How Good Are LES Model Simulations of Marine Stratocumulus Clouds? a Comparison Study of Modeling and Observation

Large Eddy Simulation (LES) models are known as an appropriate tool for investigating microphysical, dynamical and radiative properties and their interactions in stratocumulus clouds, which are known to be critically important in modulating climate change due to their large areal extent. To check the reliability of such models, we try to examine how realistic model simulated stratocumulus cloud properties are. The observation data are obtained from aircraft measurement of marine stratocumulus clouds during the Variability of the American Monsoon System Ocean-Cloud-Atmosphere-Land Study Regional Experiment (VOCALS-REx). Several observed stratocumulus clouds during the project are then simulated with two LES models: one is CIMMS (Cooperative Institute for Mesoscale Meteorological Studies) LES and the other is WRF (Weather Research and Forecasting Model) LES. Both models are 3 dimensional and bin microphysics schemes are implemented. Comparison between observation and model simulated results focuses on cloud microphysical relationships which can reveal how well entrainment and mixing processes are treated in the models. Some preliminary results indicate that both models show reasonably good performance in their assessment of thermodynamic and macro-physical characteristics of stratocumulus clouds. However, the relationships between microphysical variables such as number concentrations, liquid water content and mean volume of cloud droplets are not consistent with those observed in these clouds. The model simulated cloud microphysical relationships generally suggest homogeneous mixing while relevant scale parameters that can be estimated from environmental (thermodynamic and microphysical) conditions indicate dominance of inhomogeneous mixing especially near cloud top. This is different from the observation that showed a trait of inhomogeneous mixing near cloud top. Obviously a proper treatment of entrainment and mixing processes is required in LES models to attain greater reliability.