This is a very interesting manuscript, with a clear aim and methodology. The latter is well executed, with a comprehensive set of results and evidenced discussion. The authors make very good use of existing literature to support most of their writing. Overall, the manuscript offers an interesting and useful addition to the literature. It demonstrates well the potential of blending multiple evidence sources to develop richer understanding of hazard events. Attached I include some specific and technical comments that I believe would strengthen this manuscript. The specific comments focus on the multirisk framing, information about the cascading multi-hazards, and how the concept of 'vulnerability' relates to the database. 

I would like to thank the authors for a highly interesting and original manuscript. I very much enjoyed reading it.

Comments:

• Page 2: “A specific concern is related to complex and/or cascading hazards that can have far-reaching and multiple consequences downstream” -> I think it would be good to provide a very clear definition of multi, consecutive, cascading, etc risk. Similarly, a couple of sentences down (risks related to natural hazards are broadly defined here as the results of interactions between natural and societal components), I wonder if it would be good to explicitly define is as exposure and vulnerability (and risk) very clearly? And eg in L. 329 you talk about “cascading multi-hazards”.
• The authors speak of a multirisk database that covers centuries, but it covers hazards and impact, right? It would be good to clarify this; I also found section 3.4.1 a bit unclear.
• Out of curiosity: have definitions of individual hazard types changed over time (eg land movement or rockfall)?
• 4.2: by event, do you mean disaster? (so when a hazard interacts with exposure and vulnerability of that exposure)? In the next section you talk about impacted assets, but does this mean that you exclude vulnerability?
• Looking at the analysis in section 4.5, I wondered why we see such a peak in number of disasters around 1950-1970?
• I would be curious to read a bit more of an in depth/critical reflection on and discussion of the notion of single vs (cascading?) multi-risk and the recording (and hence perception) thereof.

Minor comments:

• 7 L. 190: studied period -> should be study period
• 275: “mobilized sources” -> do you mean used sources?
• 293: “torrents” -> do the authors mean torrential rainfall?
• You write multirisk and multi-hazard, I would hyphenate both.
• 636: I am not sure I would call the frequency figures a statistical analysis.