The objective of the manuscript is to present a new gridded precipitation dataset across mainland China using the best interpolation scheme among the 8 tested. Reliable precipitation data are important to ensure water safety and guarantee water availability and quality. Hence, efforts in creating reliable datasets are quite valuable.

However, the manuscript in its current form lacks a critical discussion on the limitation of the gridded data available and on the selected interpolation scheme. The Authors provide a list of gridded precipitation datasets available (Lines 71-104). However, a critical review of such datasets is missing. They mentioned the sensitivity to interpolation algorithms, but it is too vague. The Authors do not discuss why the scheme considered the optimal (even though it is not an optimal scheme but rather the best, based on some metrics, among the few schemes tested) leads to better goodness of fit metrics. Is such a result expected? Why is such a combination better than the others? It is simply chance? Can this scheme be transferred to other regions?

Point-by-point comments:

Abstract: I suggest the Authors revise the abstract. The primary objective of the paper (the new gridded data) and the temporal coverage (from when to when) should be better highlighted. Moreover, it should be clearer why the interpolation method selected is the best among the ones tested and how it addresses the limitations of currently available products. RMSE and other metrics as presented are not enough to judge the goodness of the method. How does this perform compared to the others? Why does it perform better?

Lines 80-81: what does the following sentence mean? “Through a fusion of remote sensing products and reanalysis datasets into in situ station data”. Remote sensing products and reanalysis data generated gauged precipitation dataset? Or gauged data were combined with remote sensing products and reanalysis data?

Section 2.3 and 2.4: Which method did the Authors use to re-grid the data? Where the raw data can be found?

Line 173: is it possible to eliminate interpolation errors?

Line 200: in the 30-year mean daily precipitation, was there any trend in the data or inhomogeneity?

Lines 245 - 325: Four interpolation methods to construct the field of ratio. Still missing how they differ, why those have been chosen, and why they provide different results.

Section 4.1 (starting line 369). My suggestion is to revise the term “optimal” for a scheme since there is no optimal scheme but simply the scheme having better metrics compared to the other schemes tested. The question of why such a combination of methods leads to better goodness of fit metrics is not answered. Why is such a combination better compared to the others? It is simply chance? Can this combination be transferred to other regions? Since the schemes perform differently depending on the topography (369-372), how do these differences affect the overall performance of the scheme? Are the metrics’ values listed (lines 375-380) average over the 45k stations used for verification?

Lines 479: Optimal interpolation scheme. Again, there is no optimal but the best among the ones tested