摘要
The first version of a global ocean reanalysis over multiple decades （1979-2008） has been completed by the National Marine Data and Information Service within the China Ocean Reanalysis （CORA） project. The global ocean model employed is based upon the ocean general circulation model of the Massachusetts Institute of Technology. A sequential data assimilation scheme within the framework of 3D variational （3DVar） analysis, called multi-grid 3DVar, is implemented in 3D space for retrieving multiple-scale observational information. Assimilated oceanic observations include sea level anomalies （SLAs） from multi-altimeters, sea surface temperatures （SSTs） from remote sensing satellites, and in-situ temperature/salinity profiles. Evaluation showed that compared to the model simulation, the annual mean heat content of the global reanalysis is significantly approaching that of World Ocean Atlas 2009 （WOA09） data. The quality of the global temperature climatology was found to be comparable with the product of Simple Ocean Data Assimilation （SODA）, and the major ENSO events were reconstructed. The global and Atlantic meridional overturning circulations showed some similarity as SODA, although significant differences were found to exist. The analysis of temperature and salinity in the current version has relatively larger errors at high latitudes and improvements are ongoing in an updated version. CORA was found to provide a simulation of the subsurface current in the equatorial Pacific with a correlation coefficient beyond about 0.6 compared with the Tropical Atmosphere Ocean （TAO） mooring data. The mean difference of SLAs between altimetry data and CORA was less than 0.1 m in most years.
The first version of a global ocean reanalysis over multiple decades （1979-2008） has been completed by the National Marine Data and Information Service within the China Ocean Reanalysis （CORA） project. The global ocean model employed is based upon the ocean general circulation model of the Massachusetts Institute of Technology. A sequential data assimilation scheme within the framework of 3D variational （3DVar） analysis, called multi-grid 3DVar, is implemented in 3D space for retrieving multiple-scale observational information. Assimilated oceanic observations include sea level anomalies （SLAs） from multi-altimeters, sea surface temperatures （SSTs） from remote sensing satellites, and in-situ temperature/salinity profiles. Evaluation showed that compared to the model simulation, the annual mean heat content of the global reanalysis is significantly approaching that of World Ocean Atlas 2009 （WOA09） data. The quality of the global temperature climatology was found to be comparable with the product of Simple Ocean Data Assimilation （SODA）, and the major ENSO events were reconstructed. The global and Atlantic meridional overturning circulations showed some similarity as SODA, although significant differences were found to exist. The analysis of temperature and salinity in the current version has relatively larger errors at high latitudes and improvements are ongoing in an updated version. CORA was found to provide a simulation of the subsurface current in the equatorial Pacific with a correlation coefficient beyond about 0.6 compared with the Tropical Atmosphere Ocean （TAO） mooring data. The mean difference of SLAs between altimetry data and CORA was less than 0.1 m in most years.
基金
the National Basic Research Program （Grant No. 2013 CB430304）
National Natural Science Foundation of China （Grant Nos. 41030854, 41106005, 41176003, and 41206178）
National High-Tcch R＆D Program of China （Grant No. 2013AA09A505）.
