Sediment fluxes in a partially stratified estuary

This dissertation investigates the sediment fluxes observed in a partially stratified estuary. In an experiment spanning from neap to spring tides, vertical profiles of velocity, salinity and suspended sediment were collected within a channel in northern San Francisco Bay.; During weaker neap tides, the water column was almost always dominated by stratification. In fact, the straining of the density field on neap ebb tides created stratification that was sufficiently strong to almost entirely damp out the turbulence responsible for bed shear stress. Suspended sediment which did appear at the study site on ebb tides was delivered via advection. In contrast, neap flood tides had a near bed mixed layer which was able to sustain bed shear stress and erode sediment even though it was capped by a density interface. During spring tides, active turbulent mixing filled the water column on every tide except for the weaker ebbs.; An acoustic velocimeter was concurrently used to measure velocity and suspended sediment concentration. Together, these measurements were used to estimate the turbulent vertical sediment flux in the near bed environment. This flux estimate correlated well with periods of erosion; it even differentiated between periods with high suspended sediment load resulting from strong advection and periods of erosion. Longitudinally spaced autonomous profilers were used to calculate the horizontal advective sediment flux. These estimates of advective flux were important to the change in suspended sediment, exceeding the erosion estimate by up to a factor of three. These two direct estimates of erosion flux and advective flux, in conjunction with a continuous deposition flux, were integrated in time. The suspended sediment calculated by this simple water column model matched observations reasonably well when the particle settling velocity was adjusted to 1.5 mm/s.; The vertical structure of sediment agreed closely with the structure of turbulent mixing field, as represented by the gradient Richardson number and the eddy viscosity. Incorporating the Munk-Anderson damping function and the turbulent active water column depth substantially improves the modeling of eddy viscosity and somewhat improves modeling of suspended sediment profiles.