Replication of atmospheric oscillations, and their patterns, in predictors derived from Atmosphere-Ocean Global Climate Model output

Abstract

Atmosphere–Ocean Global Climate Model (AOGCM) output is used for many climate change impact studies and to produce ‘predictor’ data sets for statistical downscaling methods. Quantitative and qualitative evaluation and validation are required to make informed choices concerning reliable variables and their optimum combinations for both forms of research. Previous study suggests that although mean sea-level pressure is generally well represented in models, biases associated with over- or underestimated activity for the Pacific Decadal Oscillation and the El Nino Southern Oscillation may exist within certain AOGCMs. This potential bias in indices of large-scale atmospheric variability is explored. Improvements in the replication of circulation indices are discovered between the second and third generations of the Canadian AOGCM (CGCM2 and CGCM3). With respect to reanalysis product, CGCM3 output shows less winter-time bias for the Northern Annular Mode and the North Atlantic Index than evident for the other indices under consideration. Copyright © 2010 Royal Meteorological Society