The manuscript by Valentin Dura et al. aims to study and describe the PLR, an important quantity that is, however, still poorly explored and very uncertain, especially in areas with complex topography.

Overall, I believe that the manuscript contains relevant and interesting content that deserves to be published. Furthermore, it is well written and understandable. I recommend publishing it after the authors have kindly considered the few aspects that I would recommend improving.

General comments

The only general aspect I would recommend improving is in the Results and Discussion sections.  The results are explained with so many values and details that the reader runs the risk of losing the overview. This is tried to be remedied in the Discussion and Conclusion. Neverthelss, I would try to take better care and add, after or before each section explaining much details, a sentence summarising and explaining the highlights and the general picture. Sometimes the authors have already summarised the results, but if this could be improved a little more I think it would help the reader a lot.

Specific comments

• Affiliation after the author list: wrong space at the beginning "3 Univ. Grenoble Alpes, CNRS, IRD, Grenoble INP, Grenoble, France"
• L104: please reformulate this sentence better introducing “Italy phenomenon”
• L120: I got stuck on the word massif, then realised that in the next sentence it is explained and put between apostrophes. Please introduce it better.
• Already in Fig. 1 it would be interesting to know the area in the different altitude classes as well as the number of stations.
• 1: It would be useful to reduce the white space between the subfigures and enlarge them. They are currently very small and difficult to read
• L190-191: The other datasets also cannot be defined as homogeneous (change in instrumentation, station density,...)
• L215-219: What influence does this correction have on the PLRs?
• 3 (b): Hard to see the black line. In addition the abbreviation “RG” is explained in the manuscript, but it would be helpful to mention it again in the Figure caption.
• L403: How was the value of 0.5 chosen?
• L549: “m” instead of “mp”

This is an interesting, well-written and well-thought manuscript.

The main findings are two-fold: 1) on the one hand there are interesting conclusions on the intercomparison of precipitation products, which is insightful. 2) On the other hand, there is a lot of discussion on the orographic precipitation processes as revealed in the different datasets, which are also interesting. This second aspect of reflection on the physical mechanisms is not very much emphasized in the introduction and conclusion. In that regard, the research could be better “sold” by highlighting the link between datasets and processes revealed by a fine analysis of these datasets. This is however just a comment that does not require necessary changes in the manuscript.

This said, I do have some comments that will require some minor modifications of the manuscript.

Main comments:

1. All or some of the compared precipitation products include altitudinal effects or corrections for such effects. does this create precipitation lapse rates artificially? A comment on this would be welcome.
2. In eq. 6: the conversion from beta to PLR is unclear to me. A slope is bounded between 0 and infinity, so how can it be converted to a bounded interval between 0 and 100?
   
   
   
   Also, the denominator P^bar is not defined. Is it the average or median precipitation, and over which area?
3. 389-390 and other places: I can see how there can be more variability in the PLR when one focuses on smaller scales, but not how the average R2 can change depending on scale. I may be wrong, but in principle the R2 value of a large area should be the average of its component subareas. Showing numerical values might clear any doubt.
4. Results in figure 7 (and maybe others): These results are only based on AROME, which is not forced on rain gauge data. While AROME does not perform badly in general, it is not perfect either, and is significantly different than other data-based products. Therefore, I am uncertain of the value of conclusions based only on AROME.
   
   Would it be possible to draw a similar figure as figure 7 based only on stations data, even if it means that many more catchments will have NA values?
   
   
   
   The fact that AROME does not use rain gauges is mentioned as an advantage (e.g. l. 461-462), but this poses the following question: if AROME poorly matches the high-elevation gauges, how to know whether it is because of undercatch or because of some deficiency in AROME? If the comparisons were done against stations data that are corrected for a possible undercatch, would AROME still perform as well? I am not asking to do such an exercise, but it could be discussed.
5. many of the results are given as figures, but in general it would be good to have tables with actual numerical values (for example in figure 3b the RMSE between stations and models, and also in figure 4)

Detailed comments and typos:

l.61: word missing. higher reliability?

l.180: incomplete sentence, reformulate

Figure 2: large areas seem to have a mean precipitation close to 0 (e.g. in SERVAL). is this a feature of the data or an artifact of the color scale?

l.275: 50mm is a very small altitude difference. should it be 50m?

Figure 4: would be nice to also have numerical correlation values rather than only visual scatters

l.477: resume -> summarize

section 5.4: New methods and results should not be introduced in the discussion section. I recommend moving the part on non-linear results higher in the manuscript.

Figure 9: This figure is very interesting. It shows that larger catchments have more non-linear PLRs. The question that begs for an answer is whether there exists a small subdivision of the catchments where all relations are linear, or if some portions of catchments have irremediably non-linear PLRs.

The paper analyses the relation between annual/seasonal precipitation amount with elevation represented by seven different precipitation products and focusing on an orographically complex area in France. The analysis is performed at different spatial scales and allows to asses the spatial variability and spatial-scale dependence of precipitation lapse rates (PLRs). The main conclusions are that PLRs show high-spatial variability, depending on season and altitude, and scale-dependent values, and that a high-resolution convection permitting model (AROME) could be beneficial for a precise assessment of PLRs in complex topography regions.

Based on my personal reading, I find the paper interesting and containing detailed and well-conducted analysis. It explores the important topic of the relation of precipitation and altitude, giving advancements in the comprehension of its variability and scale-dependence. I suggest minor revision before publication, for improving the clearness and readability of the results and discussion sections.

General comment:

With many analyses done (many precipitation products, massif scale, catchment scale), I know you have many results to describe. But I find difficult to not be lost while reading sections 4 and 5, and difficult to get the picture of the main findings. I suggest to, for example, split section 4.1 in two parts (for example, the second focusing just of the 4 products, from line 340), or to clearly summarize the main conclusions at the end of each sub-section (as done for example at lines 369-377).

Specific comments:

• If possible, I suggest to increase the size of some figures and text in figures. It’s difficult to distinguish lines or read the text in: Figure 1a, Figure3 (very difficult to distinguish the lines), Figure 6a,b (name of massif and numbers are very small; maybe it could help to have a table listing names, values of R^2 and PLR, and just number in the figure to identify the massif in the list)
• Figure 1: “SAFRAN” name is mentioned in the caption, not in the text. What is it referring to?
• Figure 4: Is the comparison different in the different regions? could it be possible to distinguish with colors the points for the different regions? Do they cluster? Is there a region more aligned with bisect? The precipitation products are represented by the different columns, so maybe colors could be used for giving further information on the comparison in different regions.
• Figure 7: “insufficient altitudinal variability” should be when “standard deviation is lower than 50 m” (not higher)
• Section 5.4. I suggest to move lines 512-523 and table 3 into the methodology section.
• Line 540: it’s not clear to me the meaning of this sentence.