| description abstract | The sensitivity of a regional climate model to physical parameterizations and model resolution is investigated in terms of its simulation of boundary layer stratocumulus (SCu) clouds over the southeast Pacific. Specifically, the physical schemes being tested include shallow cumulus convection, subgrid vertical mixing, cloud droplet number concentration (CDNC), and drizzle. As described in Part I, the model with standard settings captures the major features of the boundary layer in the region, including a well-mixed marine boundary layer, a capping temperature inversion, SCu clouds, and the boundary layer regime transition from the well-mixed layer near the coast of South America to a decoupled cloud layer over warmer water to the west. Turning off the shallow cumulus parameterization results in a dramatic increase in the simulated SCu clouds while the boundary layer structure becomes unrealistic, losing the decoupled regime over warm water. With reduced penetrative mixing at the top of shallow cumuli, the simulated SCu clouds are somewhat increased while the boundary layer structure remained largely unchanged. Reducing the CDNC increases the size of cloud droplets and reduces the cloud albedo but has little effect on the vertical structure of the boundary layer and clouds. Allowing more drizzle decreases boundary layer clouds considerably. It is also shown that the simulated depth of the boundary layer and its decoupling is highly sensitive to the model horizontal and vertical resolutions. Insufficient horizontal or vertical resolutions produce a temperature inversion and cloud layer too close to the sea surface, a typical problem for global general circulation models. Implications of these results for global and regional modeling of boundary layer clouds and the areas that need more attention in future model development are discussed. | |
