The paper provides a very interesting and timely methodological approach for identifying and compiling information on interrelationships among single and multi- natural hazards. Overall the paper is well written and well positioned for a place within the Special issue: Methodological innovations for the analysis and management of compound risk and multi-risk, including climate-related and geophysical hazards. Some minor considerations for the authors:

Introduction:

Overall this section is clear and concise, however, a few parts of the narrative could flow better and need clarification. 

In terms of data-scarcity on hazard impact data, and this quote: “One of the significant challenges facing urban areas in Nepal is the scarcity of hazard impact data, which is a barrier to effective DRR strategies (Chatterjee et al., 2015; SIAS, 2016).”  The first thought that came to my mind is wondering the extent which this has changed almost 10 years since the 2015 Earthquake and from the sources referenced. And if we are talking about scarcity of multi-hazard impact data, I wonder how Kathmandu differs so much from other urban settings, even in the West. 

On lines 84 and 95 the authors note the Kathmandu experiences many single and multi-hazards events without qualifying it. Maybe the examples under 110-120 could be brought up under points 84 and 95 to create a more coherent narrative. 

From lines 95 down, it is not immediate clear why the authors are highlighting the squatter settlements, and even including a visual. Is it only to illustrate population growth and growing urbanisation, or is it related to the heightened risks and impacts on vulnerable groups? This is connected indirectly to the Urban Fire bullet point, and then again later made clearer in the Results and Discussions sections with the focus on impacts to marginalised communities. I suggest the authors these connections more clear in the Introduction when introducing the section on squatter settlements. 

Methodology:

Only using English sources is certainly one of the main weaknesses of the study. But it is very good that the authors have recognised and highlighted the limitation in a number of places within the document. 

What remains unclear is what is meant by “grey literature”. Perhaps I missed it, but I find only one small explanation on line 608 which references “e.g., UNDRR reports”. Did the authors include in their literature search any other NGO/IGO and government agency reports, and perhaps preparedness plans at national/regional/local/community levels? They could likely find more hazard specific frequency magnitude/impact information through specific agencies in the regions, for instance NSET for geophysical hazards. https://www.nset.org.np/nset2012/ . These types of sources seem very relevant to have been included considering the scope of the paper, 

Line 214 notes: “We used the methodology described in the literature review in Sect. 2.1 to conduct Boolean searches for single hazards and multi-hazard interrelationships that might occur in Kathmandu Valley.” Did this also include events which have already occurred? I suggest to make that clear. 

I appreciate the thought behind pulling out the column headings into separate boxes for ease of reading in Figures 2 and 3, but I think it might create even more difficulty for the reader. Or at least it makes the presentation of the Figures look quite messy. Consider if those extra boxes are really needed. Perhaps instead just highlight the relevant text in different colours in the tables. 

Line 293 notes: ”These participants were selected based on their in-depth knowledge of the Kathmandu Valley context and existing connections built on pre-established working relationships.” What does this say about their specific backgrounds and expertise contributing to this topic? Especially given the small number or participants, this clarification is needed in the methodology section. I found the explaination (in part) only at the end of the document in the Limitations section. 

Results:

Figure 5 is great! It also made me quite curious about some of the relationships. For instance I had to look up how a hailstorm can trigger a volcanic eruption. 

Line 407 notes: “Additionally, we shared Fig. 6 with workshop participants to illustrate the value of the  multi-hazard interrelationship matrix in extracting relevant multi-hazard scenarios.” It is not clear if they were asked for feedback on that Figure, or if it was also used as a tool in the workshop? E.g. What did they think of it? Was it useful?

Line 462 notes: “Including anthropogenic hazards and related processes could form the basis of future developments of this work.” That seems like a fairly significant outcome, and should probably be highlighted in sections 4.6 and 4.7 on future work in this direction. 

Discussion:

In section 4.3.2. It could be interesting for the authors to reflect on HOW the different types of variables related to vulnerabilities and impacts (also within different spatial and temporal scales) could be considered within such a matrix. The authors do provide some considerations for this in sections 4.6 and 4.7, maybe there is a way to also provide a few of these reflections already in 4.3.2. 

4.6 and 4.7 are really nice to read. Two things that do however seem to be missing (or not explicitly enough) in terms of scalability of the matrix: should scalablity take into consideration other types of risks and hazards (not only natural)? And I miss some clear recommendations on what could be the best way forward for evolving the survey in the context of Nepal. Having done the first leg of the work, which other sources and local champions should be involved in Nepal to improve the tool?

This study presents a methodology to identify single and multi-hazards using the Kathmandu Valley, Nepal, as case study. The methodology contains different approaches, such as searching academic literature, grey literature, media, databases, and social media, but also involving directly stakeholders though workshops. It is important to have this type of data, especially for regions with less resources. I particularly liked the use of the multi-hazard data during the workshop to work on potential DRR strategies. I think the study is aligned with the scope of the journal. I have minimal technical remarks on the methods, but some on the discussions.

The manuscript is well written but also very long. I believe part of that is due to repetitive language that could be avoided and made more direct. There are too many cross-references (ex: “This will be shown in section XX”; “… this was shown in section YY”). While I would not expect significant changes on the writing for this study, my recommendation for the authors on future work is to keep things direct, linear and to the point.

Most crucially for me is that the discussion sections 4.1, 4.2 and 4.3 read as a summary or continuation of the results. Instead, the discussion should offer interpretation and reflection on the results (what do they mean?) and comparisons with past findings (literature).

Detailed comments:

Line 48 to line 52 seems to be redundant, maybe a copying mistake?

Line 78-79: Maybe briefly define technical terms, such as hazardscapes (line 78) and (Multi-hazard) interrelationship (line 79).

Section 3.2: It is an interesting demonstration of the multi-hazard table (Figure 6) for designing causal diagrams. While not the aim of this study, the actual representation of these causal diagrams should include also direct effects. This increases the complexity of the multihazard events & impacts. In the examples provided, a storm also directly causes flood and extreme temperature also directly contributes to urban fires, and they should be included in the diagram.

Section 4.1: I do not follow what is meant by frequency magnitude relationship. In fact, I did not see a discussion on the frequency, but a quick description of some hazards that happened in the past. This is not informative. I think it’s important to address: what do those numbers suggest? Why are some hazards more common than others? What should a stakeholder or reader take out from your analysis on the different hazards?

Section 4.2: Your two challenges: “Globally, there is a focus on reporting and describing single hazards instead of detailed information on hazard interrelationships. // Globally, research and understanding of single hazards are more established than for multi-hazard interrelationships.”

 These are obvious remarks, and should be the starting point of the discussion and not a conclusive remark. The following sentences are just continuations of the results, mentioning how much low cases of multi-hazards are found. That is a result, not a discussion. You should aim for answering questions like: What does the literature say about it (more specifically: is this more common in the global south? Are there regions where this is not an issue? What have other studies done to overcome this?); Why is it harder to find detailed information on hazard interrelationships? How do your findings fit in the general scheme built in the literature?

Section 4.3: Even though there is more contextualization in this section, I think it is still a long section describing results, instead of really discussing them. A direct paragraph discussing the implications of the findings without describing results would be an improvement.