Analysis and assessment of uncertainties in coastal morphodynamic modelling

During the last decades numerical models have become a powerful tool to investigate hydrodynamics and sediment dynamics. However, many key parameters and decisive numerical formulations of morphodynamic processes are bound to assumptions, estimations and uncertainties, which proceed to model results. The study aims to contribute to identify and take uncertainty of results in morphodynamic simulations into account. This is done by means of statistical analysis of key model parameters, and by the application of the ensemble forecasting method, based on Monte Carlo Simulations. The procedure is developed and applied to simulate the morphodynamic processes in the tidal flats of the Dithmarschen Bight, North Sea, Germany. Focus is placed on the migration of a tidal channel system, a physical process still poorly clarified, over observation periods of two years. Based on available data and extensive measurements of bathymetric evolution, water level, current, wave and sediment concentration, process-based models for flow, waves, sediment transport and morphodynamics are developed. Comparisons of bathymetric data with model results show that the model is capable to meet well the direction of the channel migration while the extent of migration is underestimated. Nevertheless, morphodynamic processes in tidal channels have been better resolved than those on tidal flats. Ranges of uncertainties are revealed. Analysis of model confidence showed that storm events have a significant relevance on morphodynamic results. Overall, the ensemble forecasting method proves to be a useful tool to account for uncertainties in morphodynamic modelling, to provide ranges of probability and, thus, increase the reliability of morphodynamic modelling results.