A key problem in the decadal morphodynamic modelling of tidal inlet system is on the evolution of the inlet channels, especially the unrealistic channel incision. The present study attempts to couple two processes, including the bedform roughness and sediment grain-size sorting, in the two-dimensional, depth-averaged process- based model to solve the problem. Based on the comparison of the modelling results and relevant field observations in the Dutch Wadden Sea, it is suggested that bedform roughness height predictor is a useful tool for determining spatial and temporal heterogeneous bed frictions.Within tidal channels, bedformroughness is dominated by dune roughness. In order to represent “the memory of old bedforms”, a relaxation time method was implemented. The corresponding large relaxation time (2 M2 tidal periods) smooths the tidal variations of dune roughness, and makes it respond to hydrodynamic changes on a longer time scale.When decadal morphological modelling is carried out, the model performance can be significantly improved by introducing either bedform (dune) roughness predictor or sediment sorting processes. The mechanisms are twofold. First, in the deep channels where flow velocities are high, large dunes develop and cause large bed drag coefficients. Second, coarser bed sediments within the channel area induced by sorting prevent the channel incision. If both effects are coupled in the morphodynamic modelling, the inlet channel incision is further controlled, which is not just a result of the linear superposition of both effects, because sorting can further promote the development of dunes within the deep channels. @en