Bedload Transport in Laboratory Rivers

Abstract

This chapter is a review of the erosion-deposition model of bedload transport, first proposed by Charru et al. (2004). This model, established on the basis of laboratory experiments, proposes a simplified theoretical framework describing the exchange of particles between the sediment bed and the layer of entrained grains. The latter is treated as a uniform reservoir of independent particles moving at velocity V. The exchange of particles between the sediment bed and the bedload layer sets the surface concentration of moving particles, n, and determines the sediment transport rate, q s = nV. For a steady flow over a wavy topography, the sediment flux adjusts to a change of shear-stress over a characteristic deposition length. This relaxation plays a crucial role in selecting the wavelength of bedforms such as ripples, rhomboid patterns and bars (Andreotti et al., 2012; Charru, 2006; Charru et al., 2013; Charru and Hinch, 2006; Devauchelle et al., 2010b). The equations describing bedload transport and the development of bedforms are similar in laminar and turbulent flows. This analogy explains why most alluvial morphologies formed in nature also form at a smaller scale when a laminar flow interacts with a sediment bed (Paola et al., 2009). A preliminary investigation indicates that the erosion-deposition model also accounts for the propagation of a plume of tracers entrained by bedload transport. It suggests that, after a short transient, the plume reaches asymptotically an advection-diffusion regime, in which it advances with a constant velocity while its variance increases linearly with time. The results discussed in this chapter are based on idealized laboratory experiments (uniform grain size, constant flow and sediment discharges). Rivers, and in particular gravel-bed rivers, often exhibit high grain size heterogeneities and fluctuating flow discharges. Recent progress in extending the erosion-deposition model to the transport of mixed grain sizes should allow us to investigate more realistic configurations (Houssais and Lajeunesse, 2012; Houssais et al., submitted).