Improved Hydropower Generation Estimates for 1,500 Plants across the United States
Understanding the vulnerability of electricity supply to varying weather conditions requires sub-annual energy generation observations for hydroelectric power plants. In the United States, the Energy Information Administration (EIA) serves this data need by a regular survey of utilities and plant operators, known as form EIA-923. Researchers identified a misconception regarding these data: even though this form reports annual and monthly hydropower generation, roughly 90% of the power plants lack monthly observations, and the monthly data have been reconstructed based on market signals rather than water availability. To resolve the issue, they collected plant-specific reservoir releases and river flow data to serve as proxies for monthly generation, producing a new, improved inventory of historical monthly hydropower generation estimates across the United States.
People use the EIA-923 plant-level monthly hydropower generation data to train grid-scale hydropower simulations and supply seasonally varying hydropower energy budgets to multiple types of models. EIA-923 monthly data are also incorporated into other widely used datasets. Importantly, EIA-923 hydropower data have supported numerous energy and water resource studies, as well as retrospective analyses of regional generation portfolios. These inform market projections, policy analysis, and long-term planning. Many users are unaware of the limitations of these data, as it is not widely known that most EIA-923 monthly hydropower values are not actual observations. This research reveals deficiencies in the existing data while providing a revised dataset that addresses those limitations.
The EIA conducts a regular survey (form EIA-923) to collect annual and monthly net generation for more than ten thousand power plants in the United States. Approximately 90% of the ~1,500 hydroelectric plants included in this EIA data release are only surveyed at an annual resolution and lack monthly generation observations. For each of these plants, the EIA calculates monthly generation values using the combined monthly generating pattern of other hydropower plants within the corresponding census division. Their method neglects the local hydrology and reservoir operations, rendering the generated monthly data unsuitable for many research applications. Researchers devised an alternative approach to estimating monthly hydropower using generation proxies—namely historical monthly water reservoir releases and average river discharge rates recorded downstream of each dam. The researchers evaluated the new dataset and demonstrated its substantial improvement over current estimates for most plants. The new dataset—named RectifHyd—may be used to support power system studies or analyze within-year hydropower generation behavior at various spatial scales.