Motivation:

Severe convective storms (SCS) and large hail in particular are a major loss driver worldwide. These extremes repeatedly cause considerable damage to buildings, vehicles, critical infrastructure and agriculture. Anthropogenic warming increases atmospheric moisture and thus instability, but also affects other meteorological factors that are important for the formation of hailstorms, such as vertical wind shear or melting level height. In addition, it can be assumed that the frequency of specific large-scale weather conditions or regimes, which provide the general setting of the synoptic fields for convection, will also change. In view of the large damage caused by hail, an improved understanding of hail risk, the spatial and temporal distribution of hail events and their intensity is of utmost importance for several users, such as the global (re)insurance industry.

Project activities:

The aim of the research project “Managing European Hail Risk under Climate Change (HAR-CC)”, which is performed in cooperation with Allianz Re, Munich, is to describe the effects of climate change on hail hazard and hail risk for parts of Europe. In this context, an improved understanding of the effects of climate change on hail hazard / hail risk both over past and future decades (2030 to 2050) are of interest. Therefore, high-resolution climatological analyses are prepared for different time horizons and different meteorological regions.

By applying an object-oriented approach based on specific characteristics of the events (intensity, organization, structure, spatial/temporal extent), the relationship between SCS/hailstorms and environmental conditions will be refined in order to better understand and reconcile the different trends observed in the various datasets over the last decades. This relationship between objects and ambient fields for different intensity classes and thunderstorm types will then be transferred to an ensemble of high-resolution regional climate models for different emission scenarios (SSP). Various statistical methods including machine learning methods will be applied for the relationship between object properties and model fields.

In addition, a spatially distributed, event-based risk assessment, e.g. through a catalog of stochastically generated events, is an important tool for understanding the different impacts of events with high frequency and low intensity compared to rarer, more extreme events. This will also be developed as part of this project.