On the Use of Time-Offset Model Output Statistics for Production of Surface Wind Forecasts

Abstract

An operational model output statistics (MOS) surface wind forecasting procedure has been implemented recently for 159 Canadian stations. Though the system has much in common with MOS wind forecast procedures in the United States, there are significant differences. One major difference is that the Canadian equations are not always used for the intended projections. Specifically, a time-offset MOS, or TOMOS, method is used to generate forecasts at times intermediate to those for which the equations were developed, and the TOMOS method is used in place of separate back-up equations in situations where station observations are required as predictors but are missing at model run time. This paper describes the TOMOS method and the tests that were conducted on its use prior to implementation. The results show that the TOMOS method is not only useful for the applications just mentioned, but it also has potential for production of medium range and possible long-range forecasts with little or no deterioration of forecast accuracy. The TOMOS method could be used to substantially reduce development costs for MOS wind forecast systems.