Abstract
Evaluating the climate hindcasts for 1982–2009 from the NCEP CFS Reanalysis and Reforecast (CFSRR) project using the Climate Forecast System, version 2 (CFSv2), this study identifies substantial areas of high prediction skill of the sea surface temperature (SST) in the South Pacific. The skill is the highest in the extratropical oceans on seasonal-to-interannual time scales, and it is only slightly lower than that for the El Niño–Southern Oscillation (ENSO). Two regions with the highest prediction skills in the South Pacific in both the CFSv2 and persistence hindcasts coincide with the active centers of opposite signs in the South Pacific Ocean dipole (SPOD) mode, a seesaw between the subtropical and extratropical SST in the South Pacific with a strong phase locking to austral summer. Interestingly, the CFSv2 prediction exhibits skillful predictions made three seasons ahead, more superior to the persistence forecast, suggesting significant dynamical predictability of the SPOD. An austral “spring predictability barrier” is noted in both the dynamical and persistence hindcasts. An analysis of the observational and model data suggests that the SPOD mode is significantly associated with ENSO, as an oceanic response to the atmospheric planetary wave trains forced by the anomalous atmospheric heating in the western Pacific. Although previous studies have demonstrated that the pattern of subtropical SST dipole is ubiquitous in the Southern Ocean, the SPOD has been least known and studied, compared with its counterparts in the south Indian and Atlantic Oceans. Since the SPOD is the most predictable oceanic mode in the whole Southern Hemisphere, its climate effects for local and remote regions should be further studied.
