Hydraulic computations are sensitive to bed roughness. Bed roughness is a dynamic property and changes in response to sediment dynamics. If a fine pulse covers a coarse substrate, bed roughness will drop. As flow increases on sand rivers, bed form amplitude increases, increasing bed roughness, until the river passes into a plane bed regime which results in substantial drops in n-values.
Mobile bed sediment models simulate changes in bed gradation, which the model can use to compute roughness, computing feedbacks between sediment transport and bed roughness. HEC-RAS includes three bed roughness predictors, equations and algorism that compute manning's-n from hydraulics and sediment properties, including: Limerinos, Brownlie, and Van Rijn. The bed roughnes predictors, compute new manning's n values each time step, as the bed gradation and cross section evolve, which the program uses to compute hydraulics in the next computational increment.