Sensitivity of the Thermal Balance in a General Circulation Model to a Parameterization for Cumulus Convection with Radiatively Interactive Clouds

Abstract

The dependence of the thermal balance of a general circulation model on the parameterization of cumulus convection is investigated. Incorporation of a Kuo-type cumulus parameterization into the NCAR community climate model decreases temperatures in most of the lower and middle tropospheres while increasing temperatures slightly at the tropopause, decreases both relative and specific humidities in large parts of the lower troposphere, and also reduces cloud cover and tropical precipitation. Although the Kuo parameterization represents a vertically integrated heat source, its presence in the general circulation model causes an even larger reduction in heating by the moist adiabatic adjustment, so the total heating associated with cumulus convection is less if the Kuo parameterization is used. The reduction in atmospheric temperatures relative to those at the surface with the Kuo parameterization results in enhanced heating of the lower atmosphere by surface exchange processes. These changes in convective and surface heating dominate changes in the diabatic part of the thermal balance but are moderated by changes in radiative heating associated with reduced cloudiness. Diabatic heating changes are balanced primarily by reduced mean dynamic transport of heat associated with a weakened Hadley circulation. The dependence of the circulation sensitivity to cumulus parameterization on cloud-convection feedback and the penetrative extent of convection is found to be significant. The penetrative depth of convection is especially important; since penetration by convection depends crucially on the poorly understood entrainment process, some uncertainty plagues estimates of the details of the impact of cumulus convection on the simulated general circulation. Changes in cloudiness associated with the Kuo parameterization alter radiative forcing so as to reduce the sensitivity of the community climate model to the cumulus parameterization.