| |
| | Improved Probability Forecasting of Weather |
 | | The statistical techniques are used to select the predictors from any available source, such as observations, numerical analyses and predictions, as well as other potential predictors (e.g., subjective forecasts of meteorological variables, road conditions or other weather-related variables), and to determine the optimal weights to assign to each predictor. |
 | | Our statistical ensemble methodology can be tailored to the prediction of any and all weather-related variables, including precipitation type and amount, maximum and minimum temperatures, ceiling and visibility, fog, road-surface conditions, and sea-surface conditions, to mention a few. |
 | | The mean probability predicted by the statistical/dynamical ensemble when the event occurred was significantly higher, and the mean probability predicted by the ensemble when the event did not occur was significantly lower, than that predicted by the other methods. |
| www.gfdi.fsu.edu /~pfeffer/improvfinal.html (3495 words) |
|