Mixture theory model sensitivity to effective viscosity in simulations of sandy bedform dynamics

We perform a sensitivity analysis on a three-dimensional bottom boundary layer model (SedMix3D) that uses mixture theory to simulate the flow and sediment transport over rippled sand beds. SedMix3D treats the fluid-sediment mixture as a single continuum with effective properties that parameterize the fluid-sediment and sediment-sediment interactions using several closures for the sediment phase. The effective viscosity is one such closure that includes three adjustable parameters: the intrinsic viscosity, the maximum viscosity, and the maximum packing concentration of unconsolidated sediment. The sensitivity of suspended sediment concentration predictions by SedMix3D is tested by varying the intrinsic viscosity, which is a proxy for sediment grain shape. We qualitatively and quantitatively analyze the model output of suspended sediment concentration for a range of intrinsic viscosity values typical of quartz sand. Intrinsic viscosity values ranging from 2.5 to 3.5 produce total suspended sediment concentrations that differ less than 11%. However, there is approximately a 16% difference between the suspended sediment concentrations from intrinsic viscosity values of 2.5 to 3.5 and 4.0 to 5.0. Simulations of sediment transport over bedforms performed here were not significantly sensitive to the choice of an intrinsic viscosity value in the range of 2.5 to 3.5. Using a baseline intrinsic viscosity value of 3.0, we subsequently tested two additional effective viscosity formulations. The suspended sediment concentrations predicted by the Eilers and Krieger-Dougherty formulations were very similar, but the Mooney formulation generated much less suspended sediment. We found the model to be more sensitive to variations of effective viscosity in the ripple-fluid interface than in the suspension range.