The manuscript proposes a rice mapping method to construct representative rice growth patterns based on time-series Sentinel-1 SAR data. Apparently, use of Sentinel-1 SAR data for rice mapping is very useful and seems the work is ongoing in the region for some time. Such methodology is useful, especially if it can give the rice area in advance which can help estimate the production helping in making timely trade/consumption plans.

Below are few comments that should be helpful to improve the clarity and completeness of the manuscript.

• What does this statement, .. rice growth patterns in Southeast Asia are too complex…means? Complexity refers to what? Is it not complex in other regions?
• This is nice to have high-precision rice area mapping but why is it needed. Please explain for the benefit of readers in section 1. Also is it only high precision or high resolution too?
• Were the data downloaded for whole year of 2019 of specific season? If so, the seasonal difference in rice area matters although there may be irrigation in some areas of some countries with 2 to 3 rice crops in irrigated areas and one crop in non-irrigated areas (as mentioned as Rice 1 through 4 in the manuscript). How was this seasonal difference considered in analysis in mapping the rice area? Because rice area per seasons will be different for the same country.
• Line 207-208, …the high heterogeneity of rice backscattering coefficients in Southeast Asia is caused by the high heterogeneity in climate and topography…what does mean that? Climate is obvious but what is heterogeneity in topography?
• Figure 3 is not so nice. It should be improved to be more clear to read.
• Line 223, if all flooded, why rice will differ significantly from other crops as other crops will also be flooded and possible similar backscatter, unlike for example sugarcane field which certainly may have different backscatter as they may not submerged. Please clarify the statement – the other crops.
• Figure 4, the term ‘Building’ does not reflect that square. Appropriate name is settlement. Rice is also vegetation, so vegetation may better be called as non-rice vegetation. What is band combination for optical image?
• How was number of (sample)plots of 1913 and 2032 for validation determined, any basis?
• Country’s statistics on area (despite may not align with data collection cycle) are most authentic. Table 5: the extracted rice area is only 44% of the statistics of rice cultivation area. This is rather huge difference. How to explain the feasibility of using the rice mapping method recommended by this study because of that discrepancy? It has now only Vietnam in the table. It is better to show all the countries data in the table.
• The method itself of rice mapping using Sentinel-1 SAR Data is major output of the study. If the proposed method is referring to Figure 2 flowchart, then, it still shows the need of using statistical data and available rice maps. Hence, better discussion with rationale is needed whether proposed method is to replace the existing system as efficient, accurate and even pre-harvest season rice mapping method or just additional task of mapping the rice area. Ideally, the new method is to replace or improve the existing method. It would be nice to indicate with the proposed method whether existing statistical data collection on rice is still needed. Also, advise (recommend) how this new method can be made available for the country to use? Afterall the best use of method if adopted will be for the country.

As rice land mapping is significant in Northeast Asia, the authors developed a mapping method based on time-series Sentinel-1 SAR data in this area. Based on the analysis of rice backscattering characteristics in the aimed area. Based on the analysis of rice backscattering characteristics,they combined spatio-temporal statistical features with the generalization ability to rice growth patterns, together with the adoped model approaches, the 20-meter resolution rice land cover map of five countries in mainland Southeast Asia in 2019 was obtained.The methods are helpful for agricultural management in the related areas. Similar methods can be adopted to make a regional rice cover map by repeating the experiment in this paper according to subsequent data. However, I have the following suggestions or concerns that maybe helpful for the modification of the manuscript.

My major concerns are:

1. Is the title accurate? It is strange to call rice mapping. I would like to suggest you take into careful consideration for this. Rice land mapping? Rice land cover mapping. Anyway, rice mapping is not so suitable.

2. Section 3.2: feature extraction was not told very clearly, and affected understanding for the whole processing process for mapping.In particular, the comparison between pseudo-color image and optical image (figure 4), and the role of the optical image in cartography were not clarified.

I have also some minor points as the following:

1. Abstract section: I am wondering what was the complex growth patterns for rice in tropical and subtropical regions. Since this is too vagur, would you please tell in details?

2. Introduction section: United Nation's sustainable development goals and the significances of food are not the most important for this manuscript, so please have some cut at this to make it short, and focus on the mapping methods and the advances in this field, and be noted that this is data manuscript.

3. Line 37, line 40-41, and also line 45-47: references were in wrong format or not suitable. Please also check many other places expressed like this.

4. Line 68-70: what was the difference between the two categories? The information in the second category can also include phenological information. So, confusing as far as this expression is concerned.

5. Line 70-74: Please refer to relevant literature, the expression of crop phenology should be regulated.

6. Line 106: the editing was careless: see line 106 "inSoutheast Asia".

7. Due to the limited information in table 1, I suggest make it supplement.

8. Section 2.2.2: the agricultural statistical yearbooks may have great uncertainty, which may affect its credibility. Please discuss uncertainties it brought.

9. Section 3.1: it was told the Sentinel-1 time-series data were preprocessed by using SNAP software, but you omitted the mechanism for conducting such events, and did not mention the how to determine the authority of the software.

10. Section 3.4: similar to above problem, the reason for choosing U-Net model was not introduced and affected the reliability of your processing results.   

11. Section 7, the conclusion section: the expression were not suitable, since you should tell the most important findings for the rice land mapping technology in this work, but not the evaluation and remark on that.