New Observations on Subgrid-Scale Modeling from Field Experiments: Davis99

Field measurements of velocity and temperature using two horizontal arrays of 3d-sonic anemometers were conducted in the San Joaquin Valley in California during the Davis99 field campaign. These experiments were specifically designed to help in improving subgrid-scale models for large eddy simulation (LES) of the atmospheric boundary layer. Analysis of the field data demonstrates the strong dependence of subgrid-scale model coefficients (here: Smagorinsky coefficient c_s) on stability z/L and non-dimensionalized height z/Δ , z is the height of the instrument, L the Obukhov-length, and Δ the filter size. For a given z/Δ , c_s is small in stable atmospheric conditions and inreases rapidly under convective conditions. For large z/Δ values (corresponding to better resolved flow conditions in LES), c_S shows less dependence on stability and is found to be constant at c_s ~ 0.1. For a given stability z/L, c_s increases with nondimensional height z/Δ . Conditional averaging is used to isolate effects of coherent structures on the flow. Sweep and ejection events have significant impact on the subgrid-scale dissipation of scalar variance, consequently accurate modeling of coherent structures is of importance. >http://www.jhu.edu/~dogee/mbp/sgs/combine/ cs_zoL_zoD_c_all_co.jpg</a>