Slow Granular Dynamics in River Beds: Toward a Continuous Definition of Bed Load Transport.

Abstract:

Field records and experimental studies show that the fluvial geophysical processes that shape the landscape, such as debris flows and river sediment transport, are extremely unpredictable in large part due to the nonlinear dependence of the transport rates on the structural properties of the sediment.

There is a need for a more fundamental understanding of the physical processes that control sediment transfer rates, particularly the magnitude and frequency of the bed load transport flux.

We present experiments in a simple geometry, an annular couette cell, that allows us to study the free-surface dynamics at the interface between a flowing viscous fluid and a submerged particle bed made of plastic spheres, a highly idealized river. This geometry presents an opportunity to study details of the bed structure and particle transport with a well-controlled steady shear stress during long-time experiments.

We use the refractive-index-matched laser scanning technique [Dijksman et al. 2012], to detect the particle positions on a two-dimensional vertical slice at the middle of the 15 particle wide bed, and characterize their dynamics over a range of timescales of six-orders of magnitude. We find that the particle dynamics are spatiotemporally heterogeneous, but that the overall flow field reaches a well-developed steady-state. Below the fluid flow depth, we find a wide flowing layer characterized by a fast, approximately exponential decay of the particle velocity versus depth. This layer can be associated with the active layer commonly mentioned in the literature. We find that the thickness of the flow layer increases with the applied shear stress. However, deep in the bed, the velocity profile does not indefinitely follow a exponential decay. Instead, the rate of decay of the velocity profile slows drastically, transitioning continuously to a quasistatic flow regime, with a very different exponential decay.

This study provides a new framework for understanding the threshold of bed load particle transport, as the result of a jamming process occurring at a critical depth. We test different rheological hypothesis in order to predict the entire particle velocity profile.