Statistical Prediction of Tropical Cyclone Intensity Using Dynamical and Thermodynamical Inner-Core Parameters Derived from Hwrf Analysis and Forecasts

A new multiple linear regression model for short range tropical cyclone intensity prediction is developed. Four new dynamical and thermodynamical predictors based on HWRF output are considered: (1) the horizontal advection of relative angular momentum, (2) energy exchange from the divergent to the rotational kinetic energy (Psi-Chi interactions), (3) the conversion of shear vorticity to curvature vorticity, and (4) the vertical differential of heating in the complete potential vorticity equation. Predictors were calculated using Hurricane Research Weather and Forecast (HWRF) model initial fields. Each predictor was determined to exhibit a statistically significant relationship with 12 hour intensity change in tropical cyclones by an F-test. The predictors were then used as the basis for a multiple linear regression model, following the methodology of the operational Statistical Hurricane Intensity Prediction Scheme (SHIPS). Four additional predictors, intended to represent basic storm information and environmental conditions, were included in the development of a second model. Retrospective forecasts of hurricanes in 2004, 2005, and 2006 were created for both models, and compared to operational SHIPS and HWRF forecasts. Despite relying on HWRF fields for the calculation of predictors, the new model produces better forecasts than HWRF for short term (less than 48-hr) forecasts. Additional methods were developed to extend forecasts beyond 48 hours. This resulted in a systematic improvement of HWRF forecasts. It is proposed that the new model could be used operationally as a new version of the "early" HWRF.