Climate change amplifies flooding more than extreme precipitation

To investigate this, the researchers combined climate simulations with two different hydrological models. The AWI storyline simulations by Helge Goessling, Antonio Sánchez-Benítez and Marylou Athanase provided data for the current study. The simulations reconstructed the observed event in a physical climate model under present-day and pre-industrial conditions, and they served as input for the hydrological models now used. The goal was to trace how rainfall in a catchment generates runoff and ultimately leads to flooding, and how this depends on the climate state. The differences were substantial. Without anthropogenic climate change, precipitation during Storm Boris would have been around 6–9% lower depending on the climate model, but flood peaks would have been 14–22% lower on average.

The reason for this disproportionate increase lies in how catchments respond to additional rainfall. “The soil can initially absorb some of the rain during a heavy rainfall event. But eventually the soil becomes saturated, and every additional millimetre of rain flows directly into the river”, explains Miersch. Once the soil is saturated, additional rainfall leads to a particularly sharp increase in runoff.

To assess the robustness of the results, the researchers combined two different climate models with two hydrological models, one of which uses machine learning. Despite the different methodological approaches, the same basic pattern emerged in all combinations: flood peaks increased considerably more than precipitation.

This effect is particularly relevant where flood protection measures are designed for specific flood thresholds. “If climate change increases extreme precipitation by 10%, it does not mean flooding will increase by only 10%. It could also be 20%. If flood levels exceed the thresholds for which flood protection measures are designed, extensive damage can occur very quickly”, says co-author Prof. Dr Jakob Zscheischler, a climate researcher at the UFZ and Professor of Data Analytics in Hydro Sciences at TUD Dresden University of Technology. Assessments of flood risk and flood protection that consider only changes in extreme precipitation may therefore lead to an underestimation of the increase in flood hazard.

However, the researchers deliberately limited the scope of the study to isolate how much anthropogenic climate change altered a specific event under otherwise identical conditions. The study does not establish whether climate change makes weather patterns similar to those during Storm Boris more or less likely. Changes in land use or water management were also not taken into account. Put simply, the researchers compared the same present-day landscape under different climatic conditions.

The results therefore show that changes in flood hazards associated with climate change cannot be assessed solely based on changes in precipitation. When assessing future risks and developing adaptation strategies, it is therefore important to consider the entire process from the weather event to river discharge.

The current work extends the storyline approach from the atmosphere and ocean to the rivers and hydrology. It shows that rainfall from climate models alone is not enough to assess how climate change affects flood impacts. “The study is also one of only a few to combine at least two different nudged climate models”, says AWI co-author Marylou Athanase. “Through this work, AWI contributes to an emerging intercomparison of such nudged simulations. The project is an example of successful collaboration between Helmholtz Centres, combining our AWI expertise in climate modelling with UFZ’s expertise in hydrology.”