Impact of radiative cooling and subgrid-scale mixing on the evolution of stratocumulus-topped boundary layer

In this paper, we will discuss results of a set of large-eddy simulations of stratocumulus-topped boundary layer (STBL) with modifications of processes believed to be important for the STBL structure and evolution. The experiments are based on a setup presented in Stevens et al. (Mon. Wea. Rev. 2005). The analysis expands the study of Kurowski et al. (Q. J. Roy. Met. Soc. 2009). The processes considered are the radiative cooling at the top of the cloud, the subgrid-scale mixing, and the delay of the phase change during turbulent mixing of cloudy air with warm and dry air entrained into STBL from above the inversion.

The radiative cooling is included in the model as proposed by Stevens et al. It is modified in sensitivity simulations by multiplying the radiative cooling by a factor between 0.5 and 2. Modification of the subgrid-scale mixing considers changes of the mixing length (between 0.5 and 2 of its basic value) and changes of the turbulent Prandtl number (between commonly used 0.33 and 1). The impact of the phase change delay associated with turbulent mixing is investigated by employing a scheme proposed in Grabowski (J. Atmos. Sci 2007) which delays evaporation of cloud water until the volume can be considered homogenized on the cloud microscale.

Reduction of the radiative cooling does not influence significantly domain averaged characteristics, such as the cloud cover fraction (CCF), liquid water path (LWP) and vertically integrated turbulent kinetic energy (TKE). The minor differences are observed in the vertical velocity at the inversion level since radiative cooling acts in the opposite direction than the large-scale subsidence. When cooling is increased significantly, development of STBL becomes unrealistic. The domain-averaged statistics for different values of the turbulent mixing length scale can vary up to 30%. A strong sensitivity of the solutions to the Prandtl number is also observed. The results suggest that the values taken from simulations of a dry boundary layer systematically overestimate thermal diffusion coefficient for the STBL. Both types of modification (i.e., the length scale and the Prandtl number) demonstrate that limiting thermal diffusion yields more realistic features of STBL, along with reduced decoupling within the boundary layer, when compared to observations. The delay of the phase changes during the turbulent cloud-environment mixing seems to be of less importance than the other processes.