Intermodel variances of subtropical stratocumulus environments simulated in CMIP5 models

Abstract

This paper investigates simulation biases associated with the large-scale environments of subtropical marine stratocumulus (Sc) in present climate simulations from the Coupled Model Intercomparison Project Phase 5 models. Comparison of eight major variables that strongly control the Sc clouds, including jumps of temperature and vapor across the inversion layer, indicates that these models all have similar shortcomings, such as overestimation of sensible and latent surface fluxes. The differences among the biases of these major variables between the Sc regions were also evaluated. Of all Sc regions, the Namibian Sc region showed the largest biases. The modeled simulation skill of the annual variation of low cloud cover (LCC) increases when the correlation coefficient between LCC and lower atmospheric stability is higher. This paper reemphasizes the importance, suggested in previous studies, of an improved relationship between lower tropospheric stability and LCC if we are to better predict annual variations in subtropical LCC. Key Points Detailed intermodel variability in marine stratocumulus environmentsHigh-geographical universality in skill of stratocumulus simulationImportant relationship between low cloud cover and stability of lower atmosphere