By splitting a dataset of 41 high wind cases into strong and weak classes, the synoptic patterns were examined for distinguishing characteristics. Soundings were evaluated for specific features attributable to mountain wind events. This includes strong low-level cross-barrier winds, a mid-level temperature inversion, and a critical level aloft where the cross-barrier wind is zero. Sounding analysis also revealed a distinct wet class and dry class in the mid-level dew point. By further splitting the events, correlation statistics were calculated specific to each class.
After applying the statistics on these variables for each wind event, an empirical formula was created to predict the maximum wind gust in the given conditions. The result was an overall correlation of 0.85, and an average error of 4.4kt of 6.0%. In addition, both the observed wind and empirical values exhibit a distribution density that fits the large extreme value distribution within a 95% confidence interval, which indicates physical significance. These results are promising for useful application in forecasting the severity of impending wind events.