Parameterization of the two-point velocity correlation function in a two-particle Lagrangian stochastic model and its effect on the prediction of concentration variances due to a line source

A Lagrangian stochastic model, in which a new parameterization of the two-point velocity correlation function is included, is investigated. Model simulations of temperature variances caused by a line source in an inhomogeneous wind-tunnel boundary layer are shown to be in good agreement with experimental data. A special limit of the new parameterization that is useful at large distances from the source is discussed in detail. A new technique is introduced that greatly enhances the computational efficiency of an implementation of this limit of the model and that renders long-distance simulations of concentration variances of tracers dispersing from line sources computationally treatable.