Statistical Prediction of Seasonal Mean Southern Hemisphere 500-hPa Geopotential Heights

Abstract A recently developed variance decomposition approach is applied to predict seasonal mean 500-hPa geopotential height anomalies in the Southern Hemisphere. In terms of predictability of both the winter and summer height fields, the Southern Oscillation and the Southern Annular Mode are identified as the first and second most important factors affecting the variability. Based on this study, a statistical prediction scheme has been developed. The linear trend in the leading empirical orthogonal function of the height field, the November Southern Annular Mode index, the austral spring Niño-3 index, and the November Coral Sea index are identified as the main predictors for the summer height field, while the March–May Southern Annular Mode index, the May Niño-4 index, and the austral autumn central Indian Ocean index are the main predictors for the winter height field. The predictive skill in forecasts of National Centers for Environmental Prediction–National Center for Atmospheric Research and European Centre for Medium-Range Weather Forecasts reanalysis 500-hPa geopotential height anomaly fields, in terms of a spatiotemporal anomaly correlation, is considerably higher than a single prediction achieved by a coupled general circulation seasonal forecast model.