Evaluation of the impact of satellite radiance data within the hybrid variational/EnKF Rapid Refresh data assimilation systems

On February 2014, the NOAA operational Rapid Refresh (RAP) hourly updated prediction system running at the National Centers for Environmental Prediction (NCEP) has been upgraded to version 2. The data assimilation system is now the advanced Gridpoint Statistical Interpolation (GSI) hybrid variational/Ensemble Kalman Filter (EnKF) data assimilation system. Ensemble information needed in the regional assimilation comes from the 80-member global ensemble data assimilation system. Overall, the use of the hybrid variation/ensemble assimilation procedure has produced significant forecast improvement compared to the 3D-VAR assimilation method.

Our previous studies had demonstrated that the assimilation of satellite radiance observations produced positive impact on short-range forecasts within the RAP model system. In preparation for operational implementation of RAP version 3 planned for early 2015, a series of new updates for satellite radiance data assimilation are being tested at the Global System Division (GSD) of the NOAA Earth System Research Laboratory (ESRL). These updates include the use of enhanced bias correction scheme, revised channel selection for existing sensors, assimilating observations from new instruments, and the use of the Regional ATOVS Retransmission Services (RARS) real-time data sets.

Two RAP runs, with and without radiance data have been setup at GSD for the real-time radiance data impact study. The real-time run with radiance data assimilates conventional data and all radiance data planned for the RAP version 3. Radiance data assimilated include observations made by AMSU-A, HIRS, MHS, AIRS, and GOES as well as the real-time RARS data with short data latency. The real-time run without radiance data assimilates only conventional data. Under a more controlled retrospective testing environment, the impact from satellite radiance data sets is also being examined. The impact from assimilating radiance data in the RAP system is evaluated in terms of short-term forecast skill. In this presentation, we will report on the recent radiance assimilation updates as well as results obtained from the retrospective experiments and the multi-month experiments performed in real time environment.