A Lagrangian Monte Carlo model of turbulent dispersion in the convective planetary boundary layer

A Lagrangian Monte Carlo model for predicting the dispersion of a passive tracer in a convective boundary layer is presented. The stochastic model provides a more realistic treatment of convective turbulence than previous modeling approaches. Accurate input for the dispersion prediction is provided by extensive water-tank measurements of convective turbulence. The dispersion of a large number of passive tracer particles is computationally simulated by using the Langevin equation to model the Lagrangian velocities. The behavior of the autocorrelation of the modeled Lagrangian velocities closely matches the nonexponential form computed from balloon-borne measurements in the atmosphere. A kernel estimation technique is employed to efficiently recover mean concentrations from the trajectory simulations and reduce computational requirements. The predictions of the stochastic model are in close agreement with the dispersion trends and magnitudes observed in the data.