Author(s): Ilhan ÖZgen
Linked Author(s): Ilhan Özgen-Xian
Keywords: Shallow water model, porosity, coarse grid approach, urban catchment, friction law.
Abstract: Flood events in urbanized areas are usually caused by localized rainfall events with very high intensity that occur in catchments located at the upstream of the city. The typical model chain for forecasting this type of event is a hydrological model that generates input for a two-dimensional hydraulic model. However, in recent years, the depth-averaged two-dimensional shallow water equations are applied to compute rainfall-runoff in natural catchments as well as inundation areas in city environment. The application is limited by computational constraints, which resulted from the high mesh resolution required to account for microtopography in natural and buildings in urban catchments. In this context, coarse grid approaches aim to reduce computational cost by enabling simulations on coarser meshes and introducing subgrid treatments to recover some of the information at subgrid-scale. This contribution presents a novel model chain that comprises two coarse grid approaches with specialized application domains: (1) friction law-based coarse grid approach and (2) anisotropic porosity-based coarse grid approach. The hydrological model that is usually used for these type of predictions was replaced by a shallow water model with a specialized friction law to account for microtopography. The urban flood inundation model was sped up by introducing anisotropic porosity terms to account for buildings. The model chain was applied to predict rainfall-runoff in an idealized city, based on a real rainfall event in a real natural catchment. An averaged behavior of a high-resolution model chain can be obtained with significantly lower computational cost such that the simulations run on average about 100 times faster than the high-resolution counterpart
Year: 2017