: Author Response.docx

Reviewer 2 Report

See attached file

Comments for author File: Comments.pdf

Author Response

Dear Reviewer,

We would like to thank you for your effort to detect and note these issues, and we really express our apologies for this inconvenience. Also, thank you very much for the comments that helped in improving the quality of our work. We made all the appropriate changes and to the best of our knowledge similar issues do not appear in the revised document. Apologies again for this inconvenience.

Kind regards,

Stylianos Kossieris

Author Response File: Author Response.docx