A global eddy-resolving reanalysis with the CESM2 ocean component

An ensemble optimal interpolation (EnOI) data assimilation system for the high-resolution (1/10˚) version of the ocean component of the Community Earth System Model version 2 (CESM2) is presented. As implemented within the Data Assimilation Research Testbed (DART) framework, the EnOI scheme uses a static (but seasonally varying) ensemble of pre-computed perturbations to approximate samples from the forecast error covariance and uses a single model integration to estimate the forecast mean. The EnOI scheme is used to assimilate satellite altimetry and sea surface temperature observations along with temperature and salinity in-situ observations into the ocean component of CESM2. Prior to being able to perform data assimilation at such eddy-resolving resolutions, the DART infrastructure has been reformulated to accommodate large model-states. The new data assimilation infrastructure makes large-state ensemble data assimilation possible by distributing state vector information across multiple processors on different MPI tasks. A global ocean retrospective analysis using this newly implemented EnOI-based system has been integrated for a 12-year period from 2005 to 2016. The EnOI is found to provide a practical and cost-effective alternative to the ensemble adjustment Kalman filter (EAKF) previously used for the assimilation of in-situ ocean observations into the nominal 1° ocean model.