A major challenge in river morphodynamic modelling is to properly predict the evolution of bed characteristics such as the formation of an armoured layer or grain size evolution. This paper introduces the 1D hydraulic mobile-bed model RubarBE. Its ability to reproduce the downstream fining of a gravel-sand mixture in response to bed aggradation is tested against the experiments reported by Seal et al. (1997). Contrary to most of the numerical models, the sediment grain size distribution in each layer is not represented using a multi-class model, but using the median diameter d50 and the sorting coefficient sigma. The results underline the impact of the sediment feeding rate on sediment processes and on the bed evolution. Sensitivity tests on the mesh size and sediment transport parameters were performed. The impacts of each parameter are analysed and compared with tests from the literature. Using the Meyer-Peter and Müller formula, best results are obtained for a mesh size Δx = 1 m and a lag distance Lc = 3 Δx. The front velocity is properly reproduced by the model. The downstream fining is also in good agreement with the observations. However, the model does not reproduce the vertical sorting observed in the experiments. Indeed, the coaser grain-size of the sediments stored at the surface layer cannot be recreated. As the grain size results underline the actual limits of the 1D model, some improvements of the model are proposed for representing this infiltration.Changes in morphology were assessed by DEM differencing. The DEM methodology based on the morphological unit scale provides an explicit identification of spatial patterns of erosion and deposition. The different morphological response of the reach to similar flushing events is explained by the severe incision of the main channel observed after the 2008 flood which was accompanied by the transformation of the upstream bar to an island and therefore a change on flow direction.