Abstract
The multiple mass-flux approach introduced in a companion paper aims at parameterizing the surface-generated convection in the atmosphere. The original aspects of this model are as follows. A number of convective updrafts are released in the surface layer. They initially have distinct fractional area and thermodynamic properties, determined through surface layer similarity; that is, by assuming that the thermodynamic fluxes are constant throughout the surface layer. Each updraft ascent is diagnosed using an entraining plume model. In the case of cumulus convection, this comprehensive formulation provides consistent representations of the subcloud mixing, the cloud mixing, and the associated cloudiness. In the present paper, this model is evaluated in two shallow moist convection cases, using the one-dimensional modeling framework. In the continental convection case, the diurnal cloud growth and the related subcloud layer ventilation agree with large eddy simulations. In the oceanic convection case, the model reproduces the steady-state convection structure. It also satisfactorily represents the cumulus thermodynamics (average properties, variability). These results are shown to be relatively robust with respect to several model parameters and to the vertical resolution. Implications in terms of cumulus convection analysis and modeling are discussed. Some critical issues are emphasized, in various types of cloud and mixing schemes. Model predictions, supported by results found in the literature, emphasize the complementary roles of the subcloud layer plumes, the smaller cumuli, and the bigger less diluted cumuli. These various cloud types interact through the mean environment, so that the cloud population and the mean profiles adapt to the large-scale and surface forcings.