A coupled pbl-stratocumulus parameterization scheme is developed for application in weather prediction models. The parameterizations can be summarized in the following way. The long-wave radiation scheme is based on an emissivity approach. Turbulence is calculated with a turbulent kinetic energy equation-mixing length scheme. Turbulent tendencies of temperature, moisture and cloud water are needed when a prognostic cloud scheme is used to calculate the condensation and the microphysical processes. Therefore, turbulent diffusion of conservative variables will not be applicable. Instead, turbulent diffusion of potential temperature, specific humidity and cloud water are assumed to force the cloud towards a moistadiabatic state. The moistadiabatic values are derived from two conditions, a constant eqvivalent potential temperature and a constant total water mixing ratio (neglecting precipitation) with height. In this way, the use of a prognostic cloud scheme is enabled.
A simplified night-time stratocumulus case described by Moeng et al (1996) is studied with different resolutions (25m,100 m). To be able to compare our model results with Large Eddy Simulations (Moeng et al (1996)) and 1D-results from Bechthold et al (1996) the microphysical processes were turned off. Bechthold et al (1996) proposed two criterias in order to judge the model performance, a maximum bouyancy flux at the cloud top and a quasi-linear behaviour of the total water flux. In both runs, these criterias seem to be fulfilled but in the lower resolution run the maximum value of the total water flux is much smaller than in the high resolution run. This is believed to be caused by the slightly stable layer just at cloud base. The most likely reason for this feature is that the model can not represent the fine scale structure of the heating rate profile near the cloud base. Then, the turbulence and the heat transport from the cloud base are reduced and the stable stratification is maintained. Also, entrainment is smaller in the lower resolution run which is expected since it is a small scale process. Therefore, the boundary layer does not grow at all which also is seen in Bechthold et al (1996).
In conclusion, the use of a prognostic cloud scheme is enabled by treating the diffusion terms in the above suggested way. Furthermore, two model levels with cloud water seem to be enough to catch the most important physics as defined by Bechthold et al (1996). However, defficiences in the entrainment and the total water flux exist.