A new approach to estimate bed load including the spatial distribution of shear stress: Oak Creek Revisit

Abstract:

Understanding sediment transport in gravel-bed streams is critical to the management of water and ecological resources. However, the predictions of bed load remain challenging. Most bed load equations describe the hydraulic characteristics of a given flow with the reach-averaged shear stress. However, the flow field in these rivers varies widely resulting in wide distributions of shear stress. These spatial variations in flow properties are poorly represented by the mean value and are partially responsible for inaccuracies in sediment transport estimates. We hypothesized that improved estimations of bed load can be achieved by including the complete spatial distribution of shear. In this study, we modified a subsurface-based bed load transport equation to include the load moved for relative increments of shear stress values for a given flow within a reach. The accuracy of our method was tested using the historical database collected by Oak Creek, OR. Spatially variable flow properties were estimated with the two-dimensional flow model FastMECH, which was calibrated for flow levels between 0.1 and 1.0 bankfull discharges. In these flow conditions, we observed similar shear stress distributions shapes. This shape was described using a Gamma probability function, which parameters were predicted as a function of discharge. The proposed equation predicted similar bed load transport rates to previously published methods for discharge levels that mobilize “break” the pavement layer. However, our equation is also applicable over a wider range of flows providing estimates of bedload when the surface pavement layer is still present in the channel bed. Bed load at these lower flows is not predicted with the original formulation. Our proposed equation accurately predicts the observed bed load transport rates because it accounts for sediment transport occurring in discrete regions of the channel bed that experience shear stress values much higher the average.