| description abstract | The potential utility of forecast information to improve reservoir control is well established, but physical, operational, and institutional constraints inhibit the full realization of these benefits. To help integrate forecast technology into operational water management, it is necessary to establish the opportunities and limitations of forecast use within the current system management framework, so that water managers and planners can identify in practice what constraints or aspects of control policy design most limit forecast utility. In this work, we explore the value of forecasts in the current outflow management plan of the Lake Ontario-St. Lawrence River basin, which is the first regulation policy of this system to use water supply forecasts to guide release decisions. We assess the value of forecasts for many-objective system management (e.g., upstream and downstream flood risk reduction, hydropower production, commercial navigation costs, recreational boating benefits, and wetland restoration), as a function of forecast accuracy, timing, and interactive effects, with system constraints and control rule structures. Results show that a 20%–30% reduction in forecast error over current statistical forecasts leads to observable improvements in most objectives, which grow significantly with additional forecast accuracy. Realizing these benefits is possible by increasing forecast accuracy, particularly in the springtime when tradeoffs between objectives are the greatest. However, we also identify one objective (wetland restoration) that declines with improved forecast skill, as well as operational constraints that limit forecast utility under the current outflow management plan. Overall, these results suggest that key aspects of the current control policy could be revised to better utilize forecast information, especially as forecasts improve over time. | |
