The Tibetan Plateau (TP) is a hotpot for studying soil erosion under climate change. Rainfall erosivity (or the R factor) is the most widely used parameter regarding to climate in soil erosion study. Thus, an accurate R map and related dataset is benefit to quantify soil erosion on the TP. Generally, this is a good MS and provide valuable data on TP. I recommend a moderate revision. Please find bellow my suggestions for updating & reinforcing the current paper.

1.    Abstract: I suggest the authors add more information about the dataset.

2.    Introduction: The ERA5 should be introduce in this part. For example, I notice that this product has been used to calculate rainfall of the China’s mainland. 

3.    Line 112, Please check the spatial resolution of the ERA5 data, is 25 km or 0.25°?

4.    Figure 4 and 5 : Add the unit of rainfall erosivity, which is MJ·mm·ha−1·h−1·yr−1.

5.    Part 4.2 (Line 243) Why the authors use multiplier factors to calculate new R map. Are there any references? Or the observed and ERA5-based annual rainfall erosivity show multiple relationship? Also in figure 8, why this is an optimal model with intercept of 0?

6.    I suggest an additional part as Uncertainties in the results parts. Uncertainties either from the ERA5 or from the multiplier factors should be discussed.

7.    Conclusions: I suggest this part focuses on summary of the dataset including its applications.

The authors attempted to generate a new dataset of R factor on TP. This work is useful to for water erosion modelling. I recommend a major revision. Please find bellow my suggestions for the current paper.

1. Abstract: The spatial resolution of the genenrated gridded annual rainfall erosivity dataset should be added.

2. Introduction: I suggest the authors add more information of the previous studies about the R factor on the TP.

3. Precipitation data: The detail information of the ERA5 data should be included in this section.

4. The accuracy of the ERA5 data should be examined at the weather stations.

5. The procedure of how to reconstruction R factor in 1950-2012 is important. However, nNo more information could be found in the methodology section. I would suggest the authors to rewrite this section to make this part more clearly.

This paper developed a new long-term soil erosivity dataset using the ERA5 reanalysis precipitation product, corrected by 1-min time-step meteorological observations. The core of this work lies in the use of high-frequency precipitation measurements, which however were not fully presented and discussed. For example, what’s the difference between this 1-min precipitation data and ERA-5 hourly product/other previous datasets? Furthermore, what are the contributions of precipitation frequency and intensity, respectively, to soil erosivity? These questions are important, considering the authors mentioned these issues in the introduction part. Overall, this paper does provide a valuable dataset for the TP region which is vulnerable to soil erosion but is generally superficial in clarifying the advantages and differences of this dataset against previous efforts. Therefore, I can only recommend a major revision for this paper in its current form.

Minor comments

Lines 307-309: The previous soil erosivity products are also based on observed precipitation data. I cannot see any difference of the methodology of this paper from the previous ones. Contrarily, the IDW method is actually too simple for the TP region, where the terrain is complex.