Conference Agenda

Traditional hydrodynamic flood models face significant limitations in real-time forecasting applications due to their computational complexity. High-resolution hydrodynamic flood simulations require extensive calculation times, often exceeding the skilful lead times of high-resolution short-term rainfall forecasts. This temporal mismatch proves particularly challenging for urban catchments due to their rapid response dynamics, resulting in delayed forecasts unsuitable for support of operational decision-making and impact mitigation. This study investigates a hybrid framework combining simplified hydrodynamic physics with machine learning to achieve accurate and timely flood depth mapping. While simplified models significantly accelerate computations by abstracting surface and sewer components, they sacrifice precision compared to the reference hydrodynamic simulations. To bridge this gap, a Gaussian Process (GP) regression model is trained on a comprehensive library of simulation results, enabling efficient bias correction without sacrificing computational speed. Implemented for a case study in Antwerp, Belgium, using radar-derived rainfall data, the framework demonstrated speed improvements with a factor of 30-100 depending on the simplified model configuration, while maintaining good accuracy (R² ≈ 0.85). Ongoing work expands the analysis to diverse storm events and model variants.

Extreme rainfall storms can generate such extreme flash floods that the so-called wall of water phenomenon occurs, which is an immensely fast rise of a wavefront through a street canyon or a natural water course often causing particularly devastating flood damages. This study presents a categorisation method to identify reasonable worst-case rainstorm events from station data that result in catastrophic flash flood events. The identified 24 events were shown to be considerably more extreme than typically employed design storms with a return interval of 100 years. In the selection of the extremes, special care was taken to consider different hazard attributes, which are associated with intrinsically different observed hyetograph shapes, including rectangular, short and long triangle, and double peak shape categories. Hydrodynamic simulations were carried out using the 2D robust shallow water model hms⁺⁺ to quantify the flooding impact of the 24 rainfall events in a small urbanised catchment. The simulated hazard attributes showed clear differences as a function of hyetograph shape with strong variations in peak discharges, peak water depths, hazard velocities and flashiness (shortest time to crest). The use of real heavy rainfall series is therefore essential for the generation and further analysis of wall of water events.

Multidisciplinary stormwater management is a still a main challenge of sustainable stormwater and urban pluvial floods (UPF) management. To assess an urban catchment a typical approach of coupling hydrodynamic models of above and underground systems is used. However, this is technically challenging, plus it requires significant computational and data resources. Using GAMA, a platform for developing Agent-based models, an interactive and friendly-user tool is presented in which a hydrodynamic model of sewer system in SWMM is coupled with simplified overland inundation models. The user only has to select the main parameters, like duration and return period, and run the simulation to see the inundation and a damage estimation. The hydrodynamic simulation and coupling happens in the background. The tool can be used for knowledge transfer in workshops or academia with experts, non-experts, decision makers, among others.

Heavy precipitation events frequently cause flooding, leading to significant damage in affected areas. Simulation models play a crucial role in improving heavy rain management, particularly as climate change is expected to increase the frequency of extreme precipitation events. However, conventional hydrodynamic models are computationally intensive and unsuitable for real-time applications. This study explores the use of the fast flood modelling approach "Dynamic CA-ffé" for pluvial flood management in Austria. Previously tested in two smaller case studies in Australia with promising results, this model is now applied to four Austrian case studies, including larger (up to 50 km²) and partially alpine regions. The adaptation of the model to these new contexts highlights challenges such as increased inflow from alpine surrounding areas and the integration of multiple rain gauges. To address these challenges, a simplified runoff coefficient based on land use categories, adaptation of the digital terrain model using a buildings layer and the use of raster data for precipitation were introduced. The findings demonstrate the feasibility of Dynamic CA-ffé for larger-scale and alpine applications, offering a promising tool for efficient and fast flood simulation in pluvial flood management.