Hybrid 4D EnVar for the NCEP GFS: Implementation plans and potential for outer loops

Following the successful implementation of hybrid 3D EnVar into NCEP operations for the Global Forecast System (GFS) and Global Data Assimilation System (GDAS), significant progress has been made toward the 4D extension of the algorithm. The hybrid 4D EnVar algorithm has a few attractive qualities for an operational center relative to 4DVAR, most notably the lack of need for tangent linear/adjoint models as well as reduced computational cost. NCEP is planning to implement the hybrid 4D EnVar into operations for the GDAS/GFS sometime in late 2015 or early 2016. Preliminary results at reduced resolution have been extremely encouraging. The results from various sensitivity experiments will be presented.

Beyond the initial implementation of hybrid 4D EnVar, various enhancements are still in their early phase such as improved data selection, the role of the static error covariance, and outer loop configuration. Analogous to the outer loop configuration in 4DVAR and so-called Quasi-Outer Loop of the EnKF, an outer loop can be run as part of 4D EnVar whereby the non-linear model is re-run using the increment at the beginning of the assimilation window. This may prove to be particularly important for dual-resolution configurations that use low-resolution ensemble members within high-resolution deterministic assimilation. Results from toy model experiments will be presented to demonstrate the potential impact of such a configuration within the 4D EnVar paradigm.