The vertical resolution sensitivity of simulated equilibrium temperature and water-vapour profiles

Variability of atmospheric water vapour is the most important climate feedb ack in present climate models. Thus, it is of crucial importance to underst and the sensitivity of water vapour to model attributes, such as physical p arametrizations and resolution. Here we attempt to determine the minimum ve rtical resolution necessary for accurate prediction of water vapour. To address this issue, we have run two single-column models to tropical rad iative-convective equilibrium states and have examined the sensitivity of t he equilibrium profiles to vertical resolution. Both column models produce reasonable equilibrium states of temperature and moisture. Convergence of the profiles was achieved in both models using a uniform ver tical resolution of around 25 hPa. Coarser resolution leads to significant errors in both the water vapour and temperature profiles, with a resolution of 100 hPa proving completely inadequate. However, fixing the boundary-lay er resolution and altering only the free-tropospheric resolution significan tly reduces sensitivity to vertical resolution in one of the column models, in both water and temperature, highlighting the importance of resolving bo undary-layer processes. Additional experiments show that the height of the simulated tropopause is sensitive to upper-tropospheric vertical resolution . At resolutions higher than 33 hPa, one of the models developed a high degre e of vertical structure in the vapour profile, resulting directly from the complex array of microphysical processes included in the stratiform cloud p arametrization, some of which were only resolved at high resolutions. This structure was completely absent at lower resolutions, casting some doubt on the approach of using relatively complicated cloud schemes at low vertical resolutions.