Statistical prediction of tropical storm motion over the Bay of Bengal and Arabian Sea

Statistical analog models for the prediction of tropical cyclone motion have been developed for all of the world's tropical cyclone basins. The popularity and success of these models is related to their ability to optimize the variance reducing potential of climatology and persistence, At the same time, they effectively recognize certain repetitious synoptic patterns without having to use fields of sometimes unavailable and often unreliable synoptic data over the tropics. The analog method can be simulated by the use of continuous function derived through the use of multivariate regression analysis. This method simulates all aspects of the analog process including the derivation of probability ellipses while avoiding certain operational problems inherent to the purely analog approach. This paper describes the derivation of such a model for the north Indian Ocean tropical cyclone basin. It is patterned after an earlier paper (Neumann and Randrianarison 1976) which describes the development of a similar model for the southwest Indian Ocean tropical cyclone basin.