Tropical cyclone track and intensity prediction: The generation and assimilation of high-density, satellite-derived data

The impact of recent scientific and technological advances in tropical cyclone track, intensity and structure modeling is discussed. Since the early 1990s, developments have occurred in remote sensing, data assimilation procedures, numerical models and high performance computing. In particular, there is now quasi-continuous high spatial and temporal resolution data coverage over the previously data-sparse oceans where tropical cyclones spend most of their life cycles. There has been a rapid development of data assimilation methodologies capable of using these data to initialize high-resolution prediction models. Model developments have reached a stage of maturity where the representation of many of the physical processes necessary for improved tropical cyclone track and intensity prediction are now included. Finally, available computer power has reached the teraflop range. Most operational centers have high performance computers capable of tropical cyclone modeling at resolutions necessary for skillful track and intensity simulations. This article focuses on combining all of the above developments in a tropical cyclone data analysis and prediction system. The system has produced statistically significant reductions in the mean forecast error statistics for tropical cyclone track predictions and resulted in far more realistic simulations of tropical cyclone intensity and structure. A large number of tropical cyclones have been modeled, with emphasis on those classified as being "difficult" storms to predict accurately. These difficult storms are most responsible for rapidly growing forecast errors. Our results are illustrated by case studies of such tropical cyclones. [PUBLICATION ABSTRACT]