Development of Risk-Based Reservoir Operating Rules Using Stochastic Inflow Forecasts

Developing robust reservoir operation policies under uncertainty is becoming a more important topic due to climate change, severe droughts, and increasing water demands. Deterministic hydrologic models do not always provide an adequate range of possible outcomes to consider for robust operating schemes. In this study, one Istanbul reservoir located on European continents is modelled within the GoldSim dynamic simulation environment to develop risk-based operating rules. The model produces stochastic climate time series generated by the National Weather Service’s Weather Generator (WGEN), which feeds a runoff model to generate probabilistic inflow forecasts to the reservoir models. The reservoir models use operating criteria to evaluate the risk of supply shortages and over topping the spillway. The risk of exceeding operating criteria is measured and used to dynamically update operating controls to reduce risk.