Floods we could have faced: exploring exceptional flooding using perfect storm concept

River floods that exceed historical records often come as a surprise, causing widespread damage and disruption. To enhance disaster preparedness, it is essential to estimate exceptional flood scenarios that surpass past observations. There exist a number of methods such as stochastic storm transpositions, storylines, and downward counterfactuals to explore the space of extreme floods. The perfect storm concept uses unusual combinations of causative factors to generate extreme but plausible flood scenarios. We construct synthetic floods by mixing past severe rainfall events with observed antecedent catchment states from other floods or extreme catchment states but without flood occurrence. In this study, we apply this concept to develop exceptional flood scenarios by using the meso-scale hydrological model mHM driven by 70 years of meteorological data across Germany.

Our findings indicate that plausible perfect storm scenarios, respecting flood seasonality, can produce exceptional floods exceeding those observed in the past decades. Shifting rainfall to wetter soil conditions amplifies flood severity significantly, with some cases experiencing flooding up to seven times more severe compared to the original events. Even minor temporal shifts in rainfall, such as one month earlier or later, can drastically increase flood magnitudes, highlighting the significant impact of the temporal alignment of catchment state and rainfall events on flood severity. The perfect storm approach provides a practical means for identifying and communicating plausible and intuitive extreme scenarios with low probability. By integrating this method into flood risk management, planners and policymakers can better anticipate and prepare for the impacts of unprecedented flood events, reducing negative surprises.