ALBERT

Description: Large-eddy simulation (LES) is a widely used technique in armospheric modeling research. In LES, large, unsteady, three dimensional structures are resolved and small structures that are not resolved on the computational grid are modeled. A filtering operation is applied to distinguish between resolved and unresolved scales. We present two near-surface models that have found use in atmospheric modeling. We also suggest a simpler eddy viscosity model that adopts Prandtl's mixing length model (Prandtl 1925) in the vicinity of the surface and blends with the dynamic Smagotinsky model (Germano et al, 1991) away from the surface. We evaluate the performance of these surface models by simulating a neutraly stratified atmospheric boundary layer.

Description: Globally averaged cloud changes from GCMs on average show a doubling of the Twomey effect, which is the change in cloud albedo with respect to changes in droplet concentrations for fixed cloud water and droplet dispersion. In contrast, ship-track measurements show a much more modest amplification of the Twomey effect, suggesting that the GCMs are exaggerating the indirect aerosol effect. We have run large-eddy simulations with bin microphysics of marine stratocumulus from multiple field campaigns, and find that the large-eddy simulations are in much better agreement with the ship-track measurements. The inversion strength over N. Pacific stratocumulus (as measured during DYCOMS-II) is generally much stronger than over N. Atlantic stratocumulus (as measured during ASTEX), and we have found that the response of cloud water to increasing droplet concentration changes sign as the inversion strengthens. For the different environmental conditions, we will show the overall response of cloud albedo to droplet concentrations, and decompose the response into its contributing factors of changes in cloud water, droplet dispersion, and horizontal inhomogeneity.

Description: The use of large eddy simulation, or LES, to study the atmospheric boundary layer dates back to the early 1970s when Deardor (1972) used a three-dimensional simulation to determine velocity and temperature scales in the convective boundary layer. In 1974 he applied LES to the problem of mixing layer entrainment (Deardor 1974) and in 1980 to the cloud-topped boundary layer (Deardor 1980b). Since that time the LES approach has been applied to atmospheric boundary layer problems by numerous authors. While LES has been shown to be relatively robust for simple cases such as a clear, convective boundary layer (Mason 1989), simulation of the cloud-topped boundary layer has proved more of a challenge. The combination of small length scales and anisotropic turbulence coupled with cloud microphysics and radiation effects places a heavy burden on the turbulence model, especially in the cloud-top region. Consequently, over the past few decades considerable effort has been devoted to developing turbulence models that are better able to parameterize these processes. Much of this work has involved taking parameterizations developed for neutral boundary layers and deriving corrections to account for buoyancy effects associated with the background stratification and local buoyancy sources due to radiative and latent heat transfer within the cloud (see Lilly 1962; Deardor 1980a; Mason 1989; MacVean & Mason 1990, for example). In this paper we hope to contribute to this effort by presenting a number of turbulence models in which the model coefficients are calculated dynamically during the simulation rather than being prescribed a priori.