Reservoir yield estimates with consideration of uncertainty in used data in Illinois

Abstract

Reservoir yield estimates are necessary and required for better water supply to communities especially during a severe drought. This study provides a framework to estimate reservoir yields with consideration of associated uncertainties in used data. Errors exist in inflow, reservoir capacity, evaporation and precipitation data and contribute to the overall uncertainty in reservoir yield estimates. Before combining optimization with Monte Carlo simulation, errors of each data category are assumed to follow a certain normal distribution. The framework is applied to three reservoirs in Illinois. It is found that the 95% probability intervals surrounding the estimates of reservoir yields range between -29% and +42% of the best estimate and the range is a bit right-shifted; evaporation contributes the most to the overall uncertainty, followed by reservoir capacity of small reservoirs and inflow to large reservoirs.