Studies of the Movement of Tropical Cyclones in the Southwest Pacific

Abstract

Observational data on the paths of tropical cyclones in the S.W. Pacific region, between 5°s and 30°s and between 145°E and 150°w, have been analysed by multiple regression techniques with a view to determining how accurately it is possible to predict the future displacement of a tropical cyclone using only elements from its past track and simple climatological parameters. It is found that the most important predictor, for both latitudinal and longitudinal motion, is the previous 12-hr displacement in latitude or longitude respectively. It is also shown that other predictors have some significance but that in general they do not contribute very much to the accuracy of practical prediction. In particular, terms depending on the cube of the latitude or longitude displacements appear to be of some interest as representing the retarding effect of the surrounding fluid on the motion of the vortex. Other terms are interpreted as the effects of Coriolis forces. Testing of the derived regression equations on data not used in their derivation shows that, although the accuracy of the predictions is not particularly high, they compare favourably with those obtained by methods currently used in the S.W. Pacific area. The study also contains some analyses of the motion of individual cyclones, where it is shown that rapid changes in environmental factors may be responsible for sudden changes in the direction of motion of a cyclone. It is concluded that a more satisfactory forecasting method can be obtained only by the explicit incorporation of external environmental factors into the prediction process.