Improved Streamflow in Adaptive Reservoir Operation

The main focus of this study is to improve short-term streamflow forecasts and to utilize them in adaptive reservoir operation to enhance sustainability. A combination scheme is developed to integrate reforecasts from a numerical weather model and disaggregated climate forecasts from ECHAM4.5 for developing 15-day ahead precipitation forecasts. Evaluation of the weather-climate information (WCI) based biweekly forecasts under leave-five out cross-validation shows that WCI based forecasts perform better than reforecasts in many grid points over the continental United States. WCI forecasts perform particularly better during summer months when reforecasts have a limited skill. Even though the disaggregated climate forecasts do not perform well at many grid points, the primary reason WCI based forecasts perform better than reforecasts is due to the reduction of the overconfidence of reforecasts. Since the Disaggregated Forecasts (DF) are better dispersed than Reforecasts (RF), combining them with reforecasts results in improved 15-day ahead precipitation forecasts. Improved precipitation and streamflow can be used for adaptive management of water resources system, e.g., reservoir operation.;An adaptive reservoir operation framework is proposed to explicitly incorporate ecological flow requirements. A general optimization-based decision model is presented to consider simultaneously anthropogenic water uses of a reservoir and desirable ecological reservoir releases represented by hydrological alterations parameters that capture the impacts on flow regime. Multiple uses of the reservoir, including water supply and hydropower generation, are modeled within a mixed integer linear programming model. Investigations based on a model of a reservoir in Virginia demonstrate that compared to standard releases based on the current operational practices, releases simulated using this adaptive framework perform better in mimicking natural flow or pre-impact flows. Perfect streamflow forecasts are used in this analysis. In addition, to understand how the reservoir operation model could benefit when coupled with streamflow forecasts, synthetic Ensemble Streamflow Prediction (ESP) of a month-ahead daily streamflow forecasts is integrated into the adaptive reservoir operation framework. Results demonstrate that synthetic ESP with better skill improves water allocation for intended uses of the reservoir and achieve better flow conditions in the river downstream of the reservoir. The value of updated-ESP in sustaining reservoir operation is also examined as compared to non-updated-ESP. Results show that updating forecasts further improves reservoir operation in term of meeting human needs and ecosystem health.