Multi-Temporal Analysis of Cropping Patterns and Intensity Using Optical and SAR Satellite Data for Sustaining Agricultural Production in Tamil Nadu, India

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This study analyzed the cropping patterns and intensity using optical and SAR satellite data for sustaining agricultural production in Tamil Nadu, India. This study is interesting and well organized. I have some suggestions for authors:

(1) In this study, there is a significant difference in spatial resolution between MODIS and Sentinel-1 data. Why was processed Sentinel-2 data from GEE not considered? Additionally, the impact of the spatial resolution disparity on the results in terms of uncertainty needs further exploration. It is recommended that these aspects be addressed and discussed in greater detail within the discussion section.

(2) The introduction section only discusses the importance of the topic, but lacks relevant literature on the content and methodology. Please include the necessary references to strengthen the discussion of the methods and existing research.

(3) Line 164: What is the rationale behind the selection of the threshold (-13.3)?

(4) Line 264: Why dived the CI values into six categories? Some references are needed to support this process.

Author Response

Responses to the Reviewer Comments

Thank you for reviewing our manuscript 

1. The reasons for the utilization of the MODIS datasets have been included in the discussion section of the manuscript. Additionally, it should be noted that the North East Monsoon coincides with the crop growing period in Tamil Nadu, resulting in most optical data being impacted by cloud cover. As a result, 16-day composite MODIS data was used to compare and reduce cloud cover. Furthermore, the impact of the spatial resolution disparity on the results, in terms of uncertainty, has been emphasized and discussed in detail within the discussion section.
2. The relevant literature has been incorporated to support the content in both the introduction and discussion sections.
3. The reason behind the threshold selection and the limit selection were indicated in the materials and methods section.
4. Further, the reference to support the discrimination of the cropping intensity values (CI) has also been added appropriately.

Reviewer 2 Report

Comments and Suggestions for Authors

The article presents an analysis of crop cultivation patterns and the intensity of cultivation in the Indian state of Tamil Nadu. The research was conducted using satellite imagery collected from optical and radar sensors over a period of five years (2019-2023).
Key findings of the study include:
- The area under crop cultivation varied between 52.88 and 55.85 million hectares based on Synthetic Aperture Radar (SAR) data, and between 56.87 and 62.07 million hectares based on Moderate Resolution Imaging Spectroradiometer (MODIS) data.
- Single-crop cultivation was the dominant form of agriculture, occupying approximately 51% of the total cultivated area. Double- and triple-crop systems accounted for 31% and 17%, respectively.
- MODIS data tended to overestimate cultivated areas compared to SAR data, although differences in spatial resolution did not significantly impact the analysis results.
This study emphasizes the importance of utilizing satellite data for monitoring and managing agricultural resources in Tamil Nadu and other regions.
There is practically no literature review. It was necessary to specify the following. The launch of the Sentinel-1A/B and Sentinel-2A/B satellites with a periodicity of 3-12 days and a spatial resolution of about 10 m, coupled with free access to data, created unprecedented opportunities for monitoring the Earth's surface. Based on these data, various digital online platforms for remote monitoring of anthropogenic and natural landscapes, including agricultural ones, such as Sen2-Agri, have been developed (http://www.esa-sen2agri.org /), One soul (https://onesoil.ai /) and others. Platform Sen2-Agri involves the use of optical multispectral satellite imagery Sentinel-2A/B and Landsat-8 for monitoring (Defourney et al., 2019). The free precision farming platform Onesoil uses Sentinel-1A/B radar images in addition to Sentinel-2A/B multispectral sensors, which makes it possible to refine information. The emergence and development of such online services using available satellite imagery presupposes the corresponding development of new methods and technologies of analysis and remote sensing data processing of optical and microwave wavelength ranges. Please see next - Khabbazan S., Vermunt P., Steele-Dunne S., Ratering Arntz L., Marinetti C., van der Valk D., Iannini L.,
Molijn R., Westerdijk K., van der Sande C. Crop monitoring using Sentinel-1 data: A case study from the Netherlands // Remote Sensing. 2019. V. 11. No. 16. id. 1887. 8 p.

It is somewhat surprising that the data is compared with such different spatial resolutions. Therefore, it is necessary to more fully justify the expediency of such a comparison. Why not compare Sentinel-1 data with Sentinel-2 or Landsat optical data??

Author Response

Responses to the reviewer comments:

Thank you for reviewing our manuscript. 

1. The literature review specified by the reviewer has been included in the Introduction section.
2. Despite the coarse resolution of MODIS datasets, they provide results comparable to Sentinel-1A data while being significantly less demanding regarding computational resources. Additionally, the North East Monsoon coincides with the crop growing period in Tamil Nadu, which results in most optical data being impacted by cloud cover. Consequently, the 16-day composite MODIS data was utilized to mitigate this issue, ensuring reliable comparison.
3. The indicated references have also been added to support the utility of the Sentinel 1A datasets over the multispectral datasets in the discussion part.

Reviewer 3 Report

Comments and Suggestions for Authors

Please see the attachment.

Comments for author File: Comments.pdf

Author Response

Responses to the reviewer comments 

Thank you for reviewing our manuscript 

1. The methodology used for the threshold-based classification is elaborated in detail in the methodology section. Additionally, the causes for the discrepancies in the datasets are speculated and addressed in the discussion section.
2. The literature review has also been extended.
3. The season-based composites were utilized for both the SAR data and MODIS data processing. Based on the time series spectral curves derived from these datasets, images corresponding to the peak vegetative stages were selected for further processing to determine cropping intensity and patterns. This approach was applied to classify croplands into three types and to separate them across different seasons. The thresholds for classification were set by analysing the spectral responses of various crop types during their peak growth periods. The same methodology has been clearly detailed in the methodology section of the manuscript.
4. The middle images were selected based on a spectral curve fitting procedure, wherein the peak vegetative period corresponding images were utilized for further processing of cropping pattern and intensity.
5. The lines 173 and 248 indicate that summation of the kharif, rabi and summer classified images for determining the net cropped area, wherein the zero pixels indicating the non -cropped area/fallow will be removed.
6. Considering the scope of the study and not to increase the numerical components in the manuscript, the cropped area for all three seasons in each year was not included in the main script. You can find each season's district-wise cropped area in the supplementary materials part.
7. The discussion section has been realigned to have a detailed comparison and the reasons for comparing MODIS and Sentinel 1A.

Reviewer 4 Report

Comments and Suggestions for Authors

This manuscript primarily utilizes multi-temporal datasets from optical (MODIS) and SAR (Sentinel-1A) sources to map cropping patterns and assess cropping intensity across multiple agricultural seasons in Tamil Nadu, India, from 2019 to 2023. The findings provide valuable insights into the use of optical and SAR datasets with varying spatial resolutions to evaluate the changing dynamics of cropped areas. Nevertheless, the manuscript falls short in terms of originality and scientific depth.

1. It remains unclear whether the study aims to compare the advantages and disadvantages of the two data sources or simply use both to support the analysis. The objective of the research should be clearly defined.

2. The introduction of the manuscript is loosely connected to the research subject and fails to highlight the scientific novelty of the study in comparison to contemporary research. It is advised to more clearly explain the reasons for focusing on evaluation of cropping patterns and intensity, as well as the benefits this approach offers compared to other studies.

3. The "Materials and Methods" section needs further refinement to enhance clarity and brevity, particularly regarding the processing methods for SAR data.

4. In Section 4.3, the comparison results indicate that calculations using the lower-resolution MODIS data have smaller deviations. However, this finding seems to contradict statements in lines 509–512, as well as the conclusion and abstract. It is suggested to incorporate additional references and revise the discussion, abstract, and conclusion to align with the study's implications.

Additionally, the authors should address the reasons behind the discrepancies between the remote sensing estimates and government statistics. It would be beneficial to explore whether there are variations in methodologies, data sources, or definitions of cropped area that could explain these deviations.

5.Missing link between results and discussion: The discussion needs to explicitly link back to the results presented in the previous sections. For example, how do the findings on gross cropped area, net cropped area, and cropping intensity relate to crop diversification?

6. Add appropriate units to the data in Table 5 on page 13.

7. Figures 5–10 should be revised to meet publication standards.

Author Response

Responses to the reviewer comments 

Thank you for reviewing our manuscript

1. The objectives have been restated as indicated, and the introduction section has been updated with new insights.
2. The material and methods section included a brief insight into the SAR pre-processing steps, and the materials and methods have been updated throughout.
3. Though the smaller deviations in the statistics may not be as important, considering that we are aiming to compare the datasets, such measures also indicate the means of overestimation or underestimation. Hence, statements in this regard were included, and the discussion section has been updated.
4. The changes have been made to align with the other sections. Overall, the study states that resolution might not be a constraining feature for restricting the usage of the coarser resolution dataset for the application.
5. The units for the table 5 have been updated.
6. The clarity of the figure was updated and were realigned, and the images were also given separately for inclusion through the submission system

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have made revisions based on the comments. I believe the current version is acceptable. Congratulations.

Reviewer 2 Report

Comments and Suggestions for Authors

Yes, it is much better in this form, the article can be published in this form.

Reviewer 4 Report

Comments and Suggestions for Authors

The manuscript looks good to me now.

Comments on the Quality of English Language

Minor.