This study examines the impact of elevation and other geomorphological factors on the distribution of the average annual maximum precipitation of durations of 1 and 24 hours in Italy. The topic is of high scientific interest and relevant to the readers of this journal. The paper has a practical and data-driven cut, and provides quantitative analyses of use to practitioners and local stakeholders. 

I think it can represent a suitable contribution to this journal, provided some aspects are addressed.

The list below only includes references not present in the manuscript. Their purpose is to support the reasoning and should not be considered as recommended.

1. While the paper has a clearly practical cut, I am missing an attempt toward the understanding of the physical reasons behind what is found. I think this should be at least attempted in the discussion, possibly putting the findings in perspective with what is already known on the topic.
2. The state of the art presented in the manuscript mostly focuses on the examined study region (Italy), and is implicitly limited to cover what is relevant to this specific study (i.e. mean annual maxima). However, I feel a wider perspective should be provided, especially when claiming: “The impact of the orography on extreme rainfall depths and […] are still not sufficiently understood for sub-daily rainfall events” (lines 29-30). Specifically, since not much is known, but something has been done, I think that the state of the art should be presented with some more detail. Some examples include Panziera & al 2016, Papalexiou & al 2018, Rossi & al, 2020. Marra & al 2021 is also relevant and currently cited in the manuscript but, to my view, supporting a sentence that is not directly addressed by these authors, so it should be either removed or appropriately presented.
3. One of the main findings presented in the study (lines 14-17) is that elevation alone is as important as (or less important than) geographic location (lat, lon, distance from the coast) in explaining the average annual maximum precipitation, and that the impact of orography emerges more clearly at the local scales (lines 18-20).
   
   I have some trouble with this reasoning. I’ll try to explain: (a) what the authors present as a finding seems to me rather obvious: geographic location dominates over orography, unless we expect extremes in Caltanissetta (southern Sicily, 568 m a.s.l.) to be more similar to extremes in Aosta (western Alps, 583 m a.s.l., ~1000 km away) than to extremes in Agrigento (southern Sicily, 230 m a.s.l., ~50 km driving from Caltanissetta) due to the similar elevation. I’m not sure how this can be presented as a finding; my underlying hypothesis would be exactly what here is presented as a finding, i.e. that geographic location is more important than elevation, and that elevation modulates the extremes, for example by altering some properties of the storms that occur in a given area; (b) once one accepts the above hypothesis (or finding), the need for “countrywide” analyses on which the authors put emphasis in the introduction loses part of its scientific interest, while indeed maintaining vast practical interest.
   
   To my view, this aspect might require some work on rephrasing introduction and presentation of these results.
4. As the authors mention in lines 23-26, orography affects precipitation in rather different ways when one considers the lee side and the wind side of an orographic barrier. This is expected to differently alter the characteristics of extremes at multiple durations, as shown by Avanzi & al. 2015 for the very same region examined in this study, and by other authors in other regions.
   
   I wonder why this information is not included among the factors examined in this study (for example separating typically-wind sides from typically-lee sides), and whether it could represent a relevant explanatory factor in the used models
5. As the authors mention, only few "countrywide" analyses are published, and one of them is indeed in Italy (Avanzi & al 2015) – others are for example Papalexiou & al 2018 and Panziera & al 2016. However, since countries are characterized by vastly different spatial extents, orographic settings and climatic conditions, I suggest using a more general term to define and discuss the examined scale (e.g., continental/regional/local or something on this line, as opposed to “countrywide”).
6. Lines 97-99: did you check the intermediate durations? If yes, it would be important to mention the results in the text (one sentence is probably enough). Otherwise, one could object that monotonous behaviors between 1 hour to 24 hours were indeed reported in previous studies, but also important deviations from the simple scaling for durations close to 1 hour were found (e.g., Marra & al 2021), asking whether such deviations could imply non-linear or even non-monotonic changes between durations.
7. Line 176-177: why did the authors explore a radial limit of 5 km for the computation of the geomorphological factors? Are 5 km sufficient to explain local orographic impacts at multiple durations?

Minor:

Line 11: my understanding is that the study focused on the average annual maxima, perhaps it is better to directly state this instead of “average of rainfall extremes”, because extremes are subject to a number of definitions.

Lines 70-74: it should be perhaps mentioned explicitly that this is the “simple scaling” approximation

Line 120: “compared to”

The symbol used for elevation changes between z and Z. I suggest to choose one for consistency.

Line 180 and following: usually boldface italic fonts are used to represent vectors (https://www.hydrology-and-earth-system-sciences.net/submission.html).  

Line 186-188: something sounds off with this sentence

Line 202-207: I wonder whether the equations contain too many significant digits, but I leave this to the authors to assess

References

Panziera, L., Gabella, M., Zanini, S., Hering, A., Germann, U., and Berne, A., 2016. A radar-based regional extreme rainfall analysis to derive the thresholds for a novel automatic alert system in Switzerland, Hydrol. Earth Syst. Sci., 20, 2317–2332, https://doi.org/10.5194/hess-20-2317-2016

Papalexiou, S. M., AghaKouchak, A., & Foufoula-Georgiou, E. (2018). A diagnostic framework for understanding climatology of tails of hourly precipitation extremes in the United States. Water Resources Research, 54(9), 6725–6738. https://doi.org/10.1029/2018WR022732

Rossi, M. W., Anderson, R. S., Anderson, S. P., & Tucker, G. E. (2020). Orographic controls on sub-daily rainfall statistics and flood fre- quency in the Colorado Front Range, USA. Geophysical Research Letters, 47, e2019GL085086. https://doi.org/10.1029/2019GL085086

The authors here investigate short-duration rainfall extremes in Italy focusing on improving simple relationships with elevation. The form national multiple regression equations exploiting several geomorphological  covariates  and one climatic. At the national level this did not really led to significantly improvements and thus the authors proceed to several localized spatial clustering approaches identifying this need given the clear spatial pattern in local bias. The paper was joy to read, clearly written, easy to follow and offers an addition to the literature. Therefore, my comments are only minor and optional since the paper fulfills its goal as is.

In general, there are different definition on what is considered extremes. There so many references on extremes, and sub daily extremes, mean extremes, etc.  in the introduction but it is not crystal clear what these extremes are. Are POT values, annual/seasonal maxima, etc? Please clarify.

Section 2.2. Can you show the fitting the Equations 3-6? Especially the nonlinear 5-6.

Fig 2 please try to use the minus symbol for minus and not the dash as the “--" is not clear and it should be “- −“.

Out of curiosity, have the authors tried to see if there’s potential in using the temperature as a climatological variable?

The authors might be interested in the Moccia et al. 2021 (10.1016/j.ejrh.2021.100906) study (though at daily scale) analyzing a fine resolution gridded product over Italy and investigating extremes. The bivariate choropleth in Fig13 also shows elevation-rainfall depth relationship.  

Is beta-i a vector in Eq7?

Was there some preliminary examination in leading to select a multiple linear model? Maybe pair-wise scatter plots would give valuable information and might actually lead in selecting nonlinear relationships at least for some of the covariates. It might worth creating these scatter plots. Also, yes the test showed no collinearity, but let’s see this also in scatter plots.

210. So in essence MAR acts as a proxy of elevation, right? What is the correlation between MAR and elevation?

Maybe creating subnational regions would be improved by using generic spatial clustering algorithms. I mean based on some statistical properties and not necessarily based on the geomorphological classifications. There are many of them in the literature and could offer an alternative detailed assessment on the optimal number of subregions and on their extend. I hope I have not missed this, but have the authors considered creating such region by applying spatial clustering algorithms to the bias maps?

Summarizing this is a nice paper, adding to the literature and deserves publication.