Multireservoir Allocation Framework Considering Societal and Ecological Needs in a Time–Frequency Domain

Existing reservoir management frameworks traditionally consider historical (predam) flow conditions to deliver environmental flows. Such frameworks may not be feasible because current demand and/or climate could be different from predam conditions. Hence, we developed a multireservoir framework that explicitly considers both human water demands and environmental flow requirements to minimize deviations under current hydroclimatic conditions and demand patterns. The multireservoir framework, Generalized Reservoir Analyses using Probabilistic Streamflow (GRAPS), was modified and implemented to solve the problem of minimizing the flow deviations using feasible sequential quadratic programming for three reservoirs in the Chattahoochee River Basin, Southeastern United States, which is known for its imperiled native biodiversity and productive estuarine ecosystem. Our results show that downstream reservoirs in the cascade system are less influenced by upstream reservoirs’ regulation because the downstream reservoirs receive a significant amount of natural flows. By comparing the average wavelet power spectrum at different periodicities between natural flows and downstream releases, we found that the current release policy and modified releases resulted in highly altered flows under shorter periodicities (e.g., less than 2 months) but synchronized flow variance between natural flow and downstream releases at longer periodicities (e.g., greater than 3 years). This framework of linking the multireservoir allocation model through the time–frequency analysis using wavelet power spectrum could not only advance sustainable water management policies to meet water for human and environmental needs but can also add additional value in meeting the downstream environmental demand at desired periodicities.