There is growing evidence that patterns of atmospheric intra-seasonal and interannual variability have preferred states (i.e., local maxima in the probability density function of atmospheric variables in phase space) corresponding to circulation regimes. Using atmospheric data, Corti et al. (1999) showed four distinct circulation patterns in the wintertime Northern Hemisphere. These geographical patterns correspond to conventional patterns of low-frequency atmospheric variability which include the so-called "Cold Ocean Warm Land" (COWL; Section 7.6) pattern, the negative Pacific North American pattern, and the so-called negative Arctic Oscillation pattern.
Simplified dynamical models, which represent fundamental aspects of atmospheric circulation, react to external forcing initially by changes in the recurrence frequency of the patterns rather than by changes in the patterns themselves (Palmer, 1999). Therefore, anthropogenically forced climate changes may also be expressed in an altered pattern frequency. It appears that the observed northern hemispheric changes can be associated, to some extent, with a more frequent occurrence of the COWL pattern (Corti et al., 1999), i.e., the horizontal structure of recent climate change is correlated with the horizontal pattern of natural variability.
These results indicate that detailed predictions of anthropogenic climate change require models which can simulate accurately natural circulation patterns and their associated variability, even though the dominant time-scale of such variability may be much shorter than the climate change signal itself. This regime view of climate change has recently been shown in GCM simulations (Monahan et al., 2000).
Other reports in this collection