Wind Verification and a New Method for Forecasting Winds at Two Gap Locations in Southeast Wyoming

Southeast Wyoming is one of the windiest areas in the country during the months of October-March. The strong wind has significant impacts on commercial traffic especially along Interstates 80 and 25. Two wind-prone locations where blow-overs of high-profile vehicles frequently occur are the gap areas of Arlington on Interstate 80 and Bordeaux on Interstate 25. In this study, verification stats of short-term (12-21 hour) wind forecasts were computed at these two locations over the past wind season using BoiVerify. The results show that while SuperBlend had the lowest mean absolute error overall, our Official forecasts outperformed SuperBlend when the verifying winds are highest and impacts to transportation are greatest. SuperBlend is a model forecast internal to Central Region that includes CONSALL, the consensus of all direct model output and mos guidance, and also the bias correction on CONSALL (BCCONSALL).

Despite the improvement of our wind forecasts over the models during high-wind events, the current process of “serping” in higher winds at Arlington and Bordeaux is cumbersome. It is desired that a new, more systematic method of forecasting winds at these gap locations be developed that will save time without sacrificing forecast accuracy. The second part of this research involved developing an experimental “first-guess” wind forecast (hereafter CYSWIND) that could be incorporated into the Graphical Forecast Editor. CYSWIND is computed from a linear regression best fit on the CONSALL model. Since it is also directionally dependent, it will increase westerly winds by a greater magnitude, as this is the favored direction for the enhanced gap winds. Verification of CYSWIND on an out-of-sample period reveals its superior performance, beating SuperBlend and the Official forecasts overall, and more importantly at the higher winds. CYSWIND will be utilized at our office in the upcoming wind season to assist forecasters and track its operational performance.