Ensemble data assimilation in wind industry

Wind industry has an increasing demand for high accuracy forecasts of wind and other atmospheric variables. It also requires highly reliable forecast uncertainty estimates in order to determine upper and lower error bounds, which are critically important for power scheduling. Ensemble data assimilation methods, based on Kalman filter and optimal control theory, are perfectly positioned to address such industry needs, especially because of their ability to provide flow-dependent forecast uncertainty estimates, thus flow-dependent lower and upper error bounds. In this presentation, we focus on the impact of ensemble data assimilation on the short-range (0-6 hours) wind power forecast. Performance measures of special importance for this application, such as metrics of significant ramp events, are discussed and evaluated in more detail.