Utilizing a Decision Support System to Optimize Reservoir Operations to Restore the Natural Flow Distribution in the Connecticut River Watershed
Richard N. Palmer
Reservoir development on the Connecticut River Basin has altered the natural hydrograph and detrimentally affected the region's aquatic ecosystems that thrive on natural variability. This thesis examines ways in which flows in the river and its triburaries can be returned to a more natural regime withouy negatively impacting other users. The analysis offers alternative operations to meet the various objectives as well as evaluates the impacts of climate change on the basin. This paper focuses on the upper third of the basin as a proof of concept to demonstrate the decision support system and to generate initial results for the large effort.
The upper third of the basin contains 11 sub-watersheds, six major reservoirs, and three ecological points of interest. The analysis employs simulation, optimization, and hydrology models that work in symphony. The process starts with a simulation model of the system that reflects the current reservoir operations; this assures the modeling efforts accurately portray the current conditions. The linear program optimization model is used to determine optimal operating policies and the trade-offs between reservoir objectives for three operational schemes: operating for ecological targets, hydropower, and flood control. The output from the optimization model suggests how to alter the operating polices to achieve the given objective. Downscaled climate change meteorological data (temperature and precipitation) from five General Circulation Models and two emissions scenarios are used as input into the hydrology model to produce climate impacted streamflow to examine reservoir operations under climate change. The analysis does not suggest a single operating policy, but rather a range of options to be discussed among stakeholders to best address all objectives.