Initial stages in the development of a coupled hillslope hydrology — floodplain inundation model

Abstract

Two-dimensional hydrodynamic floodplain inundation models were originally developed for civil engineering applications and have been developed to a high level of sophistication. These two-dimensional (depth averaged) schemes are capable of a high degree of spatial representation and recent developments such as their application to longer reach lengths have enabled their application in other fields including hydrology, geomorphology and ecology. These models provide a powerful tool for investigations in these fields and may be considered as a platform for further developments which incorporate specific processes occurring within the floodplain environment. It is for the case of hydrological applications that a problem has been identified in that these models do not currently consider catchment hydrology; whilst the models provide a good representation of floodplain processes in a hydraulic context, catchment hydrology is essentially treated as a black box. The only input to the system is the upstream input hydrograph (occasionally rainfall over the floodplain surface or tributary inflows are included) and output only occurs at the downstream boundary. The floodplain is assumed to be impermeable and any input from the hillslopes bordering the reach is ignored.This study discusses the initial stages of an investigation carried out to examine the significance., of contributions to the floodplain from the hillslopes bordering the reach. In order to do this, the zero flux boundary condition at the hillslope-floodplain interface is relaxed. A two-dimensional floodplain inundation model, RMA-2, is set up for a 14 km reach of the River Culm in Devon. A distributed hillslope hydrology model, VSAS3 is set up for a section of the hillslopes bordering the reach. The topography of the hillslope section is used as a geometric template, although a semi-theoretical approach is adopted whereby it is assumed that the hillslope hydrological characteristics are homogenous and VSAS3 is parameterised using values reported in the literature from field observations. VSAS3 is coupled to RMA-2 using a simple external coupling mechanism where the hillslope discharge predicted by VSAS3 is applied to elements along the edge of the RMA-2 finite element mesh. Three storm events, with different return periods are simulated using this coupled scheme. It has been shown that hillslope inflows can have a significant effect on the predictions made by RMA-2. In addition, the timing of the hillslope inflow peak relative to the arrival of the floodwave from upstream is of great importance.