Dynamic CA-ffé: A hybrid 1D/2D fast flood evaluation model for urban flash floods

• Hybrid 1D/2D pluvial flood model integrates Cellular Automata and SWMM models. • Novel coupling of a 2D steady-state surface model and a 1D transient drainage system. • Achieves 20–60x faster simulations than hydrodynamic models with reasonable accuracy. • Validated on two Australian catchments against MIKE Flood and TUFLOW models. • Innovative C++ parallelisation boosts performance for large-scale catchments. This paper presents a novel hybrid 1D/2D urban flash flood model, named Dynamic CA-ffé, which combines a static DEM-based 2D overland flow model with a dynamic 1D drainage network model. The model is validated against two Australian catchments, Scotchmans Creek and Coogee Bay. Unlike traditional hydrodynamic models that solve the full Shallow Water Equations at a high computational cost, Dynamic CA-ffé integrates a 2D steady-state model with a 1D transient model, maintaining dynamic flow exchange while benefiting from the speed of conceptual flood models. This novel coupling approach effectively imposes the influence of the momentum equation on the 2D overland flow model via the 1D drainage network model, resulting in a quasi-dynamic flood simulation. Results show that Dynamic CA-ffé achieves high accuracy in predicting flood extent and depth compared to established hydrodynamic models such as MIKE Flood and TUFLOW, but with significantly reduced simulation times (at least 20 times faster). This balance of speed and precision makes it a valuable tool for urban flood risk management, particularly in time-sensitive decision-making scenarios. Furthermore, to improve the scalability of the model for the simulation of large catchments, an innovative parallelisation technique was also implemented.