Mapping Diverse Paddy Rice Cropping Patterns in South China Using Harmonized Landsat and Sentinel-2 Data
and Baodong Xu 2,5,*
Lang Xia
Round 1
Reviewer 1 Report
This paper aims to explore an effective way for mapping diverse paddy rice cropping patterns (PRCPs) in South China. In order to reduce the cloud impacts in the PRCP region, authors employed a new harmonized Landsat and Sentinel-2 (HLS) product to capture the specific spectral and phenological features for each PRCP. The proposed feature selection and hierarchical classification (FSHC) method showed the good performance for identifying PRCPs based on field crop samples. The topic of this paper is interesting and the need for generating PRCP map is obvious. The manuscript is generally well-written and the statements are clear and logically organized. There are only several minor issues that need to be resolved before this manuscript can be accepted for publication in Remote Sensing. Please find my detailed comments below.
Page 2, Lines 73-74: It is well known that the moderated spatial resolution (> 250m) cannot be suitable for smallholder croplands caused by land cover mixture effects. Since many previous studies also tried to increase high-quality decametric-resolution observations by combining moderate- and high-resolution images, I suggest authors can introduce these studies in this paragraph.
Page 3, Lines 106-113: The evaluation of FSHC method based on field crop samples is also an important objective in this study, which can make readers know the performance of extracted PRCPs by this method. Therefore, authors can add this objective in the Introduction Section.
Page 3, Lines 123-127: Which year of GlobeLand30 product was used to extract cropland layer? Please clarify it.
Page 5, Lines 168-187: In this section, authors listed many potential VIs for identifying PRCPs in this study. However, several VIs were not specified for capturing the characteristics of paddy rice. The results in Figure 6 also showed this limitation. Thus, I suggest authors can delete several unnecessary VIs to improve the readability of this paper for readers.
Page 6, Lines 211-222: The slope data was employed to exclude non-paddy rice, which was not widely used in previous studies. Was it related to the study area? Authors can add more contents to discuss this issue in the Discussion Section.
Page 6, Lines 224-232: The evaluation part of the proposed FSHC method was also shown in Figure 3, but it was not introduced in this paragraph. Please clarify it.
Page 10, Lines 370-371: The observation date of pseduocolor image was important for readers to understand the phenological stage of different PRCPs. So, please add this information in the caption of this figure.
Page 11, Lines 393-396: Some VIs cannot show the separability between paddy rice and non-paddy rice. Did author need all VIs? Please refer to my previous comment to reduce the feature redundancy in this paper.
Pages 12-13, Lines 421-442: Since NDWI in DOY 110 or 205 also showed the large separability between paddy rice and non-paddy rice over single-cropping cropland, I do not understand why authors introduce dNDWI in this section. Generally, the complicated indicator may not derive the robustness result. Authors should perform several tests and consider that whether they need this indicator in this study.
Page 15, Lines 507-510: Figure 12b and 12c have the same x-axis and y-axis labels but no other more information to know details of each figure. Please revise this figure for clarity.
Author Response
Dear Reviewer,
We greatly appreciate your valuable comments for this manuscript. We have carefully addressed all comments and provided the itemized response to each comment. Please see the attachment to check our responses to your comments. Thank you very much!
Author Response File: 
Author Response.docx
Reviewer 2 Report
Several comments are provided in the paper. Figures are hardly readable. The need to start from an available crop map reduces the interest for this paper.
Comments for author File: 
Comments.pdf
Author Response
Dear Reviewer,
We greatly appreciate your valuable comments for this manuscript. We have carefully addressed all comments and provided the itemized response to each comment. Please see the attachment to check our responses to your comments. Thank you very much!
Author Response File: Author Response.docx
Reviewer 3 Report
The sections 3 and 4 are a little bit mixed up, in the sense that some intermediate results are already in section 3 (lines 298-301) and other methodological aspects are disseminated throughout the section 4.
In addition, the conceptual scheme of the constructed DT appears at the end of the section 4 (fig.10), which is hard to follow. I would suggest to include a graphic in section 3 describing the methodological steps (as in fig.10) and give results in Section 4 with a clear correspondance to the steps of section 3.
Author Response
Dear Reviewer,
We greatly appreciate your valuable comments for this manuscript. We have carefully addressed all comments and provided the itemized response to each comment. Please see the attachment to check our responses to your comments. Thank you very much!
Best regards,
Baodong Xu (on behalf of co-authors)
Author Response File: Author Response.docx
Reviewer 4 Report
Mapping diverse paddy rice is interesting and significant. Crop types classification in South China by remote sensing has long been challenging, due to the frequent rainy weathers and fragmented agricultural landscapes. The Harmonized Landsat and Sentinel-2 Data provides great potentials for accurate crop identification in such regions. Previous studies focused more on paddy rice mapping, while the refined rice mapping, particularly for the complex cropping and rotation system has never been investigated. This study is of great interests for the researchers in the fields of agricultural remote sensing and land use change. This paper is presented well and written clearly. The number of figures and references are suitable to the length of the paper. I just have some minor suggestions for the authors.
Line 103-105: Temporal information is crucial for crop type mapping. I suggest authors rephase this viewpoint and add reference to support it.
Figure 1: What are the meanings of the numbers in the right map of figure a?
Line 247: As cropping intensity mapping is the first step for the FSHC, I suggest authors changed the “cropping intensity mapping” to “identification of cropping intensity”.
Line 259-260: This sentence is confusing. What does the “prior knowledge” mean?
Authors need check the tense of sentences in the entire manuscript. Many descriptions on the used method should be in the past tense rather than the present tense.
Line 323-324: I think the second “paddy rice” was wrongly written, which should refer to “cropping intensity”. Please check it.
Discussions：I suggest authors discuss the potential values of these produced PRCP maps for land use planning, agricultural production management et al., which can strengthen the significance of this study and attract more readers in various fields.
Author Response
Dear Reviewer,
We greatly appreciate your valuable comments for this manuscript. We have carefully addressed all comments and provided the itemized response to each comment. Please see the attachment to check our responses to your comments. Thank you very much!
Best regards,
Baodong Xu (on behalf of co-authors)
Author Response File: Author Response.docx
Reviewer 5 Report
This study intended to map diverse paddy rice cropping patterns in South China using multi-source remote sensing data. The proposed FSHC method is practical and can be easily to apply in other regions. Particularly, the decision tree based on the optimized spectral and temporal features for PRCP mapping was developed in detail and was reasonable in practice. Authors demonstrated a good knowledge or recent development in the area of crop type mapping. I think the idea that fully capitalizes the spectral and phenological features for mapping different paddy rice types is good and implementation methodology is reasonable. In my opinion, only minor revisions are needed before this paper can be accepted for publication.
P2, L64-67: When did the “flooding signature” appear? Does the flooding signature occur in rice flooding or transplanting phase? Authors need clarify the major land cover types (water or green vegetation) that dominate the spectral reflectance of remote sensing images in this special phase.
P5, Figure 2: Clarify the meanings of “N” for HLS and field samples. If they have different meanings, two letters should be used here to avoid confusion.
P7, L252-253: Whether the training samples were independent from the validation samples? This is very important for the performance evaluation of the presented method.
P13, Figure 7: What is the meaning of the PP? The spacing of the histogram is a little small. It is hard to see the PP clearly.
P17, Section 5.2: Except for limitations and future improvements, can authors point out the future implications of PRCP maps in different research fields?
Author Response
Dear Reviewer,
We greatly appreciate your valuable comments for this manuscript. We have carefully addressed all comments and provided the itemized response to each comment. Please see the attachment to check our responses to your comments. Thank you very much!
Best regards,
Baodong Xu (on behalf of co-authors)
Author Response File: Author Response.docx
Round 2
Reviewer 3 Report
Authors revised accordingly to my comments
