In the present study we used the large-eddy simulation technique to simulate the moist convective boundary layer (CBL) with shallow cumulus clouds in order to investigate the effect of cloud shadows on the CBL over homogeneous and heterogeneous terrain. Our results show a considerable response of the CBL due to cloud shading.
Depending on solar zenith angle (SZA) shadows trigger secondary circulations with differing strength over a homogeneous surface. It is found that large SZA favor the development of secondary circulation formation. Convective clouds develop on top of warm updrafts, which evolve from solar heating of the surface. If such a developed cloud shades the region directly underneath itself, as it is the case when sun is standing in zenith, heating decreases and the updraft decays. This causes the cloud to decay as well. If the shadow is shifted beside the cloud, it triggers a heterogeneous surface heat flux with a maximum underneath the cloud. This leads to a constant updraft below the cloud and thus it appears that shading also effects cloud lifetime and size. With increasing strength of the triggered circulation clouds exist for a longer time and the number of large clouds increases.
When considering a heterogeneous surface heat flux with a striped pattern, cloud shadows affect the secondary circulations induced by the striped pattern. We considered an east-west orientation for the stripes and the sun in the south, so that the cloud shadows are shifted perpendicular to the stripes. Depending on the SZA the maximum (low SZA) of surface heat flux or the minimum (large SZA) is shaded by the clouds. While only a slight change in strength of the secondary circulation could be observed between the reference simulation without cloud shadows and the simulation with large SZA the secondary circulations get clearly reduced when a low SZA is considered. Also cloud lifetime and size are again affected by cloud shadows.
These findings show that neglecting cloud shadows, which is so far done in most simulation studies, leads to a significant error in the simulation results.