Abstract
The threshold for gravel motion shapes aquatic habitat and serves as the basis for models of bed material sediment transport. Despite decades of research on the subject, geomorphologists and civil engineers are still struggling to make sense of the substantial variability in the dimensionless shear stress associated with the onset of sediment motion, known as critical Shields stress. Many recent advances in our understanding of the threshold for grain motion come from flume experiments or reanalysis of datasets from an extremely steep natural channel. While these studies have demonstrated systematic variation in critical Shields stress with varying sediment supply, flow history effects, and grain protrusion above the bed surface, it is not yet clear to what extent these effects are important in most natural, gravel-bedded channels. We ask: how does the variability in critical Shields stress between channels compare to the variability through time, in a single channel? And how does that compare to the variability between different patches in a single channel at a point in time? In this presentation, we explore these phenomena via measurements made in small gravel-bedded natural channels in the lowlands of the Pacific Northwest. We conducted tracer particle and force gauge measurements in channels of similar slope, morphology, bed surface grain size, and flow sequencing, but varying sediment supply. These measurements give us two independent estimates of critical Shields stress. We find no significant difference in critical Shields stress between high and low sediment supply channels. This contrasts with the expected relationship based on recent experimental studies and suggests that the overall sediment supply condition of a channel may not be a first-order control on the baseline critical Shields stress in natural channels. Reassuringly, we find only modest variation in critical Shields stress between patches in a single reach at a given time, and changes from one storm to another are consistent across most patches. This suggests that temporal variability exceeds patch heterogeneity, at least in small streams.