摘要：After three hours of intense rainfall, the city of Nice was flash flooded on October 3, 2015, resulting in casualties and severe damages in property. This study presents a porous shallow water-model based numerical simulation of the flash flood event in a district of Nice, and compares the results with a high-resolution conventional shallow water model. This contribution aims to discuss practical aspects of applying a porous shallow water model to a real world case. The porous shallow water model is an integral porosity-type shallow water model. It uses unstructured triangular meshes. The conventional shallow water model is a distributed memory parallelized high-performance computing code, that uses a uniform Cartesian grid. The study site is an approximately 5 km2 spanning district of the city of Nice, France. Topography information is available in a 1m resolution and in addition, the available digital elevation model includes inframetric structures such as walls and small bridges. In the presentation of the case study, challenges of the pre-processing step of the integral porosity shallow water model are addressed. Notably, a method to semi-automatically generate “good” triangular meshes using the open-source geoinformation system QGIS and the mesh generator Gmsh is presented. During the post-processing step, the results of the porous model are mapped back onto the high-resolution topography to make the results more meaningful. The agreement between the high-resolution reference solution and the porous model results are poor. A speed up of about 10 to 15 was observed for the present case.
其他摘要：After three hours of intense rainfall, the city of Nice was flash flooded on October 3, 2015, resulting in casualties and severe damages in property. This study presents a porous shallow water-model based numerical simulation of the flash flood event in a district of Nice, and compares the results with a high-resolution conventional shallow water model. This contribution aims to discuss practical aspects of applying a porous shallow water model to a real world case. The porous shallow water model is an integral porosity-type shallow water model. It uses unstructured triangular meshes. The conventional shallow water model is a distributed memory parallelized high-performance computing code, that uses a uniform Cartesian grid. The study site is an approximately 5 km2 spanning district of the city of Nice, France. Topography information is available in a 1m resolution and in addition, the available digital elevation model includes inframetric structures such as walls and small bridges. In the presentation of the case study, challenges of the pre-processing step of the integral porosity shallow water model are addressed. Notably, a method to semi-automatically generate “good” triangular meshes using the open-source geoinformation system QGIS and the mesh generator Gmsh is presented. During the post-processing step, the results of the porous model are mapped back onto the high-resolution topography to make the results more meaningful. The agreement between the high-resolution reference solution and the porous model results are poor. A speed up of about 10 to 15 was observed for the present case.
