Improving the representation of turbulence and clouds in cloud resolving models and general circulation models

Over the past few years a new type of general circulation model (GCM) has emerged that is known as the multiscale modeling framework (MMF). The Colorado State University (CSU) MMF represents a coupling between the Community Atmosphere Model (CAM) GCM and the System of Atmospheric Modeling (SAM) cloud resolving model (CRM). Within this MMF the embedded CRM replaces the traditionally used parameterized moist physics in CAM to represent subgrid-scale (SGS) convection. However, due to substantial increases of computational burden associated with the MMF, the embedded CRM is typically run with a horizontal grid size of 4 km. With a horizontal grid size of 4 km, a low-order closure CRM cannot adequately represent shallow convective processes, such as trade-wind cumulus or stratocumulus. A computationally inexpensive parameterization of turbulence and clouds is presented in this dissertation. An extensive a priori test is performed to determine which functional form of an assumed PDF is best suited for coarse-grid CRMs for both deep shallow and deep convection. The diagnostic approach to determine the input moments needed for the assumed PDFs uses the subgrid-scale (SGS) turbulent kinetic energy (TKE) as the basis for the parameterization. The term known as the turbulent length scale (L) is examined, as it is needed to parameterize the dissipation of turbulence and therefore is needed to better balance the budgets of SGS TKE. A new formulation of this term is added to the model code which appears to be able to partition resolved and SGS TKE fairly accurately. Results from "offline" tests of the simple diagnostic closure within SAM shows that the cloud and turbulence properties of shallow convection can be adequately represented when compared to large eddy simulation (LES) benchmark simulations. Results are greatly improved when compared to the standard version of SAM. The preliminary test of the scheme within the embedded CRM of the MMF shows promising results with the simulation of shallow convection. Overall, this scheme represents a new type of flexible turbulence and cloud parameterization that is computationally efficient.