As part of the funded joint project "Model-based scenario analysis for optimizing crop production for climate change mitigation" (ModOKlim), subproject AP1 is concerned with the assessment of hail hazard in Germany (with a focus on agriculture) in the past and in the future. Together with drought and frost, hail is one of the main causes of damage to agricultural crops. In general, the risk of hail is expected to increase in the future. The main reason for this is the increase in humidity in the lower troposphere associated with the temperature increase, which favors the development of thunderstorms and thus hail formation. However, direct measurements of thunderstorm activity (lightning detection networks or radar-based analyses) do not show a clear increase in (severe) thunderstorm activity in Germany. The reasons for this discrepancy are not yet understood and will be investigated in this project.

The investigations are based on radar data of severe thunderstorm events with high hail potential. These events are classified into different intensity classes using agricultural damage reports (e.g. insurance data) and other appropriate meteorological datasets (e.g. hail size reports). Subsequently, machine learning methods are used to determine suitable combinations of thermodynamic and dynamical parameters (so-called proxies) from reanalysis data that best represent the detected hailstorms of different intensity classes. These methods are then applied to CMIP6 climate data to analyze the frequency and intensity of hail probabilities in different climate scenarios. This will make it possible to estimate the expected changes in hail hazard to agricultural crops in the future.