Long-term Morphodynamic Modelling of Coastal Evolution of the German Bight

Abstract

The morphodynamics of coastal landscapes, such as embayments, estuaries, tidal channels and sandbars, have been systematically studied for several decades by field studies, empirical formulations and, increasingly, numerical models. Process-based morphodynamic models have been recognized as useful comprehensive tools for studying coastal evolution. However, in practice, one of the difficulties of studying coastal morphodynamics arises from the broad range of temporal and spatial scales involved. Over decadal temporal scales, morphodynamic models often show limitations in reproducing natural complex morphological evolution patterns, such as the movement of bars and tidal channels. Due to the known shortcomings of the numerical modelling system, data assimilation techniques, whereby model output is combined with measurements, have sprung up recently for application in coastal morphodynamic modelling. The main objective of this research is to investigate the coastal meso-scale morphological features of tidal channel migration and channel-shoal patterns by means of process-based morphodynamic modelling (using the Delft3D modelling suite) and data assimilation (DA) techniques. The study area is in the Jade and Elbe Estuary in the German Bight, North Sea. Decadal channel migration patterns are observed in the Jade Channel and the Medem Channel of the Elbe Estuary. The predictive ability of morphodynamic modelling of tidal channel migration is firstly examined. The improvement of current model predictions by DA techniques is subsequently evaluated. For the implementation of DA methods, assumptions and approximations have to be made (often based on experience) in order to define the observation and background error covariance metrics. A systematic analysis of the user defined correlation length scale for the definition of the background error covariance matrix has been conducted. The research starts with a simplified configuration of DA by neglecting the correlation length scale (set to 0). The DA scheme is then reduced to an optimization scheme which improved the model perditions, although to a limited extent. Furthermore, the correlation length scale is extended spatially and defined in both a homogeneous and heterogeneous way. This method is referred to as a nudging method, with which the model-predicted bathymetry is nudged towards predefined true states. This study has highlighted the definition of the correlation length scale with regard to morphological features, and an optimal value of the correlation length scale is suggested with respect to the grid cell size. In order to understand and interpret the tidal channel migration and associated channel-shoal patterns in estuaries, a schematized morphodynamic model has been applied to investigate the oscillation frequency of flood/ebb dominance which controls the net sediment transport and long-term estuarine morphologies. The oscillation frequency of flood/ebb dominance is formulated and decomposed under three time scales. The long-term annual river discharge is found to be of importance for the oscillation of flood/ebb dominance. Sea level rise increases the depth and enhance the flood dominance and increase estuarine infill.