Two-dimensional and three-dimensional computational models in hydrodynamic and morphodynamic reconstructions of a river bend: sensitivity and functionality

This study assesses hydrodynamic and morphodynamic model sensitivity and functionality in a curved channel. The sensitivity of a depth-averaged model to user-defined parameters (grain size, roughness, transverse bed slope effect, transport relations and secondary flow) is tested. According to the sensitivity analysis, grain size, transverse bed slope effect and sediment transport relations are critical to simulated meander bend morphodynamics. The parametrization of grain size has the most remarkable effect: field-based grain size parametrization is necessary in a successful morphodynamic reconstruction of a meander bend. The roughness parametrization method affects the distribution of flow velocities and therefore also morphodynamics. The combined effect of various parameters needs further research. Two-dimensional (2D) and three-dimensional (3D) reconstructions of a natural meander bend during a flood event are assessed against field measurements of acoustic Doppler current profiler and multi-temporal mobile laser scanning data. The depth-averaged velocities are simulated satisfactorily (differences from acoustic Doppler current profiler velocities 5–14%) in both 2D and 3D simulations, but the advantage of the 3D hydrodynamic model is unquestionable because of its ability to model vertical and near-bed flows. The measured and modelled near-bed flow, however, differed notably from each other's, the reason of which was left open for future research. It was challenging to model flow direction beyond the apex. The 3D flow features, which also affected the distribution of the bed shear stress, seem not to have much effect on the predicted morphodynamics: the 2D and 3D morphodynamic reconstructions over the point bar resembled each other closely. Although common features between the modelled and measured morphological changes were also found, some specific changes that occurred were not evident in the simulation results. Our results show that short-term, sub-bend scale morphodynamic processes of a natural meander bend are challenging to model, which implies that they are affected by factors that have been neglected in the simulations. The modelling of short-term morphodynamics in natural curved channel is a challenge that requires further study.