Unraveling bed slope from relative roughness in initial sediment motion

Abstract Understanding incipient sediment transport is crucial for predicting landscape evolution, mitigating flood hazards, and restoring riverine habitats. Observations show that the critical Shields stress increases with increasing channel bed slope, and proposed explanations for this counterintuitive finding include enhanced form drag from bed forms, particle interlocking across the channel width, and large bed sediment relative to flow depth (relative roughness). Here we use scaled flume experiments with variable channel widths, bed slopes, and particle densities to separate these effects which otherwise covary in natural streams. The critical Shields stress increased with bed slope for both natural gravel ( ρ s = 2.65 g/cm 3 ) and acrylic particles ( ρ s = 1.15 g/cm 3 ), and adjusting channel width had no significant effect. However, the lighter acrylic particles required a threefold higher critical Shields stress for mobilization relative to the natural gravel at a fixed slope, which is unexpected because particle density is accounted for directly in the definition of Shields stress. A comparison with model predictions indicates that changes in local velocity and turbulence associated with increasing relative roughness for lighter materials are responsible for increasing the critical Shields stress in our experiments. These changes lead to concurrent changes in the hydraulic resistance and a nearly constant critical stream power value at initial motion. Increased relative roughness can explain much of the observed heightened critical Shields stresses and reduced sediment transport rates in steep channels and also may bias paleohydraulic reconstructions in environments with exotic submerged densities such as iron ore, pumice, or ice clasts on Titan.