Controls on Sediment Bed Erodibility in a Muddy, Partially-Mixed Tidal Estuary

The objectives of this study are to better understand controls on bed erodibility in muddy estuaries, including the roles of both sediment properties and recent hydrodynamic history. An extensive data set of erodibility measurements, sediment properties, and hydrodynamic information was utilized to create statistical models to predict the erodibility of the sediment bed. This data set includes >160 eroded mass versus applied stress profiles collected over 15 years along the York River estuary, a system characterized by "depth-limited erosion," such that the critical stress for erosion increases rapidly with depth into the bed. For this study, erodibility was quantified in two ways: the mass of sediment eroded at 0.2 Pa (a stress commonly produced by tides in the York), and the normalized shape of the eroded mass profile for stresses between 0 and 0.56 Pa. In models with eroded mass as the response variable, the explanatory variables with the strongest influence were (in descending order) tidal range squared averaged over the previous 8 days (a proxy for recent bottom stress), salinity or past river discharge, sediment organic content, recent water level anomalies, percent sand, percent clay, and bed layering. Results support the roles of 1) recent deposition and bed disturbance increasing erodibility and 2) cohesion/consolidation and erosion/winnowing of fines decreasing erodibility. The most important variable influencing the shape of the eroded mass profile was eroded mass at 0.2 Pa, such that more (vs. less) erodible cases exhibited straighter (vs. more strongly curved) profiles. Overall, hydrodynamic variables were the best predictors of eroded mass at 0.2 Pa, which, in turn, was the best predictor of profile shape. This suggests that calculations of past bed stress and the position of the salt intrusion can serve as useful empirical proxies for muddy bed consolidation state and resulting erodibility of the uppermost seabed in estuarine numerical models. Observed water content averaged over the top 1 cm was a poor predictor of erodibility, likely because typical tidal stresses suspend less than 1 mm of bed sediment. Future field sampling would benefit from higher resolution observations of water content within the bed's top few millimeters.