Stochastic Linear Programming for Improved Reservoir Operations for Multiple Objectives in Burkina Faso, West Africa

A network of reservoirs and diversion structures in the Comoe River Basin in southwestern Burkina Faso, West Africa, provides municipal water supply and irrigation water for sugarcane agribusiness and a population of farmers. The region is characterized by severe intraseasonal and inter-annual variability with respect to precipitation and reservoir inflows. Reservoir operations are generally conservative, even during wet years. A stochastic linear programming model is introduced which translates seasonal streamflow and precipitation forecasts, in the form of a scenario tree, into optimal release schedules for reservoir operators to implement in real-time as forecasts and system conditions change. Goals include more efficient and equitable releases, and downstream flow maintenance. A VBA-based graphic user interface (GUI) is used to ensure implementation and ease of use by operators.