Morphodynamic Modelling With Uncertain Geometry Input

For morphodynamic modelling, riverbed survey data are essential as the basis for the evaluation of temporal riverbed development, mesh creation, and model calibration. To study the effects of uncertain geometry input on these issues, datasets of different spatial resolutions were analysed. As a result, cross-profile data were derived from high-resolution survey data, which are available for a river reach in the Upper Danube in Bavaria for several periods. Finally, the prediction quality of simulations based on cross-profile and high-resolution spatial data was assessed. The analysis of both datasets shows continuous riverbed erosion but of different magnitudes. However, complex riverbed geometry due to, e.g., scours, is depicted poorly by cross-profile data. In more homogenously characterised reaches, cross-profile data significantly more closely represents the riverbed geometry than the high-resolution spatial data base. Local misinterpretation of riverbed geometry by cross-profile data leads to deviations of calibration parameters in the entire study area. Consequently, these deviations in calibration outcome effect the model predictions. In this case, cross-profile calibration generally induces higher transport capacities, leading to more erosion in the study area compared to the model based on high-resolution spatial calibration. The general shape of predicted riverbed geometries is found to be similar but with local deviations, which are not limited to areas with complex river geometry.