Quantifying the effects of spatially-distributed roughness parameters derived from Airborne Light Detection and Ranging (LiDAR) on a 2D hydrodynamic model of the Lower Yuba River, CA

Riparian vegetation in lowland rivers has an important effect on hydrodynamics and needs to be accounted for in predictive hydrodynamic modeling. Current 2D hydrodynamic models can spatially distribute vegetation, and often do so using discrete polygons to characterize ostensibly uniform vegetative patches into a small set of different vegetation class types. Roughness values are then empirically assigned to each patch type. In order for 2D models to accurately capture the complex out-of-bank flow interactions, a new methodology is needed to bridge a physical connection between vegetation presence and flow resistance across a range of scales. Airborne Light Detection and Ranging enables mapping the 1 m resolution grid of vegetation presence and canopy height. In this study a next-generation algorithm was used to spatially distribute stage-dependent channel roughness over a ∼37.5-km gravel-bed river corridor using equations from atmospheric boundary layer theory that reduce the data inputs to just estimated water depth and canopy height. Then 2D models were run for flows ranging from 0.2-20 times bankful discharge both with and without spatially distributed nodal roughness coefficients. Results were analyzed to gain insight into the stage-dependent and scale-dependent effects of vegetation on velocities, flow paths, and sediment transport capacity. At the floodplain inundating flow of 597.49 m3/s, adding spatially distributed vegetation roughness parameters caused an 8.0% increase in the wetted area within the full segment, a 7.4% increase in the mean depth, and a 17.5% decrease in the mean velocity compared with a constant nodal roughness model at the same flow. The results showed that vegetation has a strong channelization effect on the flow, increasing mid-channel velocities and diverting flow away from densely vegetated areas. On the floodplain, vegetation stands caused preferential flow paths that were otherwise unaccounted for in the unvegetated model runs.