Linear technique to understand non-normal turbulence applied to a magnetized plasma

In nonlinear dynamical systems with highly nonorthogonal linear eigenvectors, linear nonmodal analysis is more useful than normal mode analysis in predicting turbulent properties. However, the nontrivial time evolution of nonmodal structures makes quantitative understanding and prediction difficult. We present a technique to overcome this difficulty by modeling the effect that the advective nonlinearities have on spatial turbulent structures. The nonlinearities are taken as a periodic randomizing force with time scale consistent with critical balance arguments. We apply this technique to a model of drift wave turbulence in the Large Plasma Device [W. Gekelman et al., Rev. Sci. Instrum. 62, 2875 (1991)], where nonmodal effects dominate the turbulence. We compare the resulting growth rate spectra to the spectra obtained from a nonlinear simulation, showing good qualitative agreement, especially in comparison to the eigenmode growth rate spectra.