A COMPARISON OF THE ACCURACY OF OBJECTIVE TECHNIQUES FOR FORECASTING TYPHOON MOVEMENT DURING 1967

Abstract

In the past, many objective techniques have been evaluated for severe tropical cyclones in the Atlantic, but little has been done along this line in the Pacific. A computer program was developed at the Joint Typhoon Warning Center, to verify 10 separate 24-hr. forecast techniques. During the course of the 1967 season, a total of 14 different techniques were tested on 27 tropical cyclones. All techniques were adjusted to facilitate direct comparisons with the official Joint Typhoon Warning Center (JTWC) forecasts. Most methods were tested on an operational basis. All forecasts initially used operational positions with final verifications made against the tropical cyclone's best track. Although direct comparisons between the methods are difficult due to inhomogenity of sample size and differences in storms tested, all statistics and comparison figures are made against the official forecast using the same cyclones in the sample. After several storms, it became apparent that superior results were being obtained from computer steering predictions received from Fleet Numerical Weather Facility (FNWF), Monterey, Calif. After concentrating on these steering predictions and modifying them following a technique developed by Professor R. J. Renard, a final operational technique was developed called the Monterey 700-mb. A Modified. This forecasting method has shown far superior results to any other operational objective technique tested and is comparable in accuracy to the official JTWC forecast.