Cloud variability, radiative forcing and meridional temperature gradients in a general circulation model

Due to the non-linearity of cloud–radiation interaction in general circulation models (GCMs), the time-mean cloud radiative forcing (CRF) is in general different from the CRF of time-mean clouds. This implies that a change in temporal cloud variability induces a change in radiative forcing even if there is no change in time-mean cloud properties. Here we investigate this variability contribution to CRF quantitatively in the National Center for Atmospheric Research Community Climate Model 3.6 GCM. In a reference run, the variability contribution is found to account for 35% of the global-mean climatological CRF. The variability contribution peaks in the mid-latitudes and is shown to be driven by synoptic eddy activity. In a climate change experiment, where the atmospheric CO2 is quadrupled, the change in cloud variability offsets 40% of the change in CRF due to the change in mean clouds. It is found that almost all of this effect is due to variability in cloud fraction rather than in cloud water content, and it is traced to the non-linearity introduced by the model’s treatment of vertical cloud overlap. This study indicates the possibility of an eddy variability-climate feedback that has not been extensively studied and quantified in the past