Improving Operational Tropical Cyclone Intensity Guidance Suite and Situational Awareness with Better Metrics of Ocean-TC interaction

The National Hurricane Center (NHC), Central Pacific Hurricane Center (CPHC), and Joint Typhoon Warning Center (JTWC) use dynamical, statistical-dynamical, and consensus models as guidance for their tropical cyclone (TC) intensity forecasts. Despite dynamical models’ (i.e. GFS, HWRF) continuous skill improvement, statistical-dynamical models still significantly contribute to intensity forecasting, especially for rapid intensification, and are also used in the NHC consensus models that produce overall the most skillful intensity forecasts. 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) are under development. In the updated models, we are evaluating model improvements with better representation of TC-ocean interaction and the use of machine learning methods to account for nonlinear relationships between future intensity changes and predictors, which have been suggested by multiple studies.

The major component in all changes is the modified ocean-TC interaction. The ocean response to wind forcing plays an important role in TC intensity change, especially for slow-moving storms, where mixing and upwelling processes lead to sea surface temperature (SST) cooling that acts to reduce intensification. In the updated models, the ocean-TC interaction is represented by the dynamic depth-averaged temperature (Tdy) and sea surface salinity (SSS) predictors. 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 be a better metric of ocean-TC interaction compared to the currently used ocean heat content (OHC). Sea surface salinity (SSS), which is not used in operational statistical-dynamical models, was recently shown to be an important predictor for TC rapid intensification. The full ocean temperature and salinity profiles from the Navy Coupled Ocean Data Assimilation (NCODA) dataset have been added to SHIPS diagnostic files, which allows us to investigate the use of multiple ocean predictors. Preliminary results indicate that SSS and several versions of Tdy are more significant (based on a standard F-test for determining predictor significance) than many predictors used in the current operational SHIPS and LGEM models, and that the addition of these predictors could significantly improve forecast skills for SHIPS, LGEM, and SHIPS-RII.

In addition, to improve forecaster situational awareness, graphical displays of the NCODA-based SST, OHC, Tdy, SSS, and SST cooling have been developed and are available in near real-time on RAMMB-CIRA SLIDER (https://rammb-slider.cira.colostate.edu) and the internal CIRA AWIPS2 system. These products are being generated in real-time at CIRA in preparation for real-time demonstrations using NHC’s AWIPS2, and are currently provided in real-time to the Joint Typhoon Warning Center (JTWC) for demonstration and evaluation via AWIPS2. Preliminary results on the use of the new products in statistical-dynamical TC intensity forecast models and real-time demonstrations to forecasters will be discussed.