Han et al. (2025) Recombining past event precipitation and antecedent catchment states generates unprecedented floods

Identification

• Journal: Communications Earth & Environment
• Year: 2025
• Date: 2025-08-22
• Authors: Li Han, Bruno Merz, Nguyễn Viết Dũng, Björn Guse, Luis Samaniego, Kai Schröter, Sergiy Vorogushyn
• DOI: 10.1038/s43247-025-02691-6

Research Groups

• GFZ Helmholtz Centre for Geosciences, Section Hydrology, Potsdam, Germany
• University of Potsdam, Institute for Environmental Sciences and Geography, Potsdam, Germany
• Christian-Albrechts-University of Kiel, Institute of Natural Resource Conservation, Department of Hydrology and Water Resources Management, Kiel, Germany
• Helmholtz Centre for Environmental Research GmbH—UFZ, Department Computational Hydrosystems, Leipzig, Germany
• Technische Universität Braunschweig, Leichtweiß-Institute for Hydraulic Engineering and Water Resources, Division Hydrology and River Basin Management, Braunschweig, Germany

Short Summary

This study introduces a "perfect storm" approach to generate plausible, unprecedented flood scenarios by recombining historical extreme precipitation events with antecedent catchment soil moisture conditions in Germany, demonstrating that these scenarios can significantly exceed historical flood magnitudes and damages.

Objective

• To introduce a "perfect storm" approach that recombines historically observed extremes in precipitation and soil moisture to generate plausible flood scenarios that surpass past observations and are easy to understand for flood risk managers and lay people, thereby improving disaster preparedness.

Study Configuration

• Spatial Scale: Germany and headwater regions in neighboring countries (Rhine, Danube, Elbe, Weser, Ems, Oder, Meuse river basins). Hydrological model resolution: 5 km grid cells (41,800 cells). Streamflow gauges: 516.
• Temporal Scale: 1950–2021 (70 years of meteorological data, annual maximum streamflow data). Event precipitation replacement: 21 days. Soil moisture conditions: 1 day prior to event.

Methodology and Data

• Models used:
  • Mesoscale hydrologic model (mHM)
  • Regional Flood Model (RFM) for damage estimation
  • Generalized Extreme Value (GEV) distribution for flood frequency analysis
• Data sources:
  • E-OBS precipitation and temperature data (daily, 0.25 degree spatial resolution, 1950–2021)
  • HANZE database (for historical flood damages, 1950-2021)
  • Streamflow data from various national and regional authorities (516 gauges, 1950-2021)
  • Land use layers (1950–1999, 2000–2005, 2006–2011, 2012–2021)

Main Results

• The perfect storm approach generates unprecedented floods exceeding historical records at over 75% (403 out of 516) of observed gauges in Germany.
• At 370 gauges (60%), recombined floods exceed the 100-year return period, with the highest numbers in the Rhine (40%) and Elbe (30%) basins.
• Some recombined scenarios result in flood peaks with return periods exceeding 10,000 years (e.g., Hattingen, Kaub, Voltho, Neu Darchau gauges).
• Recombining May 1978 or May 2013 precipitation with April 1988 antecedent soil moisture at Kaub gauge resulted in flood return periods 600 and 400 times greater than historical events, respectively.
• Shifting extreme rainfall to wetter soil consistently amplifies flood severity.
• Moving events by just 1 month can also intensify flooding, highlighting the critical role of temporal alignment.
• The most severe perfect storm scenario was the recombination of August 2002 rainfall with April 1994 soil moisture, resulting in a flood severity index (SI_Q) of 1.97, exceeding original event severities.
• Rainfall events that historically caused no severe flooding (e.g., October 1986) could cause 1000-year return period floods under different, wetter soil moisture conditions (e.g., February 1980, April 1988, April 1994).
• Simulated flood damages for perfect storm scenarios are substantially higher than historical values, with the median damage across severest scenarios nearly 16 times greater than historical simulations.
• The perfect storm approach generally produces a wider range of flood severities compared to spatial counterfactual approaches, suggesting temporal shifts have a greater influence.

Contributions

• Introduces a novel and easy-to-communicate "perfect storm" approach for generating plausible, unprecedented flood scenarios by recombining historical precipitation and antecedent soil moisture.
• Demonstrates that such recombined scenarios can systematically exceed historical flood magnitudes and associated economic damages, providing critical insights for disaster preparedness.
• Highlights the crucial role of temporal alignment between precipitation and antecedent catchment conditions in modulating flood severity.
• Offers a practical tool for stress-testing flood protection and disaster management strategies beyond historical observations.
• Complements existing probabilistic and spatial counterfactual methods for comprehensive flood risk assessment.

Funding

• Federal Ministry of Education and Research of Germany (BMBF) through the ClimXtreme project FLOOD (Grant No. 01LP1903E)
• Federal Ministry of Education and Research of Germany (BMBF) through the KAHR project (Grant No. 01LR2102A)
• JCAR-ATRACE project (https://www.jcar-atrace.eu/)

Citation

@article{Han2025Recombining,
  author = {Han, Li and Merz, Bruno and Dũng, Nguyễn Viết and Guse, Björn and Samaniego, Luis and Schröter, Kai and Vorogushyn, Sergiy},
  title = {Recombining past event precipitation and antecedent catchment states generates unprecedented floods},
  journal = {Communications Earth & Environment},
  year = {2025},
  doi = {10.1038/s43247-025-02691-6},
  url = {https://doi.org/10.1038/s43247-025-02691-6}
}