Authors: Ying Zhang Jieshun Zhu Zhongxian Li Haishan Chen Gang Zeng
Publish Date: 2016/04/08
Volume: 49, Issue: 3, Pages: 1049-1059
Abstract
This study examined the global sea surface temperature SST predictions by a socalled multipleocean analysis ensemble MAE initialization method which was applied in the National Centers for Environmental Prediction NCEP Climate Forecast System Version 2 CFSv2 Different from most operational climate prediction practices which are initialized by a specific ocean analysis system the MAE method is based on multiple ocean analyses In the paper the MAE method was first justified by analyzing the ocean temperature variability in four ocean analyses which all are/were applied for operational climate predictions either at the European Centre for Mediumrange Weather Forecasts or at NCEP It was found that these systems exhibit substantial uncertainties in estimating the ocean states especially at the deep layers Further a set of MAE hindcasts was conducted based on the four ocean analyses with CFSv2 starting from each April during 1982–2007 The MAE hindcasts were verified against a subset of hindcasts from the NCEP CFS Reanalysis and Reforecast CFSRR Project Comparisons suggested that MAE shows better SST predictions than CFSRR over most regions where ocean dynamics plays a vital role in SST evolutions such as the El Niño and Atlantic Niño regions Furthermore significant improvements were also found in summer precipitation predictions over the equatorial eastern Pacific and Atlantic oceans for which the local SST prediction improvements should be responsible The prediction improvements by MAE imply a problem for most current climate predictions which are based on a specific ocean analysis system That is their predictions would drift towards states biased by errors inherent in their ocean initialization system and thus have large prediction errors In contrast MAE arguably has an advantage by sampling such structural uncertainties and could efficiently cancel these errors out in their predictionsThis paper is a contribution to the special issue on Ocean estimation from an ensemble of global ocean reanalysesconsisting of papers from the Ocean Reanalyses Intercomparsion Project ORAIP coordinated by CLIVARGSOP and GODAE OceanView The special issue also contains specific studies using single reanalysis systemsThe work is partially supported by the Open Project of Key Laboratory of Meteorological Disaster of Ministry of Education KLME1404 the NSF of China under Grants 41575102 and 41575085 the NSF of Jiangsu Province China under Grant BK20131431 and the Project Funded by the Priority Academic Program Development PAPD of Jiangsu Higher Education Institutions Helpful comments from Drs B Huang and ZZ Hu are highly appreciated We also thank Dr M A Balmaseda from ECMWF for providing its ocean initial conditions
Keywords: