My review can be found in the attached PDF.

My review can be found in the attached PDF.

Dear co-authors and Robert Šakić Trogrlić (editor),

Thank you for submitting your engaging manuscript, titled “Comprehensive multi-hazard risk assessment in data-scarce regions. A study focused on Burundi”. The paper presents a multi-component methodology for assessing multi-hazard risk in the context of Burundi, with potential applications to other data-scarce contexts.

I believe that the manuscript makes a useful contribution to the disaster risk reduction (DRR) landscape in Burundi through the application of multiple quantitative (as well as a brief mention of qualitative) methods. However, I agree with Reviewers 1 and 2 that the manuscript would benefit from restructuring and integrating the findings into the broader literature, including the implications for existing DRR policies and practices in Burundi. Additionally, to increase the relevance and uptake of the paper, I recommend that the authors engage with local experts and practitioners to refine the methods used in generating the multi-hazard map (Figure 12). Further details are expanded upon below.

Agreeing with the major-level comments provided by Reviewers 1 and 2, I have focused this review on medium- and minor-scale comments. I advise the following amendments:

Medium-level amendments

The discussion (Section 9) currently reads more like a results section; please integrate the findings into the existing literature. I recommend the following:

• Examine how your findings contribute to existing knowledge of Burundi’s risk context and the implications of your multi-hazard risk assessment for developing current DRR strategies.
• Offer practical suggestions on how your multi-hazard risk assessment can be integrated into existing DRR policies and practices in Burundi. I recommend that this involves consultation and knowledge exchange with local experts and practitioners to identify individual methods that can be refined, thereby strengthening the multi-hazard map (Figure 12). This engagement could be online if appropriate.
• In the additional material, please include the model code and numerical results for each component of the methods. In addition, please list the questions used in the semi-structured interviews and focus groups for the socio-economic vulnerability assessment.
• I recognise that you have described the limitations of the research throughout the methodology. However, I believe the discussion could be enriched with a summary of these limitations and how you plan to address these in future work.

Figures

General: Include the geographical area of focus in each figure caption (e.g., Burundi, Cibitoke province). This is valuable if a figure is viewed outside the context of the manuscript/paper.

Figure 1: Can the topographical categories (i.e., colours) be differentiated more clearly?

Figure 2: Be consistent with capitalisation in the figure.

Figure 4: Add north arrows and scale bars to each panel in the figure.

Figure 10: State what I, II, IIIa, IIIb, IVa, and IVb refer to in the figure caption.

Figure 11: Add units to the colline and provincial vulnerability scales. Please amend the typo as follows: “…population density increase 2050…”.

Figure 12: Include a north arrow and scale bar on at least one of the seven panels. Can you increase the resolution of this figure?

Figure 15: The final line in the figure caption is unclear: “…different global and return periods.” Should this be “…different global and ___ return periods.”?

Sections

Section 2.1, Line 93: Include the threshold criteria used in EM-DAT to determine what disaster events are recorded.

Section 2.1, Line 93: To aid the reader’s understanding of the challenges posed by hazard data scarcity in the context of Burundi, it would be useful to include a figure or table comparing the impact data documented across EM-DAT and the IOM’s Displacement Tracking Matrix (DTM) platform.

Section 5.1, Line 142: Why was the 99^th percentile of the time series chosen as the threshold for the Peak-Over-Threshold (POT) method?

Section 5.2, Lines 169-175: What is the source/are the sources of this information on the 64 flood events between 2000 and 2020?

Section 5.2, Line 180: Be cautious with acronyms: here, I presume DTM refers to digital terrain model; however, you previously defined DTM in the context of the IOM’s Displacement Tracking Matrix (DTM) platform in Section 2.1, Line 101.

Section 5.3.2, Line 228: Explain why you chose a heterogeneous Poisson Process model to address the lack of multi-temporal information in landslide observations.

Section 5.3.2, Line 235: What specific aspects of Corominas and Moya (2008) did you apply to determine the parameters of the power-law frequency-size distribution?

Section 5.4, Line 255: Provide a reference for the statement “…RP 475, which is considered the most representative return period for this hazard…”.

Section 6.1, Lines 271-272: Provide references for the following sentence: “The population in Burundi has grown from around 8 million people in 2008 (year of the last census) to around 12 million people in 2019 (projection of the National Statistics office).”

Section 6.2, Line 298: Quantify this “good match” between the results and the projections for 2019.

Section 6.3, Lines 308-310: Is there a method of validating this “mostly visual accuracy assessment”?

Section 7.1.2, Line 329: Add a brief justification for why the “HAZUS scheme from Kircher et al. (2006)” was the most appropriate choice for assigning the capacity curves.

Section 7.2, Line 356: Provide a reference for the following statement: “…an INFORM subnational risk assessment in Burundi was published in 2020…”.

Section 11, Lines 513-514: Expand on this sentence detailing the stakeholder engagement: “To mitigate this, it was attempted to validate the conceptual approach and data sources with local experts, but this was limited due to the difficulty in engaging relevant informants.”. This could be mitigated by expanding the discussion to include future work.

Minor amendments and typos

Section 5.3.2, Line 228: Change “landside” to “landslide”.

Section 6, Line 265: Mismatched parenthesis: “(see (Campalani et al., 2023)”

Section 6, Lines 264-265: The text could be rewritten to read, “Key exposure information considered in this study includes population, residential buildings, and cultivated areas; additional information on infrastructure was also collected (see Campalani et al., 2023).”

Section 6, Line 266: Should “local authoritative sources” be rewritten as “local authority sources”?

Section 6.1, Line 273: Change “In order to provide a realistic representation of current population distribution…” to “In order to provide a realistic representation of the current population distribution…”.

Section 6.1, Line 277 and in Figure 7: Change “Worldpop” to “WorldPop”.

Section 6.2, Line 287: Change “provided” to “providing”.

Section 6.2, Line 285: You have said “based on the following two information sources” and then listed three information sources.

Section 6.2, Line 291: The hyphen in “construction-types” is not necessary.

Section 6.2, Line 198: Change “…distribution data was…” to “…distribution data were…”.

Section 8, Line 397: Be consistent with UK or US spellings throughout the manuscript. Here you have written “color-coded” but on Lines 472 and 481 you have “colours”.

Section 9, Line 430: “…as expected giving the concentration of assets…” should be “…as expected given the concentration of assets…”. The same typo is present in Section 11, Line 498; please check elsewhere in the manuscript.

Section 10, Line 447: Change “national and subnational scale” to “national and subnational scales”.

Section 10, Line 451: Change “…assumes as baseline…” to “…assumes as a baseline…”

Section 11, Line 493: Please add the missing parenthesis: “…(in terms of their uppermost percentile of overlapping natural hazards.”.

Section 11, Line 495: Please change “intense precipitations” to “intense precipitation”.

Section 11, Lines 495-497: This sentence does not quite make sense, please alter: “A more comprehensive multi-hazard risk assessment was carried out, in terms of economic impact and also considering exposure and vulnerability, a different picture emerged.”

Section 11, Lines 504-505: Please remove the “on” in the phrase: “…and on consultations with…”.

Section 11, Line 528: Please change “…above 90th percentile” to “…above the 90percentile…”.

___

Many thanks for the time and effort spent in producing this manuscript. I hope that my comments will be helpful, and I wish you the best during the revision process.

Regards, Reviewer 3