Bank failure algorithms are computationally expensive. These algorithms can increase sediment model run times by an order of magnitude. Users can reduce run times by defining BSTEM parameters only at the half-cross sections where bank processes are expected. Bank failure conditions (i.e., FS less than one) are not usually instantaneous. If FS drops below one, the failure condition generally lasts for several time steps. Skipping bank failure computations will introduce error, but given the uncertainties of the model, the error may be acceptable trade off for run time.