A forecast system to predict wind power ramps was developed and applied to evaluate the similarity between sea level pressure (SLP) patterns in global-scale model forecasts and weather patterns associated with abrupt wind speed changes. The settings of the Weather Research and Forecasting model are selected according to the SLP pattern in global-scale model forecast. Wind forecast data are converted to power output, and a principal component analysis of SLP data is used to identify and forecast wind power ramps and to compute ramp probabilities and ramp alerts.