Two-dimensional turbulent subgrid models

Summary form only given, as follows. The advent of high-performance scientific computing in the last few decades has given researchers a valuable tool for understanding turbulent transport in fluids and in plasmas. However, despite the efficiency of the pseudospectral method, the direct evaluation of statistical moments of the Navier-Stokes equation by numerical simulation of high-Reynolds number turbulence is not yet possible, even in two dimensions. The effort spent resolving the dissipation scales dominates the computation, even though it is often the dynamics of the large energy-containing scales that are of greater physical interest. We use these natural constraints of two-dimensional turbulence to develop more reliable subgrid models in which the ratio of the upscale and downscale transfer depend only on the wavenumbers and not on the energies of the deleted dissipation modes. We also consider a flux extrapolation scheme, where the subgrid model is constrained to remove a wavenumber-independent amount of energy flux from the small scales, after compensating for the effects of dissipation.