the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 21 Jul 2026
Identifying regions of Europe which may be less prepared for extreme rainfall
Abstract. Societal preparedness to extreme rainfall events varies geographically. We present a method to assess assumed preparedness based on observational data alone, investigating record daily rainfall across Europe. To define assumed preparedness, we use two metrics: 1) how recently the record daily rainfall occurred; and 2) how extreme the current record was. In locations where the record occurred further in the past, societal memory of the impacts may have faded, possibly reducing people’s interest in preparedness for extreme rainfall. In regions where the current record was not very extreme, a record-smashing event is statistically more likely, yet the region has not experienced such an event so may be less adequately prepared. We show that cities such as Sofia, Barcelona, Munich, and Amsterdam may have low levels of assumed preparedness. Such locations could benefit from communication about possible risks that the population might not be aware of, encouraging adaptation before a record-breaking event arrives.
- Preprint
(1340 KB) - Metadata XML
-
Supplement
(830 KB) - BibTeX
- EndNote
Status: open (until 02 Sep 2026)
- CC1: 'Comment on egusphere-2026-4100', Rasmus Benestad, 03 Aug 2026 reply
-
RC1: 'Comment on egusphere-2026-4100', Anonymous Referee #1, 11 Aug 2026
reply
Dear authors,
Please find my comments attached.
Best regards -
RC2: 'Comment on egusphere-2026-4100', Anonymous Referee #2, 12 Aug 2026
reply
The presented study by Thompson et al. entitled "Identifying regions of Europe which may be less prepared for extreme rainfall" investigates, how well different regions across Europe are prepared regarding extreme precipitation events. For this purpose, they apply a metric on two observation-based gridded precipitation data (E-OBS, HadUK-Grid), which combines the statistical approach of the risk of a new rainfall extreme (absolute magnitude of an extreme) with the probability of a new extreme to happen (time since the last extreme occurred). The latter also includes a more social science related aspect in the form of "the memory of people" on extremes. The further back in the past the extreme event took place, the less "memory" exist. To my opinion this is a highly relevant aspect when it comes to climate communication and acceptance of adaptation measures.
The study is well-structured and easy to follow. Data and methods are well described, and the results are presented in a straightforward and easy to follow way. It fells within the scope of NHESS.
However, I have some major concerns on some specific points, which to my opinion would improve the quality of the study significantly and better highlights the importance of this type of analysis, especially regarding applications, and which should be considered in a revised version. Please find attached a detailed list of comments and recommendations.
Kind regards
-
RC3: 'Comment on egusphere-2026-4100', Anonymous Referee #3, 31 Aug 2026
reply
The authors present an interesting study on the “assumed preparedness” of different geographic locations to extreme precipitation events based on the year of the record and the “risk”, defined based on the percentiles of precipitation at a given location. I very much like the concept of the study but am a bit sceptical of the applicability of the results given the assumptions of the risk metric, especially given potential climate change-induced increases in precipitation extremes. While the authors do discuss at length the limitations of this aspect in the final section I wonder why the authors did not use a more complex approach to assess the risk of an extreme at a given location. For example, using climate projection data and calculating the probability of exceeding an extreme threshold in the future climate.
Inline comments
L48-51: Which “people” are you referring to here? I don’t think most of the general public think in terms of precipitation amounts in millimetres or percentages. They may however wonder if a flood event with large damage/loss of life is possible in their city.
L50: You state that the rainfall was 30% of the yearly total but on L37 you state rainfall totals were greater than the region’s yearly total? I suggest rewording L48-51 for clarity.
L84: Please include in the caption how 0.2 can be interpreted on the y-axis.
L102-104: I suggest being more explicit here regarding why you also apply it to the second dataset.
L109-117: I’m not sure I fully understand the rationale behind Metric 2. If I understand correctly, if the extreme is far from the 99th percentile this location would be considered to be at a low risk, why?
L116-117: The climate change signal for precipitation is not spatially constant, i.e. increases in precipitation extremes are not expected to be homogenous in space across Europe.
L144: Figure 2c. Why does this contradict statements that some areas have exceeded their yearly annual rainfall in one day during extreme events? Is this a limitation of the dataset you are using?
L145-149: I suggest to explicitly mention “convective precipitation” rather than small localised events. Furthermore, even in western Europe you can see several “spots” presumably related to isolated convective precipitation events.
L170-171: I suggest adding 13% of the data period as with the “since 2000” statistic
L205: Figure 5. Maybe I missed it and it was mentioned already, but it may be worth highlighting the limitation of the “year of the current record” approach in the spatial sense. We see areas where “sitting duck” borders “recent rarity” however presumably people are informed of local precipitation events and are therefore more prepared.
L232: Figure 6. As the most recent high-profile event, I’d expect to see Valencia on this plot.
L256: “the recent event has reduced the chance of a new record” With the climate change signal such a statement is problematic due to potential shifts in the distribution of precipitation. While limitations are discussed in the following section, I suggest to already add more nuance here.
L257-258: Surface runoff and pooling at the bottom of higher terrain were relevant for this event, the authors could elaborate here.
L266: This statement seems to ignore any climate change signal.
Minor note: The dpi of the figures should be increased a bit.
Citation: https://doi.org/10.5194/egusphere-2026-4100-RC3
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 140 | 53 | 20 | 213 | 31 | 14 | 16 |
- HTML: 140
- PDF: 53
- XML: 20
- Total: 213
- Supplement: 31
- BibTeX: 14
- EndNote: 16
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This is an interesting analysis, but I see that the paper does not look at past studies concerning record-braking events in climate research (e.g. dating back to 2003). One central point is how often record-breaking events take place, and the year of the last record is strongly affected by randomness. I have provided an overview of some of the literature on record-braking statistics and analysis in https://doi.org/10.1126/sciadv.ado3712, which I recommend the authors to look at for completeness.