Author(s): Qi Xie
Linked Author(s):
Keywords: Distributed Direct Rainfall Model; Flood Risk Study; Real-time Flood Prediction; GPU; TUFLOW
Abstract: Fast and accurate flood modeling is critical for any early flood warning (EFW) system as well as for flood risk study. So far flood management managers and flood modelers are not satisfied with traditional flood prediction methods that use empirical correlation curves or a lumped hydrologic model plus a separate 1D or 1D-2D river hydrodynamic model. With the rapid advance in GPU technology and innovation development of flood modeling software technology in recent years, Distributed Direct Rainfall Model (DDRM) technology has been showing significant advantage over traditional flood modeling methods and is used worldwide in numerous flood risk study projects and real-time flood prediction projects. The current paper introduces recent advance in DDRM technology inhered in flood modeling software TUFLOW HPC and shows a few application examples of DDRM in Flood Risk Study and Real-time Flood prediction. Key technologies such as GPU parallel computation, Quadtree mesh refinement, Sub-Grid-Sampling of DEM (Digital Elevation Model) data, Mesh Size Insensitive Turbulence Model and Sub-Surface flow simulation, etc. are briefly introduced. Data requirement and availability, computer hardware configuration, model simulation time, accuracy of model results, etc. are discussed.
Year: 2024