Résumé
Hydromorphodynamic models are powerful tools for predicting the potential mobilization and transport of sediment in river ecosystems. Recent studies have shown that they are able to predict suspended sediment matter concentration in small river systems satisfactorily. However, hydro-sedimentary modelling exercises often neglect suspended sediment properties (e.g. sediment densities and grain-size distribution), which are known to directly control sediment dynamics in the water column during flood events. The main objective of this study is to assess whether a better representation of such properties leads to an improved performance in the model. The modelling approach utilizes a fully coupled hydromorphodynamic model based on TELEMAC-3D (v7p1) and an enhanced version of the sediment transport module SISYPHE (based on v7p1), which allows for a refined sediment representation (i.e. 10-class sediment mixtures instead of 2-class mixtures and distributed sediment density instead of uniform). The proposed developments of the SISYPHE model enable us to evaluate and discuss the added value of sediment representation refinement for improving sediment transport and riverbed evolution predictions. To this end, we used several model set-ups to evaluate the influence of sediment grain-size distribution, sediment density, and suspended sediment concentration at the upstream boundary on model predictions. As a test case, we simulated a flood event in a small-scale river, the Orne river in north-eastern France. Depending on the model set-up, the results show substantial discrepancies in terms of simulated bathymetry evolutions. Moreover, the model based on an enhanced configuration of the sediment grain-size distribution (10 classes of particle sizes) and with distinct densities per class outperforms the standard SISYPHE configuration, with only two sediment grain-size classes, in terms of simulated suspended sediment concentration.