Extended Abstract (92K)P1.36
Using Single-Station Statistical Weather Prediction Software to Instruct on Forecasting Methods
Patrick S. Market, University of Missouri, Columbia, MO; and B. Pettegrew
During the Spring 2003 semester, the Synoptic Meteorology II (SM2) course at the University of Missouri-Columbia (UMC) was asked to use statistical weather forecasting software as an additional tool to use in their daily weather forecasting routine for the UMC campus. The introduction of this software constituted a technological innovation of a kind, allowing students to see a broader range of numerical weather prediction techniques. Although they understood and appreciated this single-station approach as somewhat obsolete, the software was quite helpful in opening a dialogue on such topics as the origins of model output statistics and what information powers various forecasting techniques.
Those students who signed a consent form sat for a pre-test and a post-test on forecasting methods and their origins and completed a post-project survey. Of 17 students enrolled in SM2, 12 completed all of the aforementioned requirements. Prior to lecturing on forecasting methods, a 12-point pre-test was administered wherein students were surveyed on rudimentary aspects of forecasting methods. Their average was 6.6/12. After lecturing on forecasting methods, introducing the statistical weather forecasting software and allowing students to use the program and its product for six weeks, the subjects were retested. The post-test, identical to the pre-test, revealed an average score of 11.3/12, an increase of 4.7 points; the smallest score increase was 3 points, the largest was 7.
Among other findings from the post-project survey, students recognized the limitations of the statistical approach; none agreed that the approach alone was accurate; 10 of 12 strongly disagreed that the statistical approach had become their primary forecasting tool. yet, valuable lessons were learned: 10 of 12 recognized that the quality of a forecast was often the result of the of the quality of the initial data; 7 of 12 recognized this approach as a bridge between basic forecasting approaches (climatology and persistence) and sophisticated models (Eta, GFS).
Poster Session 1, Poster Session Educational initiatives (Hall 4AB)
Sunday, 11 January 2004, 5:00 PM-7:00 PM, Hall 4AB
Previous paper Next paper