Grain-size distributions of bed load: Inferences from flume experiments using heterogeneous sediment beds

The grain-size fractions in the bedload transported over the five heterogeneous sediment beds of different values of bed roughness were computed from the flume experiments. The existence of an entrapment factor associated with the sorting observed from the bed to active layer was modeled based on the modified critical shear stress to estimate the grain-size fractions in the transport layer under given hydraulic conditions. The efficiency of these models was tested with the observed data. Subsequently, the patterns of observed grain-size distributions in the transport layer were tested to identify the distributions developed in the active layer due to sorting using three probability density functions (pdf), such as, log-normal, log-hyperbolic and log-skew-Laplace. Tests indicated that a log-skew-Laplace distribution fitted best for 49%, a log-hyperbolic for 31%, and a log-normal for 20% out of forty-five bedload samples collected under unidirectional flow with changes in flow discharge and bed roughness. The results of this study would be useful to specify the grain-size distributions in the bedload formed under different hydrodynamic conditions in various sedimentary environments.