Interdecadal Trend of Prediction Skill in an Ensemble AMIP-Type Experiment

Abstract

This study addresses the interdecadal trend in potential skill score as estimated from the 500-hPa height temporal correlation coefficient (TCC), based on a 50-yr 10-member ensemble GCM integration with observed SST. The skill scores are based on the perfect model assumption, in which one of the members of the ensemble is assumed to be true. A distinct decadal positive trend in the TCC in boreal winter (December?January?February) was found. This trend is shown to be consistent with the positive trend in the interdecadal time-scale temporal variance of SST. The geographical pattern of the differences of the TCC between each decade and the 50-yr period resembles the Matsuno?Gill pattern, suggesting that the increase in the TCC is due to the Rossby wave excitation induced by the anomalous diabatic heating caused by the anomalous SST. Similar interdecadal trends in the variance of the Southern Oscillation index and Pacific?North American pattern were found in both the observation and the simulation. The interdecadal trend in the variance of 500-hPa geopotential height over the continental United States, however, existed only in the simulation.