Sentinel-2 Recognition of Uncovered and Plastic Covered Agricultural Soil

Round 1

Reviewer 1 Report

Dear authors,

the article "Sentinel-2 recognition of uncovered and plastic covered
agricultural soil" is well written and very interesting for the journal readers. I have just few remarks:

Lines 127 - 130: concerning the first sentence It's not clear why did you use a cloudiness level if then you applied a pixel-based approach for the cloud mask. The second sentence is unclear too.

Lines 230 - 233: I suggest to explain here what the anomalies are and not in the Results section 3.4.3

Lines 239 - 240: Please add a reference for this threshold. 

Figure 5: the colour of the 'uncovered bare soil' in the legend seems different than in the plot. 

 

Author Response

We thank reviewer 1 for the appreciation.

Lines 127-130: We retrieved the scenes on the basis of cloudiness levels as a pre-selection criterion. This increases the algorithm efficiency and reduces the concerns of errors in cloud and shadow detection for highly clouded scenes. The sentence has been moved to line 147 in the section on pre-processing.

Lines 230 - 233: These lines have been updated in the text: anomalies have been explained in this section as requested.

Lines 239 - 240: The sentence has been deleted from the manuscript.

Figure 5 has been corrected.

Reviewer 2 Report

The manuscript “Sentinel-2 recognition of uncovered and plastic covered agricultural soil” reported a GEE and R-language based method to extract bare agricultural soils and plastic-covered agricultural soils based on a number of indexes and methods. Although the topic is somewhat interesting and the study might provide regional soil masks for further soil related studies including satellite retrieval of other soil properties, it lacks scientific significance with no new methods or new ideas. The audience are flooded with numerous spectral indexes and fundamental methods (e.g., spectral angle mapping). The thresholds are determined without a strong evidence. It feels like a report or a specific procedure to be followed by others who are interested in developing soil masks. Also, the writing seems to be boring and less attractive.

The subtitle “3.1 Parcel-level data preparation” sounds like data processing.

It is unusual to see a surface reflectance larger than 100% in Figure 10.

In Figure 12a, how did you know the crop types based on Google Earth maps?

Author Response

We thank the reviewer for the constructive comments. We reduced the text that was not relevant and explained the threshold selection approaches. We highlighted the aspects that distinguish our work from other research. For example, we improved the discussion section substantially and added lines 302-312.

We have modified the pre-processing section and included section 3.1 in it.

Figure 10: Values exceeding 100% are possible due to the presence of high specular reflection. A reference supporting this phenomenon has been added to the discussion section. This is an interesting aspect of highlighting extreme values and investigating their causes while such anomalies are typically smoothed out by other approaches aiming to retrieve bare soil.

Figure 12a: The crop type is from the IACS database. We clarified this in the caption.

Reviewer 3 Report

The manuscript describes a methodology of detecting artificial agricultural covers using Sentinel-2 data. The study is very detailed and well written. I appreciated that the authors described the amount of data remaining after each of the processing steps, and the graphics are of good quality and well thought of. In my opinion the manuscript could be published as it is. I would be interested to see in follow-up studies how results change in additional months throughout the cultivation period. 

Author Response

We thank the reviewer for the very appreciative comments. We retained the description of the amount of data remaining after each of the processing steps, while we reduced the text upon other reviewers requests.

 

Reviewer 4 Report

The paper deals with the detection of soil removing covering objects. This has been the main goal of the community in the last 10 years with great evolution since 2018.

The major issue in the work is that has a huge amount of citations and a very weak discussion, with only one page and few citations. This work has to have a strong discussion comparing methodologies and try to show the differences and in what the authors one got different.

 

Other comments:

• in the paragraph of objectives, many sentences are from material and methods.
• - figure 5 and 6 can come together as a, b
• figure 6. Y axis. Reflectance factor?
• figure 9, 10, merge in a, b
• data of table 2 can be inserted in the text and exclude table.
• exclude box of figure 14
• merge figure 16, 18 in one
• merge fig 21, 22
• merge fig 22, 23, 24. use only one legend of colors

Author Response

We thank the reviewer for the comments and have adopted the manuscript according to the suggestions.

We thank the reviewer for appreciating the importance of the topic. We highlighted the aspects that distinguish our approach, especially regarding the preservation of each parcel’s spatial and temporal integrity as well as analyzing extreme values and their underlying causes instead of smoothing them out. Our goal has been twofold: (1) bare soil detection and (2) artificial (mainly plastic) cover detection. We have formulated these objectives very clearly at the end of the introduction and in the discussion.

We removed irrelevant references in the introduction and methodology. We have revised the discussion section thoroughly and referred to relevant publications.

Other comments:

We have moved sentences from the paragraph of objectives to the study area description in the materials and methods section.

We excluded Figure 6 as we considered it unnecessary for the paper.

We implemented all the requested merges of figures.

Table 2 is now part of the text.

The box around Figure 14 has been removed.

Round 2

Reviewer 2 Report

This work is clearly presented, but I insist on my opinion that it is something more related to engineering. The final descision is up to the editors and the scope of this journal.

Reviewer 4 Report

The authors made proper corrections.