The ClimXtreme project “FLOOD – Investigation of the atmospheric triggers of large‑scale flood events and their impacts in Central Europe” was carried out from 1 March 2020 to 28 February 2023 under the German Federal Ministry of Education and Research (grant 01LP1903E). The work was led by Prof. Dr. Bruno Merz, Dr. Sergiy Vorogushyn and Dr. Dung Viet Nguyen from the Helmholtz‑Zentrum Potsdam and the Deutsches GeoForschungsZentrum Potsdam, Sektion Hydrologie. The project is part of Module C Impacts of the ClimXtreme consortium, which brings together several research groups to assess climate‑driven hazards in Europe.
The scientific aim was to quantify how future changes in atmospheric circulation and precipitation will alter the frequency and magnitude of extreme river floods in Germany. To this end the team expanded and validated a comprehensive flood‑model chain that links atmospheric forcing, basin‑scale hydrology, hydraulic routing and damage estimation. The model chain was calibrated against observed flood events and long‑term meteorological records, achieving high fidelity in reproducing peak discharges and flood extents across a range of German catchments.
A key technical development was the implementation of a non‑stationary weather generator that reproduces the statistical properties of precipitation and temperature while preserving the influence of large‑scale circulation patterns. The generator was trained on reanalysis data and validated against observed daily precipitation, showing strong agreement in mean, variance and extreme quantiles. It was then used to produce synthetic climate series that reflect projected changes in atmospheric dynamics and thermodynamics derived from a selection of CMIP6 climate model projections. The climate projections were carefully screened, bias‑corrected and downscaled to a 5 km grid covering Germany, providing the necessary input for the hydrological component.
Analysis of the projected atmospheric variables revealed systematic increases in temperature and specific humidity, leading to higher precipitation intensities. The non‑stationary weather generator was employed to simulate extreme precipitation events under present‑day and future climate scenarios. The simulations indicate a marked rise in the frequency of events exceeding the 90th percentile, with some regions experiencing up to a 25 % increase in the probability of such extremes.
Using the validated flood‑model chain driven by the synthetic precipitation series, the team simulated flood frequency changes for a suite of German catchments. Results show a substantial increase in the probability of 100‑year floods, with magnitudes rising by up to 30 % in the most climate‑sensitive basins under the RCP8.5 scenario. The study also quantified the associated economic damage, highlighting a projected rise in flood‑related losses across Germany.
The project’s findings have been disseminated through several peer‑reviewed publications, including contributions to the ClimXtreme LASLI project and a paper on rockfall triggers in Germany. Ongoing work focuses on publishing a detailed description of the weather generator and on a comprehensive assessment of future flood risk in Germany. The collaboration within the ClimXtreme consortium has enabled the integration of atmospheric, hydrological and socio‑economic expertise, ensuring that the results are robust and directly applicable to policy and risk management in Central Europe.