Scale-wise relaxation to isotropy in direct numerical simulations

<p>The Kolmogorov hypothesis of local isotropy is fundamental in stochastic models of turbulence and generally assumed to hold for atmospheric turbulence. According to Kolmogorov’s second similarity hypothesis, there is a range of turbulent scales (inertial subrange) that are statistically isotropic and the statistics of these scales have a universal form that is uniquely determined by the TKE dissipation rate. Recent work based on atmospheric turbulence measurements has shown that the scale-wise route turbulence takes to reach isotropy at these smallest scales is uniquely determined by the anisotropy of the energy containing eddies.</p><p>In this study we explore the connection between large-scale anisotropy and the route to small-scale isotropy through direct numerical simulations (DNS). We perform simulations of neutral flow over flat and rough (wavy) surfaces at different Reynolds numbers, to investigate the scale-wise anisotropy as a function of height from the surface and surface-roughness. The resulting trajectories of relaxation to isotropy are compared to the experimental ones and the differences between the two are explored in light of the return-to-isotropy terms and Reynolds number.</p>