In this article, we present 10 experiments carried out in a steep, shallow water ume made of an erodible bed of natural, uniform gravel. We simultaneously recorded bed load particle motion, bed and water elevation using two high-speed cameras. Particle trajectories were reconstructed with the help of a robust, original tracking algorithm, available on demand. Added together, the trajectories obtained represent more than 30 days' worth of data, with sample every 0.005 s. We share this data with the research community (https://goo.gl/p4GbsR). Because the ow was supercritical and sediment transport was moderate, two-dimensional bed morphologies spontaneously developed and migrated in the granular bed. The local ow and bed conditions over which bed load transport occurred were thus continuously changing, both spatially and temporally, precluding the use of traditional transport rate analysis. Instead, we propose an original statistical analysis that allows us to estimate how dependent particle velocity, particle acceleration, particle dffusivity, particle entrainment, and deposition rates are upon the variations of local bed shear stress and local bed slope. We found that: (i) particle velocity was linearly dependent upon the shear velocity; (ii) particle streamwise acceleration suggested the existence of two equilibrium particle velocities separated by an unstable equilibrium; (iii) particle diffusivity grew linearly with the shear velocity; (iv) the particle deposition rate was proportional to the inverse of the shear stress, and it decreased linearly with downward slopes; and (v) the entrainment rate was strongly correlated to the immediate proximity of other moving particles, but only weakly to the shear stress and bed slope. These original results could be used as inputs or parameters for modern bed load transport models and numerical simulations.
Giovanni De Cesare, David Hurther
Giovanni De Cesare, Romain Maxime Dubuis
Giovanni De Cesare, Romain Maxime Dubuis