Improving HWRF Vortex Initialization Through Assimilation of Hurricane Inner-Core Observations with the GSI Hybrid Data Assimilation System

This study aims at improving the vortex initialization of the hurricane weather research and forecasting (HWRF) model in high-resolution nested domains through 1) enhanced background error terms that are consistent with the high-resolution nested domains and 2) assimilation of inner-core observations. The major objectives are to reduce the potential spin-up and spin-down problems for high-resolution nested domains to improve hurricane track and intensity forecasts. Specifically, we emphasize enhancing the capability of the NCEP operational GSI hybrid ensemble Kalman filter and variational data assimilation (GSI hybrid) system to initialize the high-resolution HWRF model in multilevel nested domains.

Our early results show that the use of high-resolution ensemble forecasts instead of the GFS ensemble forecasts in the GSI background term has positive impact on hurricane track and intensity forecasts as it mitigates the spin-down problem during the vortex initialization. Further experiments are conducted for better assimilation of the TC Vital data and the Tail Doppler radar (TDR) observations. Several strategies, such as enhanced TDR data quality control, the use of TC structure information, as well as tuning the data assimilation method are also being tested. Detailed results will be presented during the conference.