Comparison of observational and CMIP5 Interannual Cloud Forcing Response in the Southeastern Pacific

The southeastern Pacific marine subsidence region is dominated by radiatively-important marine boundary layer clouds, which have been linked to the large cloud feedback uncertainties between general circulation models (GCMs). An analysis of the interannual response in cloud radiative forcing (CRF) to sea surface temperature (SST) in CMIP5 models is performed in the marine stratocumulus region of the southeastern Pacific and compared to satellite observations. Results suggest that there is a still a 15 Wm^-2 range in model-estimated CRF sensitivity to SST, although nearly half of the models are within the observational uncertainties. To evaluate the source of intermodel differences and model-observation differences, the interrannual sensitivity in low cloud amount and cloud liquid water path from A-Train satellite observations is compared to the latest CMIP5 estimates. Model and observational differences are linked to both the seasonal and regional distributions of clouds. Analysis of the seasonal cycle shows that models do not accurately capture the seasonal cycle in clouds and that differences are driven by large underestimates in clouds during the season of greatest observed cloud amount. Further analysis of the regional distribution of model clouds and their sensitivities shows that the nature of the CRF sensitivity underestimate varies between models; some models produce too few clouds and others producing compensating errors. Like CMIP3 models, CMIP5 models still underestimate the interannual CRF response to SST in marine boundary layer cloud regions.