General Comments:
In this manuscript the authors used several sources of historical precipitation data including point measurements and radar data to downscale precipitation projections from a Regional Climate Model (RCM) using a stochastic approach. Outputs from the RCM were first corrected by modifying the spatial covariance structure using a matrix-algebra based approach, as well as a marginal probability distribution correction using a quantile-quantile transformation. Once these corrections are applied to RCM projections for a historical and future time periods, data from the RCM model were downscaled to a finer spatial resolution. Two important applications are shown: derivation of depth-duration-frequency (DDF) curves and area-reduction factors (ARF) for future precipitation projections under RCP8.5 greenhouse gas concentrations scenario.
Although the paper is in general well written, and the topic/application aligns well with the HESS journal, certain methodological details are insufficiently described. For example, section 3.2 lacks clarity. Please consider revising these methodological gaps to enhance comprehensibility.
The current version of the paper is rather lengthy, and implementing Referee’s 1 suggestions would necessitate a comprehensive re-write, comparing multiple approaches. Although some suggestion might be beyond the scope of this work, authors might want to incorporate relevant points into the discussion, highlighting other possible approaches and potential limitations of their research.
The second referee rises an important topic regarding Inter-model comparisons and model-to-model variability in Climate Change assessments. While the manuscript does not explicitly focus on inter-model comparisons, it does present a sequence of methods for generating a statistically and physically sound downscaled rainfall data from Regional Climate Models (RCM) projections. However, is crucial to recognize that an increase in rainfall depth over all temporal and spatial scales might not universally apply. Different Global Climate Models (GCM)/RCM outputs could yield varying results.

Following I have listed some specific comments that might be useful to enhance clarity of your manuscript.

Specific comments
- Line 175: Can you please be more specific about the rank correlation type method used?
- Line 225: I would suggest rewriting this paragraph defining the recorrelation matrix components first. I understand the authors are citing the reference, but for the sake of completeness it is better to briefly define the recorrelation process.
- Line 245: Can you clarify why the recorrelation procedure improved the model dependence structure compared to the original data for both season? Is the original data the DWD interpolated or the radar data? Please clarify.
- Line 265: It is not clear how effective the extrapolation process was. There is not much information given about the threshold selection. Can you please expand on this topic?
- Line 390: When you describe the steps to downscale the EURO-CORDEX RCP8.5 data in item 6: Do you mean optimizing the kernel bandwidth?
- Line 371: Consider referencing Table A1 in the Appendix to enhance clarity regarding your variogram clustering results.
- Line 545: Please review wording in this line for accuracy.

Thanks to the authors for addressing the Editor’s and reviewer’s comments. This is a rather long paper, and some of the sections still need some re-writing and re-wording.
Here is what I found it needs revision:
- Line 165: Please change “function of of future” by function of future
- Line 195: Although it is clear that you are addressing the issue of change of support in the covariance and variance estimation (equations (1) and (2)), it is worth to complete the sentence” Both the covariance and variance decrease” by adding an additional statement like for example “ when comparing point to areal estimates.
- Line 295: In the sentence: “ The method aimed to …”. Which method are you referring to?
- Line 210: Wording of your last sentence in this paragraph might be improved: “ A similar procedure was implemented …., using the theoretical correlation of the normal score transformed data”. Which theoretical correlation are you referring to?
- Line 220: Since you changed “probability value” to “probability level”, it is better to be consistent throughout the paragraph as well in appendix A1.
- Line 230: After equation (4): I would suggest changing the definition of y_{\alpha} as the standard normal quantile corresponding to a probability \alpha.
- Line 245: Please review item 7 and consider re-writing this sentence.
- Line 250: Please consider adding a comma after: ” with different statistics”
- Figure 6: Please indicate that panel (a) is for Winter and panel b) is for Summer in the Figure legend.
- Line 280: When you say: “For the observation and historical data, a Weibull…”. Can you specify whether the data is daily, hourly?
- Line 285: Please revise the definition of (Z_m,U_m). According to your equation 7, U_m is a probability level.
- Line 330: Consider re-wording the first sentence after section 3.4.
- Figure 8: Please indicate that panels a), b) and c) corresponds to the different variogram clusters.
- Line 405: Please consider re-wording this sentence: “ were generated using random mixing. Hence, the downscaling part”.
- Line 435: Please revise the last sentence: “Hence. To account for this difference…”. Consider re-wording this sentence.
- Line 495: The sentence: “For example,the change…6-hour scale for example”. Please consider rewording this sentence.
- Line 545: Please review the wording of the sentence: “the signal in the RCP projections was not lost, the increase…”. What do you mean with the “signal was not lost”?
- Line 560: Please review: “ future scenarios will would establish…”
- Line 805: Please replace “ comparision” by comparison.

Thank you to authors for taking care of these final revisions.