The stochasticity of fluid and sediment parameters has been identified as a source of diffusion, particularly anomalous diffusion at different temporal and spatial scales of bed load particle trajectories. Data from two sets of flume experiments are presented, one data set has gravel particle trajectories tracked over a limited area and was used in identifying the influence of different shear stress conditions on diffusive processes. A new experiment was performed using spherical particles moving as bed load in an annular flume in order to address concerns about censorship effects caused by the size of the detection window. An annular flume allowed collection of practically uncensored particle trajectories over longer time period than has been previously possible in the laboratory. Three diffusive regimes were observed at distinct stages of particle motion: (i) ballistic regime at the local range; (ii) Fickian diffusion at the intermediate range; (iii) subdiffusion at the global range. Characteristic time scales separate the regimes and correlate with the mean traveling and resting times of particles. Fickian diffusion in the intermediate range is first recognized as a result of the balance between intermittent weak transport and near‐bed turbulence, as first predicted by Nikora et al. (2002, https://doi.org/10.1029/2001WR000513). In the global range, extreme values were observed in the distribution of particle resting times, suggesting that two types of distributions (related to surface motion and vertical mixing) were responsible for the subdiffusion at longer time scales. Diffusion was found to be anisotropic at all stages of particle motion.

4A complete understanding of the role of grain-scale particle-flow interaction in sediment 5 entrainment and transport has still not been achieved in spite of recent technological advancement 6 in measurement capabilities. In this study the initial motion of natural sediment particles in a gravel 7 deposit was detected and combined with simultaneous local measurements of the velocities on a 8 horizontal plane located above the bed surface using a three-component stereoscopic PIV. A series of 9 experimental tests with increasing low values of boundary shear stress were conducted. The 10 acquisition system allowed coupling between streamwise and vertical near-bed velocity and the 11 entrainment of more than 900 individual grains. Initial analysis agreed with previous observations on 12 the predominance of sweeps (Quadrant IV) and to a lesser extent of outward interactions (Quadrant 13 I) in entraining gravel particles. However, the latter were found to move sediments just as efficiently 14 as sweeps impacting on particles that had long periods of rest and so were exhibiting higher levels of 15 stability. This behavior suggests that sweep-induced lift based on Bernoulli's principle does not 16 entirely explain the generation of vertical forces on highly stable bed particles. Closer inspection of 17 the data revealed that many entrainments were correlated to occasions when stable bed grains 18 interacted with grains travelling in their close vicinity. Around 30% of the entrained population was 19 observed to initiate motion in this type of situation. For this subsample of entrainment events 20 outward interactions were found to be comparatively more effective than for the non-interference 21 case, while the relative contribution of sweeps exhibited an opposite trend. 22

No abstract

No abstract

No abstract