CALIBRATION OF A BANK EROSION MODEL ON A LARGE MIDWESTERN ALLUVIAL RIVER

Drawing upon the work of Osman and Thorne (1988), a model of bank erosion and bank failure was created for a large alluvial river located in the Great Lakes basin. The model was successfully calibrated to a dataset of 20 transects surveyed approximately five and one-half years apart. Bank erosion can be a significant component of non-point source loads, but is often neglected in watershed analysis. This model offers a simple, yet conceptually rigorous technique for estimating and forecasting first order approximations of bank erosion and bank failure losses.The bank erosion model is written in FORTRAN and utilizes a Newton-Raphson technique to solve for water depth based on bed slope, bottom width, bank height and slope, mannings n, and flow.Bed shear is calculated as a function of the average cross-section velocity and roughness height. This shear is adjusted by an empirical correction factor that accounts for the effects of river bends. The model utilizes Lick's equation (Gailani, et.al., 1991) to calculate bed erosion of cohesive sediments. The erosion routine accounts for the timedependent effect of armoring. The model checks bank stability at every time step and calculates bank failure area and mass lost if stability criteria are exceeded.The underlying assumption for appropriate model application is that the river under study is undergoing systemic adjustment (bed degradation and bank width increase) to outside forces, such as urbanization or dam decommissioning.This paper presents the model formulation, the data requirements and the process of application. Data and results from the calibration will be included as an example of an appropriate application. The model's limitations, associated uncertainty and recommendations for its use as a forecasting tool on other systems will also be covered.