Regions of Significant Influence on Unforced Global Mean Surface Temperature Variability in Climate Models

Abstract:

Decadal prediction of global mean surface air temperature (GMT) change requires that Global Climate Models (GCMs) accurately simulate the spatial signature and magnitude of unforced (internal) modes of climate variability. In this work, we document the geographic regions that are most associated with unforced GMT variability in CMIP5 GCMs at both the subdecadal and interdecadal timescales. For this purpose we define Regions of Significant Influence (ROSIs) on GMT, which are locations that have a statistically significant correlation between local surface air temperature (SAT) and GMT (with a regression slope greater than 1), and where local SAT variation leads GMT variation in time. It is found that subdecadal timescale GMT variability is most associated with SAT variation over the eastern equatorial Pacific and is consistent across GCMs and between GCMs and observations. At the Interdecadal timescale, GMT variability also tends to be associated with SAT variation over the tropical eastern Pacific in most GCMs, but the precise pattern is far less consistent from GCM to GCM and several GCMs have a much stronger GMT association with SAT variability over the Southern Ocean.

We also compare the magnitude of GCM GMT variability to that seen in observations at both the subdecadal and interdecadal timescale and we investigate the relationship between the magnitude of this GMT variability and the underlying modes associated with the GMT variability. Results show that individual GCMs have large biases in the magnitude of subdecadal timescale GMT variability but there is no systematic bias. On the other hand, GCMs appear to underestimate the magnitude of interdecadal timescale GMT variability on average. GCMs with a stronger subdecadal relationship between SAT and GMT over the tropical Pacific tend to have more variable subdecadal GMT while GCMs with a stronger interdecadal relationship between SAT and GMT over the Southern Ocean tend to have more variable interdecadal GMT.