Dear Reviewer,

Thank you for your thorough review of our manuscript and for recognising the importance of our research topic. Your comments have prompted us to critically reflect on our methodological choices, the scope of our findings, and the clarity of our writing.

Our study draws on a continent-wide survey of water managers conducted during the 2022 European drought, and our results align with existing research on drought risk management, as stated in the manuscript, while providing a rare snapshot of real-time on-the-ground management responses to an exceptional, continent-wide event. Although not immediately evident in the manuscript, this survey was the outcome of a long and complex process based on the one used in the paper "Lessons from the 2018–2019 European droughts: a collective need for unifying drought risk management" by Blauhut et al. (2021). It involved the design of the survey, its translation into 19 languages, and its dissemination across multiple countries, all thanks to a network of volunteers. The distribution strategy combined elements of both targeted outreach and broader dissemination. Initial contacts were established through members of the IAHS "Drought in the Anthropocene" group and their local collaborators, who then distributed the survey to relevant stakeholders within their networks. Where possible, this was followed by more systematic outreach at the national level - for example, in Sweden, the survey was sent to all municipalities, while in Spain and Italy to all basin authorities - in other contexts, contacts were identified through structured web searches of institutional directories and professional networks. 

This hybrid approach goes beyond a classic snowball sampling strategy – an important point we will clarify in the revised manuscript. While we agree that the sample is not the result of fully randomised or uniform sampling across countries, the inclusion of national outreach strategies and targeted expert identification lends a degree of sectoral and regional representativeness, especially where systematic efforts could be applied. We also acknowledge that the extent of coverage varies between countries, and we do not claim statistical representativeness or imply a normal distribution of the variables in the dataset and make these limitations more explicit in the revised text. However, we believe this does not undermine the overall value of the dataset, which offers rare and valuable insights into on-the-ground drought preparedness, perception, and response. This is especially significant given the lack of alternative pan-European drought datasets that systematically capture such information for the 2022 event, or similar events in the past. Given the study’s aim to capture informed professional perspectives amid an ongoing crisis, the expert-targeted approach is well aligned with its purpose and provides a strong foundation for generating policy-relevant insights. 

We took steps to mitigate the limitations of our sampling approach. For instance, we included only countries with at least ten valid responses. Moreover, our comparison of drought severity (SPEI) with perceived effectiveness and timeliness provided in the manuscript shows that the responses are not merely a reflection of meteorological conditions, but instead capture meaningful differences in management practices. While these measures cannot fully eliminate the inherent limitations of our sample, they do reflect our commitment to maximising the robustness and relevance of our analysis. These arrangements will be further discussed in Section 2.1.3 (Limitations of the dataset) of the revised manuscript.

We will revise our manuscript to more explicitly state that our conclusions offer exploratory insights rather than definitive population-wide generalisations. Although the sample may not support broad inferential statistical generalisations for all of Europe, it does enable solid descriptive statistics and comparative analyses, especially when national differences in dissemination strategies are taken into account. In line with this, we have emphasised descriptive figures and statistics throughout the manuscript and complemented them with statistical tests. In our revision, we will clarify that these tests should be interpreted cautiously and seen primarily as indications. By acknowledging these limitations transparently, we hope to encourage further research that can build on our work using more representative sampling approaches. 

We sincerely thank the reviewer once again for the constructive feedback, and we will ensure that our revised manuscript offers a clearer and more transparent discussion of our methodological choices and their implications.

Sincerely,

Riccardo Biella on behalf of all co-authors

We thank Reviewer 2 for the positive evaluation of our study’s scope and contributions, and for the constructive comments that helped strengthen the manuscript’s conceptual clarity and methodological transparency. Below we respond to each of the main points, followed by a brief note addressing the minor issues raised. Reviewer comments are listed in italic, followed by our responses in plain text. Line numbers refer to the revised manuscript with tracked changes (which will be uploaded at a later stage as requested by the revision process of this journal).

Comment – Clarify sampling limitations and avoid overgeneralization

In my view, some of the hypotheses may not be fully or precisely tested with the available data, and this should be considered when interpreting and discussing the results.

We have added a dedicated paragraph in the Limitation (Section 2.1.3, lines 271, 278-283, and Section 2.3 lines 329-331), and Discussion (Section 4.4, lines 784-789), as well as throughout the manuscript to reinforce that the analysis is descriptive in nature and that generalizations beyond the sample should be made with caution.

Comment – Add numerical examples to the abstract

The abstract could bring one or two numerical examples to exemplify their findings.

Numerical examples were added to the Abstract to better illustrate our key findings (lines 46-49, and 50-51).

Comment – Structure the introduction more clearly

I suggest structuring it into subsections to better guide the reader, as follows: (a) Contextualize the 2022 European drought; (b) Define drought risk assessment and indices; (c) Outline drought risk governance frameworks in the EU; (d) Present the novelty and main contributions of the study for the European context.

The Introduction has been restructured with clear subsections as suggested by the reviewer.

Comment – Address limitations of using SPEI-6 for regional comparisons

This approach may introduce bias, as droughts vary significantly in timing, duration, and severity across Europe.

We added a brief note in Section 2.2 (lines 300-304) explaining the use of a single time window across diverse European regions with variable drought patterns and clarifying its limitations.

Comment – Clarify how preparedness, awareness, and effectiveness were assessed

It is unclear which questions or items were used to evaluate each of these aspects.

We now clearly specify the questionnaire items used to derive these variables in Section 2.1.1 (lines 200-228), making the link between survey items and analytical variables explicit. This was briefly repeated in the Results (Section 3.4.1, line 592).

Comment – Justify how Hypothesis 3 is tested and interpreted

I’m not fully convinced that the authors are effectively testing Hypothesis 3. Could the authors further elaborate or clarify their reasoning behind this assessment?

We agree that the analysis would have been more robust with additional data on perceptions, particularly from "non-drought" years. However, the primary aim of this test is to explore whether perceptions of effectiveness are shaped by long-term memory (that is, whether there is a lasting influence from earlier events such as the 2018 drought) or whether they are more influenced by recent experiences and short-term memory, such as the severity of the 2022 drought. Our working assumption is that perceived effectiveness should primarily reflect levels of preparedness, including factors such as water governance, management plans, and available financial resources, rather than the severity of individual drought events. However, the strong correlation we observe between 2022 drought severity and perceived effectiveness in 2022 suggests that recent events exert a dominant influence on managers' perception. This indicates that short-term memory may introduce a bias in how effectiveness is evaluated. We briefly clarified this in Section 2.3 (lines 337-339).

Minor Comments:

We have corrected typographical errors, clarified percentage inconsistencies, fixed figure captions, and adjusted the introduction for flow. We thank the reviewer for spotting these errors.

Final Statement

We thank the reviewer again for their valuable and constructive feedback. The manuscript has been significantly improved as a result. We hope that the revised version meets your expectations and look forward to your final decision.

With kind regards,

Riccardo Biella

On behalf of all co-authors

We thank Reviewer 3 for their positive evaluation of the manuscript and for recognizing its clarity and organization. We also appreciate the thoughtful comments and suggestions, which have helped us improve the precision and transparency of key methodological aspects. Below, we respond to each point in detail and indicate how the manuscript was revised accordingly. Below we provide a detailed, point-by-point response to all reviewer comments. Reviewer comments are listed in italic, followed by our responses in plain text. Line numbers refer to the revised manuscript with tracked changes (which will be provided at a later stage in accordance with the revision process of this journal).

Comment – Consider categorizing responses by mission (responsive vs. strategic)

Responses to the survey are categorised according to state, type of structure... but not according to mission (responsive versus strategic). We can expect real differences. Why didn't you make this distinction in the analyses?

We appreciate this suggestion. In the revised manuscript, we clarify in Section 2.1.1 (lines 250-256) that drought management plans (DMPs) were classified as either responsive (short-term) or strategic (long-term). This was also more clearly described when presenting the structure of the survey (Section 2.1.1, lines 202-203, and lines 211-215). While both plan types were included in our analysis, we now also acknowledge in the Discussion (Section 4.4, lines 789-792) that many organizations reported having both types of plans and forecast type, which limited our ability to disaggregate their individual effects on preparedness and effectiveness.

Comment – Discuss variability in national forecasting systems

Do all states have a national forecasting system? These aspects must be discussed in a more developed section on the limits of this work.

We have added a sentence to the Discussion (Section 4.4, lines 792-795) acknowledging that forecasting systems vary across countries, and that the absence or presence of a national forecasting system could influence how organizations perceive and implement preparedness measures.

Comment – Address possible outlier status of Sweden (SE)

The answers for SE seem very particular (e.g. F3). How can you explain it? They may influence correlation (e.g. Fig. 7b, Fig. 10c-d).

Sweden stands out in the dataset, as Swedish respondents rated the effectiveness of drought measures significantly higher than most other countries. While the exact reasons for this perception are not fully clear, several plausible explanations exist: (1) Milder drought conditions: In 2022, drought impacts in Sweden were generally less severe than in many other parts of Europe and were largely confined to the southern regions. This may have influenced the overall perception of response effectiveness. (2) Increased preparedness following past events: Sweden has faced several droughts since 2015, with the severe 2018 drought serving as a key turning point. In its aftermath, substantial investments were made to improve drought preparedness, particularly at the municipal level. These included hiring climate and water strategists, raising public awareness, building system redundancies, and establishing backup water reserves. As shown in other studies as well, such efforts have demonstrably improved drought preparedness, which likely contributes to the more favourable evaluations of effectiveness.

Despite Sweden's relatively high ratings, there is no statistical justification to treat it as an outlier. Its values remain within the commonly accepted range for identifying outliers (i.e. within 1.5 times the interquartile range above the 75th percentile). Therefore, excluding Sweden from comparative analysis or discussing it as a "special case" is not appropriate, as natural variation across countries is to be expected. Moreover, Sweden had a relatively high number of respondents, which adds to the reliability and stability of its ratings.

Comment – Clarify timeliness calculation and use of median

How was timeliness computed? How are answers 'Before March' considered in the calculation? The mean cannot be calculated using truncated samples; in this case, using the median is more appropriate.

In response, we revised the Methods Section to explain how categorical responses (e.g., “Before March” or “After September”) were numerically approximated to enable analysis. This certainly poses a limitation to the analysis, and we have more explicitly acknowledged that (Section 2.3, lines 312-320).

Following the reviewer’s suggestion, we also replaced to mean timeliness with median. This meant replacing the values throughout the text and figures where appropriate (Figures 6, 7 and 8), as well as all references to the values of timeliness of in Section 3.3.4.

Despite changing from mean to median, no significant changes in the main analysis occurred. This is mainly due to the main analysis being carried out using the unpaired Wilcoxon test, which does not depend on the value of the mean. Only two minor differences occurred. First, the median difference in response timeliness went from 1 month when using the mean, to 2 months when using the median (as stated in the Abstract, line 48, and Results, lines 566-567). Second, the significance of the link between SPEI6 and Timeliness decreased for the 2018 drought, and increased for the 2022 drought. Consequently, the relation between the SPEI6 of the 2022 drought and the Timeliness is now significant. This was clarified in the Results (Section 3.3.3, lines 556-557), and Discussion (Section 4.3, lines 748-754).

Comment – Typographical and formatting issues

We have carefully reviewed and corrected all typographical and formatting issues noted, including removing or correcting the incorrect section reference, and fixing punctuation and spacing errors.

Final Statement

We thank both reviewers again for their valuable and constructive feedback. The manuscript has been significantly improved as a result. We hope that the revised version meets your expectations and look forward to your final decision.

With kind regards,

Riccardo Biella

On behalf of all co-authors