Sediment Transport Modelling for Coastal Morphodynamics

The accurate prediction of sand transport rates in the coastal zone depends not only upon the choice of sediment transport model but also, rather critically, upon the formulation used to predict the bed roughness (ks). Sand transport rates are modelled here using both a One-equation, turbulence-closure, TKE-model and also Bijker’s sand transport model. Following some remarks about the nature and importance of sand ripples, results corresponding to wide ranges of wave and current conditions are obtained based upon both prescribed and predicted values for ks. The substantial difference between the respective results highlights the importance of predicting ks as accurately as possible in situations in which the bed roughness is not known. An example involving waves incident on a plane sloping beach is used to illustrate the variations in ripple dimensions, and hence ks, likely to be found in a typical coastal domain. Here Bijker’s model has been implemented within the TELEMAC Modelling System to predict the longshore sand transport rate. The inclusion of local variations in ks within the study area approximately halves the longshore transport rate compared with results based on the use of a constant, prescribed value of ks throughout the domain. The possible consequences of bed roughness variations are discussed briefly in the context of morphological modelling. INTRODUCTION The accurate quantification of local sand transport rates in the marine coastal environment is a prerequisite for the prediction of seabed changes and coastline evolution. In an intercomparison with laboratory data in the EU MASTII G8-M Project (1992-5), it was found that four different ‘research’ models predicted net sediment transport rates beneath asymmetrical waves, and also in collinearly combined wave and current flows, to well within a factor of 2 of the measured values (Davies et al., 1997). In addition, cycle-mean sediment concentration profiles were predicted to good accuracy. However, the range of conditions investigated was quite narrow, involving sheet flow (i.e. plane beds) only. A more recent series of model intercomparisons, and model comparisons with field data, carried out over a wider range of conditions during the EU MASTIII SEDMOC Project (1998-2001), suggested far less convincing agreement between the models and the data, and indicated that large gaps still remain in our knowledge of sand transport processes (Davies et al., 2002). Initially, in the SEDMOC study, seven ‘research’ models were intercompared over a wide range of wave and current conditions, corresponding to both plane and rippled sand beds. These models included both one-dimensional vertical (1DV) formulations, varying in complexity from eddy viscosity and mixing length models to a full two-phase flow formulation, and also 2DHV formulations capable of representing vortex shedding above sand 1) School of Ocean Sciences, University of Wales (Bangor), Menai Bridge, Anglesey LL59 5AB, UK. [email protected] 2) Laboratoire National d'Hydraulique et Environment, Electricité de France, 6 quai watier, BP 49, 78400 Chatou, France. [email protected].