Abstract
The role of the parameterization of vertical convection in calculating the climate impact of doubling CO2 is assessed using both one-dimensional radiative-convective vertical models and in the latitude-dependent Hadley-baroclinic model of Lindzen and Farrell (1980). Both the conventional 6.5 K km−1 and the moist-adiabat adjustments are compared with a physically-based, cumulus-type parameterization. The model with parameterized cumulus convection has much less sensitivity than the 6.5 K km−1 adjustment model at low latitudes, a result that can be to some extent imitated by the moist-adiabat adjustment model. However, when averaged over the globe, the use of the cumulus-type parameterization in a climate model reduces sensitivity only ∼34% relative to models using 6.5 K km−1 convective adjustment. Interestingly, the use of the cumulus-type parameterization appears to eliminate the possibility of a runaway greenhouse.
