Performance and the Optimal Integration of Sentinel-1/2 Time-Series Features for Crop Classification in Northern Mongolia

Round 1

Reviewer 1 Report

Comments regarding manuscript „Mapping of 10m crop types in northern Mongolia based on 10- day interval features of Sentinel-1/2”.

Generally, the manuscript is well written and comprehensive and concerns new methods of mapping crops in agriculture.  However, there are some comments. Provide abbreviations and unite to all quotations. Description of M-index (from l.319) is a part of Methodology not Results. Limit self-citation to 5 references.

Author Response

Dear Reviewer, we appreciate valuable comments and suggestions about this paper, which significantly help us improving the quality. Please see the response document.Thanks a lot. 

Author Response File: Author Response.pdf

Reviewer 2 Report

File attached

Comments for author File: Comments.pdf

Author Response

Dear reviewer, we appreciate valuable comments and suggestions about this paper, which significantly help us improving the quality. The responses can be checked from annex.Thanks a lot. 

Author Response File: Author Response.pdf

Reviewer 3 Report

Mapping of 10m crop types in northern Mongolia based on 10- day interval features of Sentinel-1/2

This manuscript applied the Random Forest (RF) classifier and Google Earth Engine (GEE) cloud computing for crop type mapping using Sentinel-1 (S1) and -2 (S2) imagery. In this research, the highest overall accuracy of S2 (93%) is slightly higher than the joint S1 and S2 (92%) classification. Also, the study result revealed that the highest accurate crop type mapping could be achieved from 150 DOY (11 May) to 260 DOY (18 September) using S2 data.

Although the topic of the manuscript is of wide interest in remote sensing community and the research design is innovative, some parts of the manuscript must be improved.

Review summary

While the topic of this manuscript is in principal suited for Remote Sensing, I am seeing several serious methodological flaws in this study. My major concerns are:

• Title: its interpretation can be very confusing, do crop types have dimensions of 10 m? Also, add 'imagery' after Sentinel-1/2?
• Figure 4: is it really necessary to show e.g., thumbnail in the upper right corner that is really hard to read (I suppose that it shows list of locations)
• LN 151: please convert 3:2 ratio to 60:40
• Section 2.3. – overall throughout the manuscript grammar and spelling must be checked, 'its processing' is very strange formulation, perhaps rename it to Proccesing of the satellite data
• Section 2.3.1 – are Level-1C S2 imagery used for crop type classification or were they converted to Level 2A? Furthermore, how were 20 m spatial bands resampled to 10 m?
• LN 184: Start of the sentence sound very strange :“ Swir1 and Swir2 reflectance bands, Swir1 and Swir2, has a good ability to separate crops.“
• LN 194 – 199: please add Equation numbers and e.g. Nir in LN 198 is wrongly spelled, once more, make a grammar revision of the manuscript
• LN 235: what does natural domain mean?
• Caption of the Figure 5: according to the figure caption, this framework is suitable for use only in 2018?
• LN 249: future and past tense in one sentence about methodology used in the manuscript?!
• LN 252: mentiones Equation which are not properly introduced in Section 2.3
• LN 264: a total of 198 metrics, in Title they are names as features, please adjust the term throughout the whole manuscript
• Again LN 302 – 305, Equations are not properly introduced, and terms (e.g., Y_ij are not explained). Also, the Authors mention that five indices are listed, however there are OA, UA, PA and F1-score showed
• Figure 6 caption sounds very strange/unfinished. Also, what does grey buffer zone around each crop mean?
• Figure 9: again, is it really necessary to emphasize the year of the crop classification? Furthermore, how was weight contribution in Section 4.3.2 calculated?
• What is a meaning of rapid separation in the end of the Conclusion: „Regarding crop type mapping, the study showed the rapid separation between spring crops in (..)“
• Also, the research compared S1, S2 and S1+S2 classification scenarios, however I did not find accuracy regarding each classification scenario

I think that the research design of this manuscript is good, but still some additional changes need to be made, in order to be published in this journal. Section Conclusion needs a minor revision since it needs to be: „In the Conclusion section, state the most important outcome of your work.“ In a current form, this statements lack in the manuscript. Once again, grammar/spelling and Figure caption need to be revised throughout the whole manuscript

Author Response

Dear reviewer, we appreciate valuable comments and suggestions about this paper, which significantly help us improving the quality. Please check it from annex. Thanks a lot.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

I believe the authors have adequately addressed the comments.

Author Response

I believe the authors have adequately addressed the comments.

Response: Thank you for your appreciation.

Your comments and suggestions significantly improved the quality of this work. Thanks a lot.

Author Response File: Author Response.docx

Reviewer 3 Report

Performance and the optimal integration of Sentinel-1/2 time-series features for crop classification in Northern Mongolia

The aforementioned manuscript has improved from version_01 to version_02. However, I still have some minor concerns, and they are listed, as follow:

• Figure 6: what is an amplitude / how it has been calculated?

Figure 8: still, I think that there should be a Table with listed OA and Kappa values for three classification scenarios (perhaps mean OA and Kappa, or OA and Kappa for DOY 300)

• Section 4.3.2: the Authors have noted which function calculated the contribution of the 96 input features, but what is that measure for RF classifier – MDA, MDG,..?

The manuscript has improved between the two versions, but still in some parts of the manuscript English language and style needs to be corrected.

Author Response

Thank you for your valuable comments and suggestions. We carefully check and revise the points you mentioned. I hope the revision addresses your concerns. Thanks.

Author Response File: Author Response.pdf