ABSTRACT

Researchers have been estimating the long-term variability of wind power since the 1940s, when large-scale wind power was first considered in detail. Thomas, an early proponent of large-scale wind power, examined 25 years of wind speed data from seven weather stations around the United States and estimated that 'the annual average velocity for any one year seldom drops more than 10 or 15 percent below the long-term average velocity'. First, we use the Climate Forecast System (CFS) reanalysis, which has higher spatial and temporal resolution than most of the reanalyses used in previous papers. Second, we analyze locations of existing wind plants in the United States. The variability of annual energy generation for individual wind plants is larger than the variability of their aggregate. Third, we more accurately estimate hub-height wind speeds by accounting for atmospheric stability and adjusting model outputs based on comparisons with measured data.