On the Impact of Bed-Bathymetry Resolution and Bank Vegetation on the Flood Flow Field of the American River, California: Insights Gained Using Data-Driven Large-Eddy Simulation

Abstract

High-fidelity models can simulate many detailed hydrodynamic features of flow in natural rivers. However, it is challenging to explicitly capture the effects of highly complex bathymetry and bank line vegetation on the simulation results of models. Often detailed hydrographic data representing the channel bed is unavailable, requiring the use of more simplified representations of the channel bathymetry. Furthermore, explicit modeling of vegetation is generally not accounted for in most three-dimensional (3D) numerical models, which could lead to significant error in obtaining accurate results. In this study, we have investigated these issues by comparing the effect of (1) different sonar survey resolutions of the channel bed, and (2) trees on the large-eddy simulation (LES) results of a test case in a reach of the American River, California. This comparison has shown that the higher resolution survey reduced the near bed velocities and bed shear stresses and increased the number of vortical structures in areas with highly detailed surface structures. The hydrodynamic results from the tree-resolving LES were compared against a baseline LES case that excluded trees. The results indicate that the inclusion of trees in the simulations can significantly alter the flow dynamics of the river by decreasing the flow momentum and bed shear stress along the banks, while increasing the momentum and stresses in the middle of the river. While trees appear to increase turbulent fluctuations near the bank, these fluctuations probably do not contribute to erosion processes. These findings have the potential to improve the accuracy of high-fidelity modeling of natural streams and rivers and add to the general understanding of the flow dynamics in natural waterways.