The interplay between reservoir storage and operating rules under evolving conditions

Reservoir storage helps manage hydrological variability, increasing predictability and productivity of water supply. However, there are inevitable tradeoffs, with control of high frequency variability coming at the expense of robustness to low frequency variability. Tightly controlling variability can reduce incentives to maintain adaptive capacity needed during events that exceed design thresholds. With multiple dimensions of change projected for many water supply systems globally, increased knowledge on the role of design and operational choices in balancing short-term control and long-term adaptability is needed. Here we investigated how the scale of reservoir storage (relative to demands and streamflow variability) and reservoir operating rules interact to mitigate shortage risk under changing supplies and/or demands. To address these questions, we examined three water supply systems that have faced changing conditions: the Colorado River in the Western United States, the Melbourne Water Supply System in Southeastern Australia, and the Western Cape Water Supply System in South Africa. Moreover, we parameterize a sociohydrological model of reservoir dynamics using time series from the three case studies above. We then used the model to explore the impacts of storage and operational rules. We found that larger storage volumes lead to a greater time before the shortage is observed, but that this time is not consistently used for adaptation. Additionally, our modeling results show that operating rules that trigger withdrawal decreases sooner tend to increase the probability of an adaptive response; the findings from this model are bolstered by the three case studies. While there are many factors influencing the response to water stress, our results demonstrate the importance of: i) evaluating design and operational choices in concert, and ii) examining the role of information salience in adapting water supply systems to changing conditions.