The paper examines public-sector compensation in disaster recovery, focusing on the 2020 Gloria storm in Catalonia. It highlights the importance of public compensation for uninsured assets and provides insights into financial aid distribution for disaster recovery. The study reveals that fluvial and coastal hazards caused over 80% of recorded damages, with infrastructure sustaining the highest losses. Public funding prioritized community assets and critical infrastructure.

The paper details the data used and the methodology with great precision and presents clear and concise results. The case study seems very appropriate to me. In addition to good conclusions. For all these reasons, I recommend its publication. I simply add some personal recommendations that the authors may or may not consider:

• In line 226, Gumbel is mentioned but its use is not justified as it is done with GEV previously.
• In figure 6, it would be advisable to add the letter labels to know what the caption refers to (a, b, c…).
• In the conclusions, add future contributions following the line of research.
• If possible, a graphical diagram of the methodology used in the paper.

 Congratulations to the authors for their work, I found it very interesting.

This paper presents an analysis of public-sector compensation following Storm Gloria in Catalonia, Spain. The manuscript consists of three analyses focusing on: (1) losses, (2) hazards, and (3) the intersection of losses and hazards. A direct economic loss database was derived by categorizing public funds that were allocated towards recovery. The hazard analysis focused on three hazards (meteorological, coastal, and fluvial) and consisted of evaluating the spatial extent of hazards and determining hazard return periods. The intersection of losses and hazards evaluated the relationship between losses and hazard return period. The results shows that fluvial and coastal hazards are the major driver of losses for Storm Gloria, and that infrastructure sustained the largest losses. The manuscript shows no relationship between losses and hazard return period.

This manuscript is organized, well written, and easy to read. The reviewer recommends revisions prior to acceptance in Natural Hazards and Earth System Sciences. The reasons for this recommendation, and suggestions for improvement of the manuscript, are outlined below.

Major comments:

1. Line 148: The introduction to the methods section outlines a three-step methodology; however, the methods section only consists of two subsections (3.1 Direction Loss Analysis and 3.2 Hazard Analysis). Is there anything that can be said in the methods about the third step in the methodology? This doesn’t need to be long. It would help readers know what to expect further in the manuscript. As it stands now, line 151 (the one-sentence description of the third step) isn’t clear to me.
2. Line 229: Why is the return period for meteorological hazards of each pixel assigned to the nearest return period available from the Catalan meteorological service (2, 5, 10, … 500-yr)? That is, in the example given, why is the interpolated return period of 29-year event re-assigned a return period of 20-years? This does not appear to be consistently done across hazards (see line 255 where the return period for fluvial hazards is estimated as 12-years and not re-assigned).
3. Line 269: Can the authors provide a justification for using Block Maxima method as opposed to something such as peaks over threshold?
4. Fig. 6f: Am I viewing this figure correctly in that the largest possible number of effected elements in each municipality is only 9? I would assume there are more elements/assets in a municipality. It would help readers if the total number of entries in the compiled database is clearly presented throughout the manuscript (maybe a new column Table 1 with "number of elements" or shown somewhere in Fig. 4?).
5. Fig. 7: Are the return period maps shown in Fig. 7 all produced using the same linear distribution that is shown in Fig. 3 (e.g., the 48-hr duration)? If so, why are the 1-hr and 12-hr maps using the 48-hr duration return period estimate? If the maps are produced using their own duration return period estimates, why are these so drastically different? I would assume that there should be more similarity in return periods across for different durations.
6. Line 488: The authors suggest that there are “potential geographical variations in historical records across these [north-south] regions”. Is this not something that can be directly determined rather than using terms “suggesting” and “potential”? It appears the authors have hourly wave data at 52 nodes (Table 2) and could perform an extreme value analysis at each node to determine if there are indeed north-south variations in the historical records. This isn’t something that needs to be shown per se, rather something that could be definitively stated. It’s additionally not clear why the return periods of significant wave heights (I’m assuming this is referring to the observed significant wave heights for Gloria) are used to justify this statement. That is, significant wave height can vary along the coast for a single storm.

Minor comments:

1. Line 12: Could this sentence be rephrased? I was expecting “losses” to be computed from “damages”; however, I did not see this in the manuscript.
2. Line 29: The authors use USD here, but EUR elsewhere (e.g., line 289). If possible, I’d suggest staying consistent across this manuscript.
3. Fig. 1: I’d suggest the authors consider adding a small inset map showing where Catalonia is located in Spain / Western Europe. This is not a requirement, but something to consider.
4. Line 286: I’d suggest the authors consider removing “on one side” (e.g., “…we present the density distribution (Fig. 4) and the ___ (Fig. 5)”.) This currently reads as if the two plots shown are on the same figure.
5. Line 318: “first-level” not “firs-level”.
6. Fig. 6: Each subplot is missing a, b, c, etc.
7. Fig. 8: Is it possible to label the other rivers that the authors discuss in the text? Namely the Tordera, Fluvià, and Besòs. This would make it easier for the readers to identify which rivers the authors are discussing without having to refer to Fig. 1. Otherwise, this is a great figure.
8. Fig. 10 and line 504: Can the authors provide more information on what is meant by “standardized return period”?
9. Line 561: “have” should be “has”.
10. Line 618: Check on the word “fist”. Could this be removed, or should this be “first”?