| description abstract | A recent scheme to predict tropical cyclone (TC) activity over the western North Pacific partially failed in 1997 and 1998, during which a warm and a cold event of the El Niño?Southern Oscillation (ENSO) occurred, respectively. This paper presents results of two approaches to improve on such predictions. The first is to include new predictors that are related to ENSO based on some recent research, and the second is to provide an updated prediction by incorporating monthly values of predictors in April and May of the current year. The results suggest that new predictors related to ENSO can indeed be identified, which include temporal changes in the Southern Oscillation index, strength of the Australian monsoon, and intensity of the subtropical high in the South Pacific. These predictors, together with those selected from the original prediction scheme, are combined to form a modified scheme that in general gives better forecasts of TC activity. The updated scheme that includes April and May predictors further improves the accuracy of the predictions. Real-time predictions from both schemes for the year 2000, which were made in April and June, are found to be largely accurate. Both schemes show better skill compared with the original one. | |
