Development and research of GSI-based hybrid data assimilation system for HWRF: Performance of dual resolution and single resolution hybrid systems with moving nests assimilating airborne observations

GSI-based hybrid data assimilation (DA) system is further extended with the operational regional Hurricane Weather Research and Forecast (HWRF) modeling system to improve high-resolution convection-allowing tropical cyclone prediction. A two-way dual-resolution and single-resolution hybrid system for HWRF with moving nests were developed. In the dual-resolution configuration, the control hybrid analysis and forecast were conducted at a higher resolution (3km) than the EnKF analyses and forecasts (9km). During the data assimilation cycling, the 3km domain is movable following the TC center. In the single-resolution configuration, the 3km movable control analysis ingests 3km movable EnKF ensemble rather than the 9km EnKF ensemble. The performance of the newly developed systems were studied by conducting experiments with retrospective hurricane cases and by running experiments in near-real time for 2014 hurricane season assimilating airborne observations. Retrospective experiments revealed that the new dual-resolution two-way hybrid data assimilation system performed better than the baseline 9km single-resolution hybrid DA system. Series of specifically designed experiments were conducted to isolate the impact of the differences of the DA systems on the forecasts. It was found for maximum wind and minimum sea level pressure forecasts, such improvements were due to both (1) running the 3/9km moving nest during free forecasts and (2) dual-resolution analysis during the DA cycles. For track forecasts, such improvements are due mostly to running the 3/9km moving nest during free forecasts. Tests of the 3km single-resolution hybrid with movable nests also demonstrated promising results. Performance of the newly developed system during 2014 hurricane season will be presented at the conference also.