Assimilating atmospheric observations into the ocean using strongly coupled ensemble data assimilation

Abstract

The local ensemble transform Kalman filter (LETKF) is used to develop a strongly coupled data assimilation (DA) system for an intermediate complexity ocean-atmosphere coupled model. Strongly coupled DA uses the cross-domain error covariance from a coupled-model background ensemble to allow observations in one domain to directly impact the state of the other domain during the analysis update. This method is compared to weakly coupled DA in which the coupled model is used for the background, but the cross-domain error covariance is not utilized. We perform an observing system simulation experiment with atmospheric observations only. Strongly coupled DA reduces the ocean analysis errors compared to weakly coupled DA, and the higher accuracy of the ocean also improves the atmosphere. The LETKF system design presented should allow for easy implementation of strongly coupled DA with other types of coupled models.