Most GCMs simulate similar IPO-related cloud changes, with a strong east-west asymmetry in the Pacific. The high cloud cover and longwave cloud forcing changes associated with IPO could be understood through the variations of omega (large scale motion), and low cloud cover and net cloud forcing changes associated with IPO could be understood through the variations of sea surface temperature (low level stability). The consistency across models and the fundamental large-scale control make the evaluation of model biases possible. On the other hand, most of robust cloud changes associated with global warming exhibit a zonally symmetric structure, with the latitude of maximum and minimum total cloud fraction shifting poleward and the height of high cloud rising. The inter-model spread of global warming related cloud changes are larger and more complicated than those associated with IPO.
Our results suggest that the phase shift of IPO during late 1990s contributes significantly to the observed cloud trend during satellite era. However, one has to consider the anthropogenic signal to get the full picture understanding of the observed cloud changes.