Subgrid-scale (SGS) modeling is a long-standing problem in the large-eddy simulation (LES) of atmospheric boundary layer. It is a critial componement of a successful simulation. A number of SGS models have been proposed for various levels of requirement, such as Smagorinsky's (1963) eddy viscosity model, Mason and Thomson's (1992) stochastic backscatter model, Sullivan, et al.'s (1994)two-part eddy viscosity model, and Branko's (1997) nonlinear model. A modified
Smagorinsky first-order closure has been used in the Terminal Area Simulation Syatem (TASS-LES). While it could successfully simulate buoyancy-dominated atmospheric boundary layer flows, simulations for shear-dominated and neutral boundary layer flows are somewhat inferior. For the later, two-part eddy viscosity model of Sullivan et al. has been modified and implemented into the TASS-LES model.
In this SGS model, eddy viscosity is separated into a fluctuating part and a mean-field part, and the equations of computing mean-field eddy viscosity are based on the surface layer similarity theory. The turbulent kinetic energy equation is not included and the model still uses the simpler first-order closure scheme. The new model uses the same expression for estimating mean-field eddy-viscosity at the flist grid level above the surface as derived by Sullivan et al.. At any other height, howeve, a different expression based on the SGS momentum flux profile is used. With this SGS model, the TASS-LES is used for simulations for shear-dominated weakly convevtive flow and neutral flow. The results are compared with other first-order SGS models, as well as the surface layer similarity relations. The new model might be useful for simulating stable boundary layers.