Using a statistical representation of subgrid cloudiness to improve the Community Atmosphere Model

This study describes the implementation of a new parameterization for stratiform condensation and cloud fraction within version 5 of the Community Atmosphere Model (CAM5). At the core of this scheme is the assumption that subgrid-scale variations in liquid water potential temperature and the sum of water vapor and liquid follow a bivariate Gaussian distribution. The fact that subgrid variability is explicitly defined in this scheme is leveraged to enforce consistency between condensation, cloud fraction, radiation, and microphysical processes. Comparison against the stock version of CAM5 is made. Although differences between the two versions are small, the new scheme is an improvement because it allows better consistency with other components of the model and has a more physically-defensible basis.