Valuing seasonal streamflow forecasts in power system operations

The value of seasonal streamflow forecasts for the hydropower industry has long been assessed by considering metrics related to hydropower production. However, this current approach overlooks the role played by hydropower dams within the power grid, therefore providing a myopic view of how forecasts could improve the operations of large-scale power systems. There are, in particular, two points worth stressing. First, the inherent uncertainty of streamflow forecasts could be easily propagated into the grid, especially if the power system is highly reliant on hydropower. Second, the relationship between water and power systems is not unidirectional: failing to capture feedback mechanisms may add uncertainty to exercises aimed at characterizing the value of seasonal forecasts. To fill in this gap, we developed a novel modelling framework that (i) hard-couples a reservoir system model with a power system model, and (ii) is subject to reservoir inflow forecasts with different levels of accuracy. We implement the framework on a real-world case study based on the Cambodian grid, which relies on hydropower, coal, oil, and imports from neighboring countries. In particular, we evaluate the performance of both systems in terms of power production costs, CO₂ emissions, and the amount of curtailed hydropower. Through this framework, we demonstrate that the value of streamflow forecasts is affected not only by their skill, but also by the dynamic behavior of the coupled water-power system.