Improvement of grand multi-model ensemble prediction skills for the coupled models of APCC/ENSEMBLES using a climate filter

Abstract

Twelve coupled model simulations of two multi-model ensemble (MME) systems for boreal winters from 1983 to 2005 are used to improve the climate prediction. From grading the relative capability of each simulation in reproducing the observed link between the tropical El Niño-Southern Oscillation (ENSO)-related Walker circulation and the Pacific rainfall, we find an optimal MME suite with improved prediction skills. This study demonstrates that the climate filter concept, proposed by us in a recent work, is not only useful in improving the MME prediction skills as compared to a single MME system, but also the skills of a grand MME that encompasses two well-performing MMEs. © 2013 Royal Meteorological Society.