Abstract
To reproduce the dynamical states of global climate on seasonal and inter-annual scales, a four-dimensional variational (4D-VAR) data assimilation system with a coupled ocean-atmosphere global model has been successfully developed in Data Research Center for Marine-Earth Sciences (DrC), Japan Agency for Marine-Earth Science and Technology (JAMSTEC). Our 4D-VAR coupled data assimilation (CDA) system with a three-month assimilation window runs routinely in every three months since January 2010 and generates sets of reanalysis data and ensemble three-year climate prediction data. To realize further efficient production of reanalysis data sets, DrC develops a production support system for synthesizing reanalysis data. It comprises the three subsystems for (a) automated observational data acquisition with a “cron” process, (b) quality control and data processing, and (c) synthetic visualization of observational and reanalysis data. With the support system, we evaluate performance of the sets of the three-month reanalysis and three-year ensemble prediction data. The majority of the data sets hold higher reproducibility in the three-month assimilation periods and predictability in the first one-year ensemble prediction periods over the tropical ocean states in the Pacific and Indian oceans. The support system, in particular, for synthetic visualization plays a crucial role to evaluate the reanalysis and prediction data.
