Clark University

Rivers shape continental surfaces; they erode continental reliefs, transport erosion products and finally deposit them in plains. As a consequence, understanding physics of sediment transport is of fundamental importance in earth sciences. However, sediment transport occurs at the dynamical interface between fluid and granular mechanics, which makes its full understanding challenging. Over the last century, the complexity of river sediment transport dynamics has been mainly investigated from the hydrodynamics perspective. In particular, the noticeable fluctuations of flux of particles carried close to the bed called bed load has been associated with the flow turbulence fluctuations. We propose here to rather investigate the granular aspects of sediment transport. Bagnold (1956) was one of the first to propose a simple argument to predict the base of the bed load layer, based on a yield stress able to dislodge particles inside the bed; yet today too few sparse datasets support neither quantify this hypothesis. As an attempt to remedy this situation, we have performed new bed load experiments in a annular flume where we immerse a layer of particles whose refractive index has been matched with the fluid one. Then, using a laser sheet positioned at the middle of the channel, we image the vertical bed structure under different flow shear stresses. The bed being sheared by a laminar flow, we investigate sediment transport features without any turbulence effect. This unique setup allows us to detect particles and compute their average concentration and velocity profiles over long times and thus to record very slow vertical dynamics of particles of a fully 3D infinite river-like system. We found that the system exhibits two distinct characteristic elevations : one at the surface, where the particles concentration decreases abruptly, and a second one deeper inside the bed, where after a first fast decay with the distance to the surface, the velocity profile exhibits a sharp change of trend to present deeper a very weak dependance with the bed elevation. This second critical elevation occurs at a fixed normalized shear rate a submerged version of the Inertial Number that we interpret as a vertical transition from a granular flow regime to a creeping regime. These results lead to fundamental implications for the rivers and landscapes dynamics. We will also discuss at the end how they are situated at the corner of different recent granular rheology problems. Co-authors: Morgane Houssais, Carlos P. Ortiz, Doug J. Durian and Doug J. Jerolmack Tuesday, November 25, 2014, 2:00 pm Room BP 211/Math-Physics Building