Statistical ensemble prediction of the tropical cyclone activity over the western North Pacific

Abstract

This paper presents a statistical model to forecast seasonal tropical cyclone activity. In order to give a comprehensive view of seasonal tropical cyclone activity, we include not only the number of total tropical cyclones but also the number of typhoons and the NTA (Normalized Typhoon Activity) index as the predictands. The model is based on a multiple linear regression model in which the final predictors are selected with respect to minimizing the prediction error rather than simply fitting with past data. The model is expanded into ensemble prediction by considering the uncertainty of the single and deterministic forecast. The probability forecast based on the ensemble model shows reasonably good skill with respect to reliability and relative operating characteristics.