Estimation of Nitrogen Content in Winter Wheat Based on Multi-Source Data Fusion and Machine Learning

Round 1

Reviewer 1 Report

The authors have estimated the crop nitrogen content of winter wheat using UAV data and especially using the MS, RGB, and thermal infrared (TIR) images. The study has used three different machine learning algorithms for predicting N content in wheat crop experimental sites situated in China. The concept is well written and formulated but there are some issues that need to address. The specific comments are given below. Accordingly, a major revision of the manuscript has been recommended.

Major Comments:

1)      Abstract: Line 17-18, needs to rewrite. It says RFR achieved the highest prediction accuracy but also highlights differences in accuracy among the three algorithms. Also, add the Unit of MAE

2)      Introduction is not properly contextualized. For instance, it starts with UAV-based plant N content estimation which does not say about what spectral bands are known to be best for leaf N content. Add some insights into it

3)      In section 2.2.2: in places of resolutions, Write 1280 x 960 pixels. Also mention spatial/ground resolution for RGB, MS and Thermal  images

4)      In section 2.3.2: Instead of VF, one should use existing standard abbreviations like Fractional vegetation cover (FVC) throughout the ms.

5)      Table 3: Are all indices used  in the predicting ML models

6)      Results: Replace sub section (L225) by N content of winter wheat. Everywhere mention in the text about the unit of N content (e.g. 22 mg/g in L228). Also, clarify microgram or milligram.

7)       Also add units of MAE and RMSE throughout the ms

8)      Fig 2: I suggest, also adding a running mean line would be more useful for a, b, and c. 

9)      Clarify what is RLCM used somewhere in ms?

10)  What is the significance of the study? It should be added in the list line to the conclusion

Minor Comments

1.      L254: "obtained by the RGB data is comparable to that of the MS data". Check this as full stop missing

2.      L283: "shown in Figure 9 RFR generally". Check this as full stop missing

3.      Many of the abbreviations are used but not defined. Check thoroughly

Author Response

Dear editor,

Thank you very much for your comments and suggestions. Those comments are all valuable and very helpful for revising and improving our paper, as well as the important guiding significance to our research. We have studied comments carefully and have made correction：

1.Comment: Abstract: Line 17-18, needs to rewrite. It says RFR achieved the highest prediction accuracy but also highlights differences in accuracy among the three algorithms. Also, add the Unit of MAE

Response: The main text has been revised as you have suggested and the unit of MAE (mg/g) has been added

2. Comment: Introduction is not properly contextualized. For instance, it starts with UAV-based plant N content estimation which does not say about what spectral bands are known to be best for leaf N content. Add some insights into it

Response: The introduction has been duly supplemented in the main text. (L41, 42)

3. Comment: In section 2.2.2: in places of resolutions, Write 1280 x 960 pixels. Also mention spatial/ground resolution for RGB, MS and Thermal  images.

Response: The units of resolution and the spatial resolution of the various sensor images have been added in section 2.2.2 as you suggested.

4. Comment: In section 2.3.2: Instead of VF, one should use existing standard abbreviations like Fractional vegetation cover (FVC) throughout the ms.

Response: Already corrected the VF to FVC according to your suggestion.

5. Comment: Table 3: Are all indices used  in the predicting ML models

Response: All indices in Table 3 were used in the prediction model.

6. Comment: Results: Replace sub section (L225) by N content of winter wheat. Everywhere mention in the text about the unit of N content (e.g. 22 mg/g in L228). Also, clarify microgram or milligram.

Response: The title of section 3.1 has been changed from 'subsection' to 'Relationship between CSC and N content of winter wheat'. The mg in mg/g in this paper is milligrams, which has been reflected in the text (L103).

7. Comment: Also add units of MAE and RMSE throughout the ms

Response: Units of MAE and rRMSE have been added in the text as mg/g and %, respectively.

8. Comment: Fig 2: I suggest, also adding a running mean line would be more useful for a, b, and c. 

Response: Thank you for your suggestion, we have changed b and c in Figure 2 to the average of daily climate data according to your suggestion。

9. Comment: Clarify what is RLCM used somewhere in ms?

Response: I am very sorry that I mistyped a GLCM in section 2.3.4 as RLCM, thus causing your misunderstanding, and have corrected RLCG to GLCM in the text.

10. Comment:  What is the significance of the study? It should be added in the list line to the conclusion

Response: The significance of this study has been added to the conclusion section of the text as you suggested.

In the main text, a period was added where it was missing. In addition, the entire text was thoroughly checked and all abbreviations were defined.

Reviewer 2 Report

Dear,

follows my suggestions and considerations for the manuscript.

Comments for author File: Comments.pdf

Author Response

Dear editor,

Thank you very much for your comments and suggestions. All these comments are valuable and helpful for us to revise and improve the thesis. We carefully studied the comments and made changes, including revising the paper title, adding references, defining abbreviations, adding R2 and equations, revising the figure, and adding content.

Round 2

Reviewer 1 Report

After reading the revised manuscript, I find that the authors have addressed all comments and queries. The manuscript has improved significantly.

Reviewer 2 Report

Dear Authors,

All corrections, suggestions have been made.

Congratulations on the final acceptance of the paper.