Computational study of inertial range spectral transfer by ExB and polarization drift nonlinearities

Spectral transfer processes in drift wave turbulence models, including fluid models for trapped particle modes, are governed by two nonlinearities, the polarization drift nonlinearity no∇·vp(1) which dominates at short scale (kθρs approx-gt 1), and the ExB nonlinearity vE·∇n which dominates at large scale. Here, vp(1) = Bo-1(c/q)bxvE·∇vE is the nonlinear polarization drift, vE is the ExB drift, and n is the fluctuating density. The polarization drift nonlinearity admits two quadratic invariants (typically energy and enstrophy or mean square vorticity). This gives rise to a dual cascade process analogous to the cascade of 2-D Navier Stokes turbulence, in which enstrophy is cascaded to small scale and energy to large scale. The ExB nonlinearity is typically dominated by the dissipative contribution of nonadiabatic electrons to the density fluctuations and conserves only energy, which is cascaded to small scale. In addition, the ExB nonlinearity is capable of energy transfer which is strongly nonlocal in wavenumber space. In this work the authors report on initial results of a numerical study of the interplay of turbulent transfer by these two nonlinearities within a single drift wave model. From spectral and pseudospectral simulations, turbulent transfer properties in various parts of the spectrum, as well as the net transfer characteristics, are inferred from equilibrium spectra. Direct measurement of local and nonlocal inertial transfer into or out of selected regions of wavenumber space is accomplished and compared with the equilibrium spectrum and theoretical results. These measurements are made for undriven, undamped inertial flow from initial spectra which are consistent with similarity transfer from stirred large scales to dissipated small scales