Search

Two schemes for modeling condensation in warm nonprecipitating clouds are compared. The first one is the efficient bulk condensation scheme where cloudy volumes are always at saturation and cloud water evaporates instantaneously to maintain saturation. The second one is the comprehensive bin condensation scheme that predicts the evolution of the cloud droplet spectrum and allows sub- and supersaturations in cloudy volumes. The emphasis is on the impact of the two schemes on cloud dynamics. Theoretical considerations show that the bulk condensation scheme provides more buoyancy than the bin scheme, but the effect is small, with the potential density temperature difference around 0.1 K for 1% supersaturation. The 1D advection–condensation tests document the high-vertical-resolution requirement for the bin scheme to resolve the cloud-base supersaturation maximum and CCN activation, which is difficult to employ in 3D cloud simulations. Simulations of shallow convection cloud fields are executed applying bulk and bin schemes, with the mean droplet concentrations in the bin scheme covering a wide range, from about 5 to over 4000 cm−3. Simulations employ the microphysical piggybacking methodology to extract impacts with high confidence. They show that the differences in cloud fields simulated with bulk and bin schemes come not from small differences in the condensation but from more significant differences in the evaporation of cloud water near cloud edges as a result of entrainment and mixing with the environment. The latter makes the impact of cloud microphysics on simulated macroscopic cloud field properties even more difficult to assess because of highly uncertain subgrid-scale parameterizations.

This paper presents an approach to locally predict homogeneity of the subgrid-scale turbulent mixing in large-eddy simulation of shallow clouds applying double-moment warm-rain microphysics. The homogeneity of subgrid-scale mixing refers to the partitioning of the cloud water evaporation due to parameterized entrainment between changes of the mean droplet radius and changes of the mean droplet concentration. Homogeneous and extremely inhomogeneous mixing represent two limits of possible scenarios, where the droplet concentration and the mean droplet radius remains unchanged during the microphysical adjustment, respectively. To predict the subgrid-scale mixing scenario, the double-moment microphysics scheme is merged with the approach to delay droplet evaporation resulting from entrainment. Details of the new scheme and its application in the Barbados Oceanographic and Meteorological Experiment (BOMEX) shallow convection case are discussed. The simulated homogeneity of mixing varies significantly inside small convective clouds, from close to homogeneous to close to extremely inhomogeneous. The mean mixing characteristics become more homogeneous with height, reflecting increases of the mean droplet size and the mean turbulence intensity, both favoring homogeneous mixing. Model results are consistent with microphysical effects of entrainment and mixing deduced from field observations. Mixing close to homogeneous is predicted in volumes with the highest liquid water content (LWC) and strongest updraft at a given height, whereas mixing in strongly diluted volumes is typically close to extremely inhomogeneous. The simulated homogeneity of mixing has a small impact on mean microphysical characteristics. This result agrees with the previous study applying prescribed mixing scenarios and can be explained by the high humidity of the clear air involved in the subgrid-scale mixing.

This paper discusses an extension of the approach proposed previously to represent the delay of cloud water evaporation and buoyancy reversal due to the cloud–environment mixing in bulk microphysics large-eddy simulation of clouds. In the original approach, an additional equation for the mean spatial scale of cloudy filaments was introduced to represent the progress toward microscale homogenization of a volume undergoing turbulent cloud–environment mixing, with the evaporation of cloud water allowed only when the filament scale approached the Kolmogorov microscale. Here, it is shown through a posteriori analysis of model simulations that one should also predict the volume fraction of the cloudy air that was diagnosed in the original approach. The resulting model of turbulent mixing and homogenization, referred to as the λ–β model, is applied in a series of shallow convection simulations using various spatial resolutions and compared to the traditional bulk model. This work represents an intermediate step in the development of a modeling framework to simulate characteristics of microphysical transformations during entrainment and subgrid-scale turbulent mixing.