The use of HWRF ensemble versus global ensemble in GSI hybrid analysis for inner-core data assimilation

A hybrid ensemble-variational data assimilation system has been implemented for operational HWRF since 2013 hurricane season. This system uses the global ensemble from the GFS hybrid EnKF/Var system in GSI hybrid analysis. The background error covariances estimated from the global ensemble capture the TC vortex features better than static background error covariance used in 3DVAR scheme. Although, this hybrid method helped improve inner-core data assimilation for some cases, inner-core observations are still not optimally assimilated into operational HWRF. The global ensemble is not sufficient in representing the forecast errors of the HWRF model, which runs at a much higher resolution. A hybrid data assimilation system that uses HWRF generated ensemble to estimate background error covariance is being developed. The impact of using of the HWRF ensemble vs. the globel ensemble in hybrid analysis for inner-core data assimilation will be presented at the conference.