CLOUD AND WATER-VAPOR MOTION VECTORS IN TROPICAL CYCLONE TRACK FORECASTING - A REVIEW

Tropical cyclone track prediction remains a vexing problem in meteorol ogy, particularly for numerical weather prediction. While there has be en significant improvement in forecast skill in recent years, errors i n prognosis, particularly for recurving cyclones still remain unaccept ably high. Consistent with track prediction being to a significant ext ent an initial value problem, there has been, in recent years, cogent evidence that, a combination of high resolution numerical modelling, t he use of appropriate assimilation techniques and the exploitation of high spatial and temporal resolution observations can improve the accu racy of tropical cyclone forecasts. Before landfall, tropical cyclones have their genesis and move over the data-sparse tropical oceans. Her e the prediction of their movement is an application for which remotel y sensed data are quintessential. In this context, this paper examines the increasingly important contribution of cloud and water vapour mot ion vectors to tropical cyclone prediction and evaluates their import to accurate prediction in terms of both the numerical modelling charac teristics and the data assimilation techniques employed. Overall, it i s shown that cloud and water vapour drift winds have made a significan t contribution to the tropical cyclone track forecasting problem when used with conventional intermittent assimilation techniques, such as 6 -hourly cycling, and, more recently, with continuous assimilation tech niques such as 3- and 4-dimensional variational assimilation. These co ntinuous assimilation schemes appear to have the potential to use near continuous asynoptic wind data in the most effective way.