Enhancing European windstorm return period estimates for (re)insurance

Assessing windstorm hazard return periods is crucial for the (re)insurance industry due to the large losses these events can cause. Accurately estimating return periods for specific wind gusts is essential. Traditionally numerical model simulations over multiple years of windstorm events are used for this purpose.  However these models may contain biases, such as over-calibration to certain periods (e.g. the 1990s) or major loss events (e.g. Daria and  Lothar).  Return periods from a numerical model are compared to an existing statistical model and differences explored. From these differences, it is possible to adjust the numerical simulation model output to match the known statistical distribution more closely.  The adjustment method adheres to the yearly structure of the numerical simulation model output. It is shown to provide a suitable adjustment for a variety of locations, providing a good use case for the (re)insurance industry. The method is flexible, allowing for more simulated years than the numerical model’s output.  This method is applicable to most locations within the European domain, particularly in areas more exposed to extratropical cyclones.