Holger Schüttrumpf

Increasing extreme weather events pose significant challenges in hydrology, requiring tools for preparedness and prediction of intense rainfall impacts, especially flash floods. Current risk reduction measures for pluvial flood risk management rely on flood hazard maps, but inconsistencies in transregional standards that are used for risk assessment hinder cross-regional comparisons. While there are existing guidelines for the development of pluvial flood hazard maps, there is still a lack of holistic modelling systems that enable harmonised predictions of the impacts of heavy rainfall events. Furthermore, sensitive city data (e.g., critical infrastructure, sewer network) exist in many municipalities, which cannot be readily disclosed for modelling purposes. In this work, we propose an approach using distributed analytics to distribute computation commands to existing hydrodynamic models at different locations. In combination with harmonising model adapters, we enable the generation of harmonised pluvial flood hazard maps of different regions to tackle the inconsistencies and privacy concerns. We apply our approach to four adjacent urban areas in the Rhein-Sieg Kreis of North Rhine-Westphalia. Our results demonstrate the ability of our approach to produce cross-regional pluvial flood hazard maps, supporting disaster preparedness and management in regions prone to extreme weather events and flash floods.