3-D simulation of clouds with subgrid fluctuations of temperature and humidity

A new scheme for simulating clouds, which considers subgrid temperature and humidity fluctuations, is proposed. This is done by using local estimates of first and second moments of the joint specific humidity–potential temperature-distribution in a Monte-Carlo-integration of the relevant processes. Simulation of cirrus and low level cloudiness has been done within the frame of EUCREX (European Cloud and Radiation Experiment). With the new scheme, a part of the given grid volume can be supersaturated and condensation or ice formation occurs without a supersaturation of the entire volume to be realized. This produces more frequent and earlier cloud water (ql) and cloud ice (qi) and results in an increase in time-average cloud activity. The domain averaged values ql and qi using the new scheme remain larger most of the time and are appreciable after the time they are approximately zero when no subgrid fluctuations are considered. The assumption that local condensation can occur in the presence of turbulent fluctuations, even though the average state is below saturation, is part of a general effort to improve sub-scale cloud modeling. It appears to be necessary for a model which uses a larger grid volume and therefore needs more accurate predictions for the subgrid turbulence.