FV3-JEDI: Progress toward Interfacing the GFS and GEOS Weather Forecast Models with JEDI

The Joint Effort for Data assimilation Integration (JEDI) - led by the Joint Center for Satellite Data Assimilation (JCSDA) - is an inter-organizational endeavor to develop a common framework for performing data assimilation on model native grids. Over the past year or so much progress has been made towards interfacing JEDI to models based on the Finite Volume Cubed-Sphere (FV3) dynamical core. These models include the FV3 based Global Forecast System (GFS), soon to be used operationally by the National Weather Service, and the Goddard Earth Observing System (GEOS), used by NASA's Global Modeling and Assimilation Office (GMAO). We will provide details of the development strategies for the infrastructure needed to interface both of these models with JEDI and perform data assimilation. This includes all geometry, state and increment operations; static background error covariance modeling using the B matrix on Unstructured Mesh Package (BUMP); offline use of ensemble perturbations for hybrid background error covariance using BUMP; linearized model based on the tlm and adjoint developed by GMAO; parallel IO; interfacing to the Unified Forward Operator (UFO); interfacing to the CRTM and interfacing to the forecast models though ESMF. In this paper we will present some early cycled results using AMSUA, Radiosonde, Aircraft and GNSSRO observations using hybrid 4D-Var data assimilation and show some results from an offline hybrid 4DEnVar experiment. We will discuss the ongoing work and outline short- and long-term plans.

In addition to the model interfacing to perform data assimilation we have also begun implementing the Forecast Sensitivity Observation Impact (FSOI) tool into JEDI. We demonstrate the use of the FV3-JEDI system for computing observation impacts for both GEOS and GFS systems, with GSI still in use for the forward and adjoint assimilation steps.