Predicting Bedload Transport in Spatially Heterogeneous Flows: Statistical Characterization based on Large-Eddy Simulations

Bedload transport is governed by complex, nonlinear processes that span a wide range of spatial and temporal scales, from the scale of individual sediment grains to the reach scale at which it is typically estimated. Even under uniform flow conditions, bedload transport exhibits strong variability due to intense turbulent fluctuations and near-bed coherent structures. In more complex flows, such as those involving large immobile boulders, the flow becomes highly nonuniform, and sediment motion is also affected by spatial variations in near-bed turbulence induced by the macroroughness elements. Despite these complexities, bedload transport is often estimated using a global shear stress derived from the depth-slope product, neglecting the temporal and spatial variability of the flow and sediment fluxes. This work aims to improve our understanding of the grain-scale mechanisms that control bedload transport in nonuniform, turbulent flows and to incorporate the relevant temporal and spatial variations into transport predictions. To achieve this, we conducted large-eddy simulations (LES) over a fixed rough bed with an array of boulders coupled with the discrete element method (DEM) to resolve the trajectories and interactions of mobile sediment grains. The resulting high-resolution, grain-scale numerical data allow for a detailed statistical characterization of bedload transport in spatially heterogeneous flows. Furthermore, through the analysis of probability density functions and the fitting of large-scale stochastic transport models, we aim to incorporate the inherent temporal and spatial variability of bedload fluxes and sediment dynamics driven by turbulence into bedload predictions. From this analysis we offer new insights into the processes controlling bedload transport in complex environments and provide a statistical framework to improve transport models in natural streams.