Fluctuating flow releases on regulated rivers destabilize downstream riverbanks, causing unintended, unnatural, and uncontrolled geomorphologic changes. These flow releases, usually a result of upstream hydroelectric dam operations, create manmade tidal effects that cause significant environmental damage; harm fish, vegetation, mammal, and avian habitats; and destroy riverbank camping and boating areas.

This work focuses on rivers regulated by hydroelectric dams and have banks formed by sediment processes. For these systems, bank failures can be reduced, but not eliminated, by modifying flow release schedules. Unfortunately, comprehensive mitigation can only be accomplished with expensive rebuilding floods which release trapped sediment back into the river.

The contribution of this research is to optimize weekly hydroelectric dam releases to minimize the cost of annually mitigating downstream bank failures. Physical process modeling of dynamic seepage effects is achieved through a new analytical unsaturated porewater response model that allows arbitrary periodic stage loading by Fourier series. This model is incorporated into a derived bank failure risk model that utilizes stochastic parameters identified through a metaanalysis of more than 150 documented slope failures. The risk model is then expanded to the river reach level by a Monte Carlos simulation and nonlinear regression of measured attenuation effects. Finally, the comprehensive risk model is subjected to a simulated annealing (SA) optimization scheme that accounts for physical, environmental, mechanical, operations, and flow constraints.

The complete risk model is used to optimize the weekly flow release schedule of the Glen Canyon Dam, which regulates flow in the Colorado River within the Grand Canyon. A solution was obtained that reduces downstream failure risk, allows annual rebuilding floods, and predicts a hydroelectric revenue increase of more than 2%.