Impact of improved representation of horizontal and vertical cloud structure in a climate model

Abstract

Many studies have investigated the effects that misrepresentation of sub-grid cloud structure can have on the radiation budget. In this study, we perform 20-year simulations of the current climate using an atmosphere-only version of the Met Office Unified Model to investigate the effects of cloud approximation on model climate. We apply the “Tripleclouds” scheme for representing horizontal cloud inhomogeneity and “exponential-random” overlap, both separately and in combination, in place of a traditional plane-parallel representation with maximum-random overlap, to the clouds within the radiation scheme. The resulting changes to both the radiation budget and other meteorological variables, averaged over the 20 years, are compared. The combined global effect of the parameterizations on top-of-atmosphere short-wave and long-wave radiation budget is less than 1 W m−2, but changes of up to 10 W m−2 are identified in marine stratocumulus regions. A cooling near the surface over the winter polar regions of up to 3°C is also identified when horizontal cloud inhomogeneity is represented, and a warming of similar magnitude is found when exponential-random overlap is implemented. Corresponding changes of the same sign are also found in zonally averaged temperature, with maximum changes in the upper tropical troposphere of up to 0.5°C. Changes in zonally averaged cloud fraction in this location were of opposite sign and up to 0.02. The individual effects on tropospheric temperature of improving the two components of cloud structure are of similar magnitudes to about 2% of the warming created by a quadrupling of carbon dioxide.