REMOTE ACOUSTIC MEASUREMENT OF SUB-AQUEOUS GRAVITY-DRIVEN GRANULAR FLOWS

Abstract

The central goal of this thesis is to characterize the velocity structure and concentration within gravity-driven sub-aqueous granular flows using high-resolution acoustic remote sensing technologies. The experiments were carried out with both glass beads and sand. The results have implications for studies of granular flows in general and of bedload dynamics in coastal ocean and fluvial environments in particular. The geoacoustic properties of water-saturated sediments, which determine how the incident sound interacts with the granular medium, are investigated through measurements of sound speed and attenuation within the medium and the reflection coefficient at the sediment-water interface. The measurements are made in the scattering regime: i.e., 0.5<ka<1.2, where k is the acoustic wave number in water and a is the median grain radius. The results, as well as those reported in the literature, confirm the (ka)^4 dependence of attenuation and negative dispersion predicted by the multiple scattering theory of Schwartz and Plona (1984). Scaling the data by a factor depending on porosity and grain density is shown to substantially reduce the spread among the available sound speed estimates. The measured roughness of the sediment-water interface is Gaussian and the measured reflection coefficients are consistent with the Eckart (1953) prediction for a rough surface with Gaussian-distributed roughness. A single-scattering model of reflection from the sediment-water interface is developed and found to reproduce the statistics and spatial variations in the reflected amplitude, including the decorrelation lengths associated with the speckle pattern in the reflected pressure field. Sub-aqueous gravity-driven granular flows of O(1) cm thickness are investigated at mm-scale vertical resolution using a MHz frequency coherent Doppler sonar. The measurements are made for flow over both fixed roughness and erodible beds. Good agreement is obtained between the observed velocity profile and that predicted from an analytic low Reynolds number viscous fluid model. The model fits to the measured velocity profiles, combined with the Bagnold (1954) relation, yield estimates of effective viscosity and porosity within the moving layer that are respectively 400 (1100) times larger than that of water and 16% (7%) larger than the stationary glass beads (sand).