The Evaluation of Several Cyclone Prediction Techniques

Abstract

A comparison is made of several schemes for making predictions of the location and intensity of cyclones. One scheme (due to K. W. Veigas and F. P. Ostby) is based on a linear model developed by a stepwise regression technique; another (due to J. J. George and P. M. Wolff) is based on an empirical study of linear and non-linear relations between selected meteorological variables. A third scheme, making use of climatological averages, was developed to use as a basis for determination of skill. The prediction methods are compared on the basis of operationally meaningful criteria. The climatological procedure gave consistent, reasonably accurate results over the entire Northern Hemisphere. The procedure of George and Wolff was superior to climatology throughout the hemisphere. No significant difference in accuracy was detected between the Wolff-George and the Veigas-Ostby procedures in the region for which Veigas-Ostby was developed (United States East Coast).