AN EVALUATION OF 12–HR STATISTICAL FORECASTS OF THE 1000- TO 500-MB THICKNESS

Abstract

The usefulness of a number of variables as predictors of the 1000-500 mb thickness is explored. These predictors are utilized, in several combinations, to form multiple regression equations from which four series of 12-hour thickness forecasts are prepared. Verification scores are presented and compared with the scores of 12-hr numerical thickness predictions for the same period.