Paddy Rice Phenological Mapping throughout 30-Years Satellite Images in the Honghe Hani Rice Terraces

Round 1

Reviewer 1 Report

This study deals with an exciting approach to monitoring land cover changes using long satellite data series. However, it needs to improve the introduction. The text does not show any research challenges, and I suggest the authors improve the study's scientific presentation. In the current mode, the description sounds like a technical report. Next, I describe some more specific points that need adjustments:

Title:

The title contains information unnecessary, such as the name of the tool used in data processing (GEE), and there is essential information that needs to be included, such as the period and type of data used. I suggest a more direct description, such as: Rice paddy phenological mapping throughout 30-years satellite images in the Honghe Hani Rice Terraces

Abstract:

Line 17 - Listed as a World Heritage Site in 2013, the Honghe Hani Rice Terraces is a representation of coexistence between natural and cultural systems.

I suggest changing the sentence to a more direct form: The Honghe Hani Rice Terraces is a representation of coexistence between natural and cultural systems. This region was listed as a World Heritage Site in 2013.

The sentence between lines 25-28 is so long. I suggest to break in two or three sentences.

Don’t start the sentence of line 29 with percentages numbers. Reorder this sentence.

In the sentence lines 56-58, the words “paddy rice” happen three times. Please, Could you fit this sentence?

Avoid paragraphs with only one sentence (line 341-344 and 472-475). You must develop them.

Author Response

1. This study deals with an exciting approach to monitoring land cover changes using long satellite data series. However, it needs to improve the introduction. The text does not show any research challenges, and I suggest the authors improve the study's scientific presentation. In the current mode, the description sounds like a technical report. Next, I describe some more specific points that need adjustments:

Response: Thank you for the positive comment. We have carefully read all comments and suggestions, responded to them, and implemented the corresponding revisions. We have improved the research challenges in the first and second paragraphs of the Introduction to highlight the two challenges in paddy rice mapping. One challenge is the similarity of spectral features between paddy rice and other crops in the first paragraph of the Introduction. Another challenge is the impact of cloud contamination on the suitable time window of paddy rice in the second paragraph of the Introduction.

2. Title:

The title contains information unnecessary, such as the name of the tool used in data processing (GEE), and there is essential information that needs to be included, such as the period and type of data used. I suggest a more direct description, such as: Rice paddy phenological mapping throughout 30-years satellite images in the Honghe Hani Rice Terraces

Abstract:

Response: We have revised the title from “Revealing the Paddy Rice Area Changes in the Honghe Hani Rice Terraces Based on Phenological Features and Google Earth Engine” to “Paddy Rice Phenological Mapping Throughout 30-years Satellite Images in the Honghe Hani Rice Terraces”.

3. Line 17 - Listed as a World Heritage Site in 2013, the Honghe Hani Rice Terraces is a representation of coexistence between natural and cultural systems.

I suggest changing the sentence to a more direct form: The Honghe Hani Rice Terraces is a representation of coexistence between natural and cultural systems. This region was listed as a World Heritage Site in 2013.

Response: We have changed “Listed as a World Heritage Site in 2013, the Honghe Hani Rice Terraces is a representation of co-existence between natural and cultural systems” to “The Honghe Hani Rice Terraces represent the coexistence between natural and cultural systems. Despite being listed as a World Heritage Site in 2013…”.

4. The sentence between lines 25-28 is so long. I suggest to break in two or three sentences.

Response: We have divided this sentence into two sentences. Among them, the first sentence describes the separability between paddy rice and other land use types at different phenological periods; the second sentence describes adding phenological information to improve mapping accuracy and stability.

5. Don’t start the sentence of line 29 with percentages numbers. Reorder this sentence.

Response: We have changed “10.65%, 8.81%, and 5.71% of paddy rice were converted to forests, shrubs or grasslands, and other croplands in the Honghe Hani Rice Terraces from 1989-91 to 2019-21, respectively” to “In the past thirty years, 10.651%, 8.810%, and 5.711% of paddy rice were respectively converted to forests, shrubs or grasslands, and other croplands in the Honghe Hani Rice Terraces”.

6. In the sentence lines 56-58, the words “paddy rice” happen three times. Please, Could you fit this sentence?

Response: We have changed “Previous studies used phenological features to map paddy rice by detecting the surface water and green vegetation change of paddy rice to discriminate paddy rice and other land use types” to “Previous studies used phenological features to map and discriminate paddy rice from other land use types by detecting its surface water and green vegetation change”.

7. Avoid paragraphs with only one sentence (line 341-344 and 472-475). You must develop them.

Response: For the sentence on lines 341-344, we have divided this sentence into two parts. The first sentence described the application of data input and classifier to map the historical distribution of paddy rice from 1989-91 to 2019-21. Another sentence described the average mapping accuracy of seven periods, and we also provided the detailed accuracy results in Table S2.

For the sentence on lines 472-475, we have divided this sentence into three parts. The first part described the impact of other driving forces on paddy rice losses in the Honghe Hani Rice Terraces. The second and third parts described the impact of the Grain for Green Program and landslide risk on paddy rice losses in the Honghe Hani Rice Terraces.

Author Response File: Author Response.docx

Reviewer 2 Report

This paper attempts to provide an improved historical mapping of Paddy Rice in the Honghe Hani Rice Terraces. This is particularly important in combining local knowledge and remote sensing tools to improve the mapping accuracy of Paddy Rice changes. Despite the importance of the topic, I find several areas for improvement in the manuscript that must be addressed before being considered for publication. The authors could have been more specific and more clearly explained how the interviews and questionnaires were utilized to derive the results shown.

Major Comments/Questions

1.      First of all, how the three-year aggregation data was performed needs to be clarified. Different temporal aggregation approaches exist to merge temporal data (e.g. median, mean, etc.)

I found it vague that the authors stated in line 259, “.., we merged Landsat spectral bands and vegetation indices of three-year periods….”. The authors should provide more clarification on how it was done. The paper should be self-contained, and the readers should find all the necessary information to understand the methodology used. Addressing this issue will tremendously improve the manuscript.

The authors may want to check the following article that is relevant to the methodology referred to in the manuscript:

[1] Carrasco, L., Fujita, G., Kito, K., & Miyashita, T. (2022). Historical mapping of rice fields in Japan using phenology and temporally aggregated Landsat images in Google Earth Engine. ISPRS Journal of Photogrammetry and Remote Sensing, 191, 277-289. https://0-doi-org.brum.beds.ac.uk/10.1016/j.isprsjprs.2022.07.018

2.      Correlation is different from causation. Two highly correlated variables don’t necessarily mean that one variable is affecting/causing the outcome of the other and vice versa.  Other tools can be utilized other than the coefficients of correlation.  

Pearson correlation cannot be used to derive implications (lines 381-382 and 384-386), such as “ …..positive correlation of annual precipitation and paddy rice (0.587) indicated that increased temperature and reduced precipitation, e.g., warming or continuous drought, may lead to paddy rice losses in the study area.”

Please refer to these values as relationships and not causation.

3.      Lines 49-52: Please provide references for your statement: “We found …….., leading to problems in identifying the driving mechanisms behind paddy rice distribution changes.” ?

4.      To my knowledge, Landsat TM stopped in 2011; however, in lines 193 to 195, the authors seem to suggest that TM data from 1989 to 2012 were used. Did you use images from 2012? Please provide the image references (Path/row and dates) if so. If not, how the gap between TM and OLI was addressed?

5.      Lastly, the manuscript needs more work, especially in the Discussion, Applications, and Conclusion sections. The writing style must be improved, and redundant explanations must be addressed. I found the manuscript rough, and it needs improvement.

Specific comments:

Line 142: Please replace ..shown that .. by ..show that ….

Line 181: Table 1 needs better formatting to allow smoother reading.

Line 184: Figure 2: For easier and quicker cross-comparison between vegetation types, I suggest you have the same scaling (vertical axis) for each index.

Line 388: Please correct the figure number to 8.

Line 205: the EVI formula is missing the gain factor of 2.5. See reference [38] you cited for EVI.

Comments for author File: Comments.pdf

Author Response

1. This paper attempts to provide an improved historical mapping of Paddy Rice in the Honghe Hani Rice Terraces. This is particularly important in combining local knowledge and remote sensing tools to improve the mapping accuracy of Paddy Rice changes. Despite the importance of the topic, I find several areas for improvement in the manuscript that must be addressed before being considered for publication. The authors could have been more specific and more clearly explained how the interviews and questionnaires were utilized to derive the results shown.

Response: We appreciate the reviewer’s valuable comments and suggestions. We have responded to all the comments and suggestions point-by-point below. Please refer to the specific responses.

Major Comments/Questions

1. First of all, how the three-year aggregation data was performed needs to be clarified. Different temporal aggregation approaches exist to merge temporal data (e.g. median, mean, etc.)

I found it vague that the authors stated in line 259, “.., we merged Landsat spectral bands and vegetation indices of three-year periods….”. The authors should provide more clarification on how it was done. The paper should be self-contained, and the readers should find all the necessary information to understand the methodology used. Addressing this issue will tremendously improve the manuscript.

The authors may want to check the following article that is relevant to the methodology referred to in the manuscript:

[1] Carrasco, L., Fujita, G., Kito, K., & Miyashita, T. (2022). Historical mapping of rice fields in Japan using phenology and temporally aggregated Landsat images in Google Earth Engine. ISPRS Journal of Photogrammetry and Remote Sensing, 191, 277-289. https://0-doi-org.brum.beds.ac.uk/10.1016/j.isprsjprs.2022.07.018

Response: We have carefully read this manuscript and added material to describe the details of the merged temporal data as follows:

• Highlighting the reason why we need to merge indices and bands of three-year periods: “To ensure we had enough Landsat images with lower cloud covers (< 5%) to analyze the driving forces of paddy rice area changes.”
• Explaining the method of merging temporal data: “We calculated median values of Landsat spectral bands and vegetation indices during the FTP (DOY from 1 to 126) and the growing and harvesting period (GHP, DOY from 126 to 300) data input within three-year periods”.
• Showing the results of image collection numbers and cloud cover: “We obtained 11 Landsat image collections of three-year periods: 1989-91, 1992-94, 1995-97, 1989-00, 2001-03, 2004-06, 2007-09, 2010-12, 2013-15, 2016-18, and 2019-21. Among them, except for 2001-03 (cloud cover = 27.455%) and 2013-15 (cloud cover = 5.771%), the cloud covers of the other Landsat image collections are lower than 4%”.
• Providing the paddy rice mapping code in Appendix A of the Supplement.

2. Correlation is different from causation. Two highly correlated variables don’t necessarily mean that one variable is affecting/causing the outcome of the other and vice versa.  Other tools can be utilized other than the coefficients of correlation.  

Pearson correlation cannot be used to derive implications (lines 381-382 and 384-386), such as “ …..positive correlation of annual precipitation and paddy rice (0.587) indicated that increased temperature and reduced precipitation, e.g., warming or continuous drought, may lead to paddy rice losses in the study area.”

Please refer to these values as relationships and not causation.

Response: We have adjusted this section by only displaying the correlation coefficient between paddy rice and its driving forces and changed Figure 8 from a heatmap to linear regression diagrams with correlation coefficients. In addition, we have provided the results of questionnaires to support driving force analysis in sections 2.7 and 3.4.

Figure 8. The correlation coefficient (r) between paddy rice and its driving forces. (a-d) are the correlation coefficients between rice paddy with annual temperature, annual precipitation, GDP, and proportion of primary industry in four counties from 1989-91 to 2019-21, respectively.

3. Lines 49-52: Please provide references for your statement: “We found …….., leading to problems in identifying the driving mechanisms behind paddy rice distribution changes.”?

Response: We have changed “warming (r = -0.73), drying (r = 0.59), and proportion of primary industry reduction (r = 0.70) mainly led to paddy rice losses” to “Lower agricultural profits and drought leading to problems in identifying the driving mechanisms behind paddy rice distribution changes” in the Abstract section.

4. To my knowledge, Landsat TM stopped in 2011; however, in lines 193 to 195, the authors seem to suggest that TM data from 1989 to 2012 were used. Did you use images from 2012? Please provide the image references (Path/row and dates) if so. If not, how the gap between TM and OLI was addressed?

Response: We have corrected the time of TM from “1989 to 2012” to “1989 to 2011”. The time of OLI from “2013 to 2021”. Therefore, by referring to the manuscript of Carrasco, L., et al. (2022), we used ETM+ images to compensate for the missed image in 2012. “Because the band numbers, wavelength ranges, and spatial resolution of the Enhanced Thematic Mapper Plus (ETM+) are similar to those of TM”. We have also provided the handling code in Appendix A of the Supplement material.

In addition, we have provided the path/row in the study area “the path/row of Landsat satellites in the HHRT are 128/44, 128/45, 129/44, 129/45, and 130/44”, and the image numbers in different phenological periods and Landsat sensors from 1989-91 to 2019-21 in Table S1 in the Supplemental material.

Table S1. The image numbers in different phenological periods and Landsat sensors from 1989-91 to 2019-21. TM is the Thematic Mapper Sensor, ETM+ is the Enhanced Thematic Mapper Plus sensor, OLI is the Operational Land Imager sensor, FTP is the flooding and transplanting period, and GHP is the growing and harvesting period.

5. Lastly, the manuscript needs more work, especially in the Discussion, Applications, and Conclusion sections. The writing style must be improved, and redundant explanations must be addressed. I found the manuscript rough, and it needs improvement.
   Response: We have improved this manuscript by deleting redundant content, especially in the Discussion section. In addition, we have revised the Conclusion section and emphasized the specific results of this study.

Specific comments:

Line 142: Please replace ..shown that .. by ..show that ….

Response: We have replaced “…shown that…” with “…show that…” in this manuscript.

Line 181: Table 1 needs better formatting to allow smoother reading.

Response: We have improved the format of Table 1.

Line 184: Figure 2: For easier and quicker cross-comparison between vegetation types, I suggest you have the same scaling (vertical axis) for each index.

Response: We have used the same scaling (vertical axis) in Figure 2, the range of NDVI, EVI, and NDSVI from 0 to 1, and the range of LSWI from -0.2 to 0.6.

Line 388: Please correct the figure number to 8.

Response: Thank you. We have corrected the figure number to 8.

Line 205: the EVI formula is missing the gain factor of 2.5. See reference [38] you cited for EVI.

Response: Thank you. We have corrected the EVI formula.

Author Response File: Author Response.docx

Reviewer 3 Report

The paper is well written.

Just look out for areas to improve readability by checking for spellings to improve the grammar.

Comments for author File: Comments.pdf

Author Response

The paper is well written.

Just look out for areas to improve readability by checking for spellings to improve the grammar.

Response: We appreciate the positive comments. We have improved the English to a native language level and revised the linguistic and spelling errors by rewording several unclear sentences in the revised manuscript.

Author Response File: Author Response.docx