Abstract
Shallow cumulus clouds are individually small, only hundreds of meters in horizontal extent,but exert an influence far outweighing their size. Scattered cumulus fields cover vast regions of the Earth’s surface and are an essential component of the climate system by way of lower-tropospheric mixing and radiative effects. Because of their small size, shallow convection occurring inside a climate-model grid box must be parameterized. As climate model grid sizes continue to decrease, the standard approach for parameterizing shallow convection via the mass-flux method begins to break down as the underlying assumptions of quasi-equilibrium and small cloud area fraction begin to become violated. Future representations of shallow cumulus may rely heavily on the physics of individual clouds, and this research aims to gain a better process-level understanding of cloud-scale dynamics and mixing processes. This dissertation uses large-eddy simulation (LES) to composite clouds of similar depths andexamine detailed cloud-edge structures and mass flux as a function of vertical wind shear. The LES configuration is shown to reasonably reproduce lidar-observed vertical velocity structures. The cloud updrafts are found to be predominantly opposed by a downward-directed pressure gradient force with down-shear portions of cloud updrafts having a weakened downward-directed pressure gradient force, suggesting all else being equal, the down-shear updraft would be more vigorous. However, reduced down-shear vertical velocities are a result of buoyancy dilution associated with an organized entrainment-to-detrainment couplet that develops as as sub-cloud layer shear is advected into the cloud layer by the updraft. Total net mass flux through the cloud interface is insensitive to shear and the clouds do not behave as entraining plumes, but rather, weakly detraining plumes. A modern mass-flux parameterization performs poorly in comparison to LES vertical velocities and would benefit from the addition of the pressure gradient force term in the vertical velocity budget equation.