Dear Nafisa Nuari Islam and co-authors,

Thank you for submitting your manuscript "Community-Scale Assessment of Flood-Related Public Health Vulnerability Using Multi-Criteria AHP in Northwestern Bangladesh". I apologize for the delay after the first round of reviews. I have taken over the handling of your manuscript and will serve as the editor for the next steps.

Your study addresses an important and timely topic at the intersection of flood hazards and public health vulnerability, with an interesting micro-scale framework and policy-relevant insights. To guide a constructive revision, I would like to invite you to submit a revised manuscript and a detailed authors’ reply to both reviewers. Please also include a track-changes version (this can be part of the authors’ reply).

Beyond the reviewers’ suggestions, I encourage you to address the following comments to improve methodological clarity, internal consistency, and reproducibility. I am aware that several changes will require some effort. However, I am convinced that doing so will meaningfully strengthen the manuscript. I do not expect every comment to be implemented, but please provide a concise, point-by-point response outlining the changes made and the rationale where you choose not to revise.


1. Material and Methods: Methodological clarity, internal consistency and reproducibility of sampling
========================================

(1a) Union count: the text says that data were collected from "six unions", but then seven are listed. Table 1 contains 6 unions. Please clarify and verify whether the correct set was used in the analyses.

(1b) Name inconsistencies: multiple spellings appear (e.g., Purba Chatnai/Chhatnai; Tepa Kharibari/Tepakhribari). Please standardize the spelling.

(1c) Sampling description inconsistency: The abstract state that 315 hoseholds were randomly selected, whereas unions were purposely selected by flood classes. This approach is reasonable, but please add a brief rationale for the selection of these Upzilas in the context of low, moderate and high flood-affected regions.

(1d) Please provide explicit descriptions of variables used in formulas. For instance, please add information on the years for x_0 and x(t) in the text, I think this will facilitate reproducibility and readability.

(1e) Projected population vs. households: Formula 1 explicitly refers to the projected population, yet the household number was used. Please whether household counts were derived from population projections, and how this was implemented.

(1f) Timeframe: The objectives explicitly emphasize the 2020 floods, while data are from a 2019 survey. Please clarify how 2019 data supports 2020-focused claims, or reframe the scope to match the data.

(1g) The stated ideal sample size of 270 is not reproducible with the information provided in the text. Please clarify: (i) the target parameter type (e.g., proportion; in my view it is not obvious here to what aspect the term "standard deviation" refers in this context), (ii) the assumed value used for variability (e.g., p used in p(1−p)), (iii) the z-score used for 90% confidence, (iv) whether any design effect or finite population correction was applied, and (v) if and how multi-union sampling affected the calculation. Without this information, the computation remains non-transparent.

(1h) The statement "Each data variables were collected through binary responses (yes=1, no=0) and Likert scales (table-2)" is misleading. According to Table 2, many of the variables are nominal or ordinal (not Likert items). Please correct this description.

(1i) Please correct the likely autocorrect typo in Table 2 ("15-Jun" should probably read "6-15").

(1j) Min–max nornalization requires numeric inputs per household. Please describe how nominal categories were coded to numbers. Also, it should be made clearer which categories were treated as benefit vs. cost (as this affects corresponding direction used for normalization). For reproducibility, consider adding (e.g., as a supplement or in the appendix):
1. The recoding scheme from categories to numeric values at the household level (e.g., ordinal score, one-hot encoding).
2. If applicable: The rationale for any ordering assigned to categories.
3. The benefit/cost designation for each indicator
4. Optional: The min and max used in the normalization for those derived variables.


2. Material and Methods: AHP
========================================

(2a) The consistency ratio formula is incorrectly presented as CR = CI/CR, which is self-referential. Please correct the formula and check the CR calculations.

(2b) The AHP description introduces both "A" and "AA" matrices with identical definitions of elements a_{ij}. This is confusing, please streamline to a single clear definition and remove redundancy.

(2c) Please add a concise description of how expert judgments were aggregated (e.g., geometric mean) and summarize inter-expert variability in the results section. This will help assess robustness.

(2d) Vulnerability formula notation: Consider cleaner notation for category indices (e.g., V_C for each category; VT = sum of VC across all categories) to avoid undefined symbols.

(2e) In the formula in section 2.7, please use epsilon (ε) for the error term instead of the Euro symbol (€).

(2f) Subsection 2.8 is rather short. If the spatial interpolation is retained, please justify the use of IDW and provide methodological details (e.g., power parameter, search radius, number of neighbors). Include a validation (e.g., cross-validation) and discuss limitations and artifacts common to IDW (e.g., point-centered structures), as the manuscript currently does not cover this.


3. Results: Statistical reporting and interpretation
========================================

(3a) Please add the sample size (n=315) to the relevant tables and figures in section 3. This is an important information, especially since results are presented as percentages.

(3b) I suggest to clarify in the captions of Figure 3 and Table 5 that these percentages are based on reports of respondents.

(3c) Figure 4: Please reconsider the color scale for vulnerability, the different shades of green are hard to differentiate.

(3d) Please distinguish exploratory results from inferential claims. Avoid terms like "significant variations" (l.310) without a statistical test.

(3e) Figure 5 shows patterns that warrant a discussion of IDW limitations, manual class breaks, and parameter choices (e.g., power parameter, alternative classifiers). Please include rationale and sensitivity checks, and discuss point-to-surface inference limitations for household-level indices. See also comment (2f).

(3f) The riverside (right-bank) vulnerability claim is persuasive qualitatively. Could you support this with simple spatial summaries (e.g., zonal means by distance-to-river bands) to complement the qualitative evidence?

(3g) The regression table (Table 8) states p = 0.00 and "all factors hold only a 1% error". I suggest to replace with exact p-values (or p < 0.001) and define what "1% error" refers to (e.g., significance level, CI coverage). Please specify the model type, diagnostics, and assumptions (e.g., multicollinearity, residual checks).

(3h) "The model shows that it includes 99.6% of data (R²)". This sentence should be corrected to reflect that R² denotes the proportion of variance explained. Please consider reasons for the high R² (e.g., aggregation, limited variability in some indicators) and discuss potential overfitting.

(3i) Coefficients are close in magnitude. Please discuss practical significance, and consider presenting standardized effects or confidence intervals to support the ranking of drivers.


4. Discussion and conclusion
========================================

Please consider expanding the discussion to more fully reflect your methodological choices (see comments above). In addition, you could examine the interplay between structural and non-structural factors, discuss how institutional support can mitigate high flood exposure within a DRR framework, and articulate the broader scientific contribution of the study. It would also be valuable to highlight the transferability of your insights to other flood-prone contexts and outline how the knowledge gained can inform policy and practice beyond the study area.


5. Presentation and referencing
========================================

Please review the submission guidelines before resubmitting the manuscript, and ensure that formal requirements are met (https://www.natural-hazards-and-earth-system-sciences.net/submission.html).

(5a) Referencing: When multiple references are listed at a position, all references should be put in parentheses separated by semicolons. In addition, several reference entries appear malformed or truncated in the current version (e.g., fragmented citation text), and some contain conflicting information (e.g.: "Ferdous & Mallick, 2019" contains the info "(June 2018)" after the volume number); please ensure that the reference list and in-text citations are complete and consistently formatted.

(5b) Figure captions: Figures should be numbered "Figure 1" instead of "Fig01".

(5c) Tables: Please harmonize styling and font; use horizontal lines only above/below the table and between head and body, per guidelines.

(5d) Formulas: Please number the formulas, and reference them explicitly in the text to avoid possible ambiguity.

(5e) Please spell out acronyms upon first use (e.g., WASH)

(5f) Please check for and resolve minor editorial issues (terminology, union names, typographical errors) for readability.


As stated above, I do appreciate that several of the requested revisions will take time and care, but I am confident that addressing them will substantially enhance the clarity, rigor, and impact of your study. Please don’t hesitate to contact me if any questions arise during revision.

I look forward to receiving your revised manuscript and authors’ reply, which I will then send for further review.
 
Regards,
Matthias Schlögl
[NHESS Editor]

Dear Ms. Nafisa Nuari Islam and co-authors,

Thank you for submitting your revised manuscript. I have now received two independent reviews. I continue to view the article as promising: it addresses a highly relevant topic in an under-researched and vulnerable setting, and it has the potential for meaningful societal impact.

Thank you for your continued efforts on this manuscript. Both reviewers have offered thoughtful suggestions, and I would be grateful if you could incorporate them in your next revision. In particular, please refine the discussion to situate your findings more clearly within the broader NHESS context and to highlight their wider relevance. While I understand that a full sensitivity analysis may not be feasible at this stage, I encourage you to expand your consideration of the sensitivity and robustness of your results in the discussion section. A more structured and balanced treatment of the study’s limitations would indeed strengthen the manuscript. Please note that this is not a critique of your study design, but rather a request to provide an appropriately framed discussion of the applied method (i.e., AHP).

I remain positive about the manuscript’s suitability for publication in NHESS once these points are addressed.

I look forward to receiving your revised manuscript along with a detailed response to the reviewers’ and editor’s comments.
Upon reviewing your revision and response, I will determine whether further peer review is necessary.

Kind regards,

Matthias Schlögl
[NHESS Editor]

Dear Ms. Nafisa Nuari Islam and co-authors,

Thank you for submitting your revised manuscript. I have received one report from a previous reviewer and have re-read the paper myself. To avoid further delay, we will proceed without a second review at this stage.

We will, however, need one more round of revision. Please address the following points:

(1) Discussion: As already mentioned in previous feedback, I'd like to ask you to refine the discussion to situate your findings more clearly within the broader NHESS context and to highlight their wider relevance. In particular, connect your results to hazards and risk reduction debates (e.g., how household-scale AHP results can inform multi-level DRR planning, early warning-to-response pipelines, and health systems resilience), beyond the current case framing and policy bullets. The manuscript already argues transferability; please comment on the the generalizable mechanisms and diagnostic use (structural deficits vs. capacity gaps vs. coordination failures) and briefly illustrate how they could be applied in other flood-prone contexts. Clarify how your micro-scale results complement district/regional analyses and the methodological streams you cite, including how benefit/cost normalization and weighting choices map into the composite and why this matters for integration with larger-scale assessments.

(2) Limitations: The paper now lists methodological and scope limitations as bullet list; please reorganize them into paragraphs more suitable for the discussion section - or at least provide some better context and framing around the bullet points. For instance, you could distinguish data limitations (self-reported recall; lack of hydrological monitoring), methodological limitations (no formal sensitivity analysis; expert-subjectivity despite CR controls), and external validity (single cross-section; riverine context; institutional context time-variance). Where possible, you could indicate the likely direction of bias (e.g., recall bias toward recent/severe events) and implications for interpretation of the AHP-weighted vulnerability scores. This would clarify how readers should generalize the findings.

(3) Interpretation: Please review the causal language when interpreting findings. Indicate/discuss more clearly where evidence is observational/correlational (e.g., correlations among categories, union-level comparisons), and where inference is conceptual. Make explicit that the phase-specific pathways are proposed mechanisms supported by the observed patterns and expert-weighted structure rather than tested causal models.

(4) Figures: Please ensure to use a consistent font in all figures (currently, different serif and sans-serif fonts are mixed). Correct spelling and formatting issues (e.g. fix the line break in "infrastructur-e" in Figure 2). Also, I suggest to use colors (1) more sparsely in general, (2) more consistently across figures (currently, every figure does have a different color palette), and (3) in a manner suitable for readers with color vision deficiencies.

Please submit a clean, tracked-changes version and a brief response letter outlining how each point has been addressed. I look forward to receiving your revision.

Kind regards,

Matthias Schlögl
[NHESS Editor]

Dear Ms Nafisa Nuari Islam and co-authors,

Thank you for your thorough responses and for your continued work on the manuscript. I am pleased to inform you that your manuscript has been accepted for publication in NHESS.

I have one minor technical note for the production files: when preparing the publication-ready figures, please ensure the font in Figure 3(a) matches the others (Calibri).

Our production office will be in touch shortly to guide you through the final steps, including typesetting and proof review.

Thank you again for choosing NHESS for your work.

Best wishes,

Matthias Schlögl
[NHESS Editor]