<p>Given the wide range of institutional and market contexts in which hydroelectric dams are operated, determining the value added from improvements in hydrologic forecasts is a challenge. Many previous examples of hydrologic forecasts being used to optimize hydropower production strategies at dams focus on a single reservoir system or watershed, with a key assumption that the marginal value of hydropower production is exogenously-defined (dams are &#8216;price takers&#8217; in markets for electricity that exhibit no market power). In some cases, this may accurately reflect current institutional boundaries and decision making processes. However, with increased attention being paid to how more coordinated grid management strategies, including management of hydropower assets, could facilitate deep integration of renewable energy, it is critical to understand how the use of improved hydrologic forecasts could produce wider grid-scale benefits, including &#160;lower costs and emissions. In this study, we quantify the value of streamflow forecasts to a centralized power system operator in charge of coordinating sub-weekly operations of hydropower assets, using the Western U.S. as a case study. We propagate flow forecasts through realistic models of reservoir operations and models of bulk power systems/wholesale electricity markets. Our results shed light on how the value of flow forecasts to grid operations can vary across regions and power systems. They also highlight the potential for conflicts between firm-specific objectives (profit maximization) and system-wide objectives (minimization of costs and emissions) when determining value added from hydrologic forecasts. &#160;</p>
Market segmentation for profit maximization using machine learning algorithms
