A Three-Dimensional Variational Analysis Using Sequential Filter

A three-dimensional variational (3D-Var) data assimilation (DA) scheme which introduces sequential filter is designed to improve state estimation from observations. The sequential filter employed in this new hybrid scheme, operating on the background field instead of on the observation field as usual, is under the framework of ensemble adjustment Kalman filter (EAKF). Under perfect global barotropic spectral model, numerical experiments show that the new 3D-Var can retrieve structures of true field appropriately and outperforms traditional 3D-Var.