Europe has experienced a sharp rise in catastrophic flash-flood losses, highlighted by record-breaking downpours such as those in Valencia (2024) and Western Europe (2021), when the amounts of rainfall experienced were higher than anticipated. This raises a key question: Where in Europe are similar events most likely to occur?

Applying extreme value theory (EVT), we analyze rainfall data over a wide range of European weather stations to estimate precipitation levels associated with various return periods. We use an automated method to estimate extreme return levels at scale, supplemented by manual methods to improve fits where necessary.

This white paper presents our findings and demonstrates how insurers and governmental agencies can harness EVT to structure parametric insurance covers, sharpen solvency assessments, and guide resilient infrastructure investment.

Discussion points:

• Three recent floods: Using EVT to gauge the extent to which rainfall in these events could have been expected
• Introduction to EVT: Block maxima and peak-over-threshold methods
• Application of EVT to precipitation data: Methodology, implementation and results
• Practical use cases: Parametric insurance, regulatory frameworks, and government planning and resilience