Leading modes of Asian–Australian monsoon interannual variability as represented in CMIP5 models

The Asian–Australian monsoon (AAM) exerts a strong impact on the regional climate. The ability of climate models to simulate the AAM is of great importance. This study evaluates the performance of 26 models, which participate in the Coupled Model Intercomparison Project phase 5 (CMIP5), in representing features of the leading modes of the AAM interannual variability. This is achieved by diagnosing quantitative metrics derived from a Season-reliant Empirical Orthogonal Function (S-EOF) analysis, and plausible external forcing contributing to the simulation results are presented. Our results show that the CMIP5 models can generally capture the spatial patterns of the leading modes of AAM interannual variability during El Niño development periods because of strong air–sea interactions. The variance percentage explained by the first two S-EOF modes is found to be closely related to the amplitude of the El Niño-Southern Oscillation (ENSO) in the models. In contrast, simulation of the periodicity remains challenging, and the models show deficiencies in the accurate capture of periodic characteristics. Models with realistic ENSO simulations typically perform well in simulating S-EOF1, and the key to improving the simulation of S-EOF2 is the greater intensity of the simulated field of background heating over the Northwest Pacific warm pool. The multimodel ensemble (MME) performs better than most of the 26 individual models. Models in the MME with better ENSO simulation and stronger warm pool heating are found to be better able to represent the relationships between the first two leading S-EOF modes and ENSO, although biases remained.