Creepy landscapes : river sediment entrainment develops granular flow rheology on creeping bed.

Abstract:

To predict rates of river sediment transport, one must first address the zeroth-order question: when does sediment move? The concept and determination of the critical fluid shear stress remains hazy, as observing particle motion and determining sediment flux becomes increasingly hard in its vicinity. To tackle this problem, we designed a novel annular flume experiment - reproducing an infinite river channel - where the refractive index of particles and the fluid are matched. The fluid is dyed with a fluorescent powder and a green laser sheet illuminates the fluid only, allowing us to observe particle displacements in a vertical plane. Experiments are designed to highlight the basic granular interactions of sediment transport while suppressing the complicating effects of turbulence; accordingly, particles are uniform spheres and Reynolds numbers are of order 1. We have performed sediment transport measurements close to the onset of particle motion, at steady state, and over long enough time to record averaged rheological behavior of particles. We find that particles entrained by a fluid exhibit successively from top to bottom: a suspension regime, a dense granular flow regime, and - instead of a static bed - a creeping regime. Data from experiments at a range of fluid stresses can be collapsed onto one universal rheologic curve that indicates the effective friction is a monotonic function of a dimensionless number called the viscous number. These data are in remarkable agreement with the local rheology model proposed by Boyer et al., which means that dense granular flows, suspensions and bed-load transport are unified under a common frictional flow law. Importantly, we observe slow creeping of the granular bed even in the absence of bed load, at fluid stresses that are below the apparent critical value. This last observation challenges the classical definition of the onset of sediment transport, and points to a continuous transition from quasi-static deformation to granular flow. These results provide a new perspective to connect the transport laws for soil creep, landslides/debris flows and river transport. Although our experiments are highly idealized, evidence from other studies suggest that our observations may be directly relevant to natural systems. Finally we show that our findings are robust for mixed grain sizes.