Three equilibrium doubled CO2 experiments have been performed using the Bureau of Meteorology Research Centre atmospheric general circulation model. These experiments used identical versions of the model, apart from changes in the convective parameterization and the horizontal resolution. The penetrative convection parameterizations used were variants of the Tiedtke (1989) mass flux and Kuo (1974) schemes. The shallow convection was also varied in strength. In the control climate the mass flux scheme produces a warmer, moister troposphere, with substantially more high cloud than the Kuo scheme. The precipitation distributions agree reasonably with observations overall, although the mass flux scheme gives improvements in some regions.

The increase in resolution is generally found to have a smaller impact upon the climate than the change in convection. Under a doubling of atmospheric CO2, the equilibrium responses of all three experiments were extremely similar in surface and tropospheric temperatures and humidity changes. The model response is at the low end of the scale of simulated climate change, consistent with a strong negative feedback found due to clouds. This feedback is similar to that found in earlier fixed season experiments. It appears to be insensitive to differences in cloud cover simulated in the control climates of the present experiments.