Tropical Cyclone Predictability

The scientific objectives of the effort are to apply several techniques developed within the program, including Monte Carlo ensemble approaches, techniques adapted from non-linear systems analysis and optimal combinations of forecasting approaches to estimating the intrinsic limits to predictability for (i) tropical cyclone mean forecast position errors in the first instance, and (ii) to tropical cyclone intensity in the second part of the program. These intrinsic limits exist because the equations governing the behavior of all atmospheric systems including tropical cyclones are deterministically chaotic. Thus errors in the initial conditions and model formulation lead to error growth that eventually reduces the skill of the forecasts to zero. The intrinsic limits are to be compared with the results being obtained in practice and the size of the disparity represents the gains in predictive skill that are still achievable. It is of fundamental importance to have some idea of how large the gap is between that being obtained and the ultimately achievable in order to justify the continued allocation of resources to the various problems. A substantial part of the research objectives is developing a data assimilation and prediction scheme that can produce improved forecasts of tropical cyclone intensity.