The Potential of Seasonal Forecasting of Flood Frequency for the Insurance Sector

This Joint Research Project between Prof. Dr. Bruno Merz of the GFZ German Research Centre for Geosciences and AXA Group Risk Management aimed to explore the feasibility of providing seasonal forecasts of flood frequency for Europe that are relevant for the insurance industry. The main focus was to investigate the connection between catchment state, climate state, and flood frequency in European regions. The project aimed to determine the extent to which seasonal forecasts of flood frequency are viable from an insurance perspective and to identify the role of catchment state and climate state in forecasting flood frequency. Prof. Dr. Merz was analyzing catchment state and climate state indicators to understand their impact on the distribution of maximum seasonal streamflow and their ability to predict flood frequency for the months ahead. Preliminary conclusions suggested that catchment state indicators, which act as proxies for catchment wetness, may be more effective predictors of flood generation across Europe than climate state indicators. 

In 2019, the project involved investigating the variability of catchment or climate state and its effect on the distribution of maximum seasonal streamflow at around 650 stations in Europe. The study found that catchment state indicators appeared to be more effective predictors than climate state indicators, with antecedent precipitation and antecedent streamflow showing strong influence for up to two months ahead of the forecast season. The results also indicated that the immediate preceding season/month provides the best results for flood probability forecasting, with some skill persisting for higher time lags, and further analysis is needed to understand the increased skill for higher time lags.