Date Awarded


Document Type


Degree Name

Master of Science (M.Sc.)


Virginia Institute of Marine Science


Carl T Friedrichs

Committee Member

Courtney K Harris

Committee Member

Linda C Shaffner

Committee Member

Nina Stark


The erodibility of estuarine sediment beds has a number of ecological and societal implications, including alteration of benthic habitats, light-limitation of primary production, and reintroduction of pollutants to the water column, as well as impacts on dredging operations and on the fate of potentially dangerous objects on the seafloor. The objectives of this study are to better understand controls on bed erodibility in estuarine environments, including the roles of sediment grain size, water content, percent organics, and bed fabric, as well as effects of tides, storms, salinity distribution, river discharge, and location within the estuary. An extensive set of erosion experiments, with supporting sediment properties and hydrodynamic information, was utilized to create statistical models to predict bed erodibility. This data set included measurements from over 150 Gust microcosm erodibility cores taken over a span of 15 years at various sites along the York River Estuary, Virginia. For this study, erodibility was defined in two ways, the first being the mass of sediment eroded at an applied stress of 0.2 Pa (“eroded mass”), and the second being the normalized shape of the profile of eroded mass as a function of applied stress (“erosion profile shape”). By considering these two different definitions of erodibility, the goal was to determine how erodible the sediment is at a single commonly occurring stress (0.2 Pa) and then infer how the underlying bed may erode at higher stresses due to stronger tides or storms. The dataset was analyzed in its entirely and also in data subsets that corresponded to spatially and hydrodynamically distinct areas throughout the river. A series of multiple linear regressions were formulated to predict eroded mass at 0.2 Pa along with the shape of the erosion profile for each data subset. In models with eroded mass as the response variable, the explanatory variables with the strongest influence included tidal range squared, percent organic content, salinity, river discharge, and water level anomaly. With the exception of water content, the signs of the relationships between eroded mass and each of the explanatory variables were largely consistent with that anticipated based on previous literature. Results showed that the hydrodynamic properties, such as greater tidal range (which increases erodibility due bed disturbance) and decreased salinity (which increases erodibility in due to new deposition), together influenced eroded mass more than the sediment properties. Furthermore, eroded mass at 0.2 Pa was by far the most important variable in explaining the erosion profile shape (linearly increasing for high erodibility cases, and exponentially increasing for low erodibility cases), with more than double the influence of any other explanatory variable. The control of eroded mass at 0.2 Pa by hydrodynamic variables, and the strong relationship between eroded mass and erosion profile shape means that relatively accurate predictions of both erodibility parameters can be made based solely on hydrodynamic variables. Further discussion includes potential explanatory variables that were not part of the dataset, but may influence erodibility, as well as future steps for improved experimental design, such as more randomized sampling and more precise measurement of sediment water content.




© The Author