Extended Abstract (124K)17A.2
A simplified dynamical system for tropical cyclone intensity evolution
Mark DeMaria, NOAA/NESDIS, Ft. Collins, CO
Improvements in tropical cyclone intensity forecasting over the last few decades have lagged those of track prediction. Because of the complexity of the processes involved with intensity changes, simple statistical forecast models have remained competitive with three-dimensional coupled ocean-atmosphere prediction models. Further improvements in operational statistical forecast models are limited by the ability to represent the nonlinear intensity evolution by a system of linear regression equations. As an alternative to linear regression, a simple dynamical system for the evolution of the maximum sustained surface winds of tropical cyclones is developed. This system is based on a logistic growth equation (LGE) when the storm center is over water and an empirical inland wind decay model when the storm center is over land. The LGE model relaxes the solution towards the maximum potential intensity (MPI) when atmospheric conditions are favorable, or decays the solution to zero when the atmosphere is unfavorable. The MPI is estimated from the thermodynamic properties in the storm environment, and the relaxation rate is determined from the vertical wind shear and a measure of convective instability in the storm environment. The storm evolution is governed by a first order differential equation, which allows the adjoint model to be easily determined. The adjoint model is used to choose the free parameters in the model so that the prediction best matches observed intensity changes. The application of this dynamical system to operational intensity forecasting will be described, along with methods to incorporate advance satellite temperature and moisture retrievals in the prediction. The fit of the free parameters in the LGE model also provides insight into the roles of dynamic and thermodynamic processes in tropical cyclone intensity change.
Session 17A, Tropical Cyclone Intensity Change III: Statistical-Dynamical Forecast Methods
Friday, 2 May 2008, 8:00 AM-9:45 AM, Palms GF
Previous paper Next paper
Recorded presentation