Despite continued improvement of dynamical model intensity forecast skill (i.e. GFS, HWRF, HAFS, COAMPS-TC), statistical-dynamical models still provide useful intensity forecast guidance, especially for rapid intensification. The statistical-dynamical intensity forecasts are also used as input to the National Hurricane Center (NHC), Central Pacific Hurricane Center (CPHC), and Joint Typhoon Warning Center (JTWC) consensus models that usually produce overall the most skillful intensity forecasts.
This presentation will discuss the modified versions of the operational Statistical Hurricane Intensity Prediction Scheme (SHIPS), the Logistic Growth Equation Model (LGEM) models, and the SHIPS Rapid Intensification Index (RII) models that use modified metrics for ocean-TC interaction and machine learning. In the current operational versions of these models, ocean-TC interaction is represented by sea surface temperature (SST) and ocean heat content (OHC). In the new formulation, sea surface salinity (SSS) and dynamic depth-averaged temperature (Tdy) predictors derived from the Navy Coupled Ocean Data Assimilation (NCODA) dataset are added. Tdy takes into account the upper ocean temperature and salinity structure, static stability effects, and the storm's intensity, size, and translation speed, and has been shown to provide different and better measures of ocean-TC interaction compared to the currently used OHC. The new representation of ocean-TC interaction also allows for a coupled version of LGEM to be developed, where the SST cooling depends on the predicted intensity. Preliminary results indicate that the use of the new ocean-TC predictors could significantly improve forecast skill for SHIPS, LGEM, and SHIPS-RII. Independent verification of the modified models, comparison with some of the best-performing operational models, and real-time demonstrations to forecasters will be discussed.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner