Numerical simulations of flow reversal in Rayleigh--B\'enard convection

We investigate numerically the statistical properties of the large scale flow in Rayleigh--B\'enard convection. By using an external random perturbation on the temperature field, we were able to decrease the effective Prandtl number of the flow while keeping the Rayleigh number relatively small; this increases the Reynolds number thus making possible the numerical investigation of the long--term flow statistics. We also propose a simple and quantitative explanation for the experimental findings on the statistical distribution of flow reversals and reorientations.