Turbulence Memory’s Possible Influence on Tornado Intensity

In large-eddy simulation (LES), eddy viscosity type subgrid-scale models and boundary conditions assume an instant equilibrium between resolved mean shear and unresolved mean stress. This neglects the effect of so-called “turbulence memory”, the lag in time of mean stress to mean shear, which has been observed in laboratory experiments and direct numerical simulation (DNS). A new wall model that accounts for turbulence memory is implemented in LES of idealized tornadoes performed using Cloud Model 1 (CM1). The scheme exploits the fact that the acceleration of air parcels in a vortex is predominantly transverse (as opposed to streamwise). Depending on the magnitude of the transverse acceleration time scale relative to the turbulence time scale, the radial convergence of angular momentum toward the axis of the vortex may be enhanced by the transverse stress “remembered” by turbulence, which, in turn, affects tornado intensity.