Toward a Cloud Analysis and Forecasting System Leveraging JEDI

Cloud forecasting via numerical weather prediction, using initial conditions from a cycling data-assimilation (DA) system, remains a challenge owing to both scientific and technical obstacles. To address the technical obstacles and provide a tool for advancing the science of cloud assimilation, NCAR and US Air Force, in collaboration with JCSDA and the Met Office, are developing a new, modular DA system that is not tied to a specific model and that facilitates sharing of DA infrastructure, components and expertise between the partners. The initial phase of the project aims to evaluate the Joint Effort for Data assimilation Integration (JEDI) as the platform for this DA system.

The evaluation is based on the implementation, using the Object Oriented Prediction System (OOPS) and other elements of JEDI, of prototype ensemble-variational assimilation capabilities for the Model for Prediction Across Scales (MPAS) and the new dynamical core in development at the Met Office. This presentation reports results from the MPAS/OOPS prototype, together with various contributions to the shared JEDI framework, including automated software testing, modeling for static background error covariances, implementation of additional radiative transfer algorithms in the Unified Forward Operator, and testing of ECMWF's Observational DataBase (ODB) format for observation ingest and feedback.