Lagrangian stochastic dispersion model simulations of tracer data in nocturnal flows over complex terrain

A three-dimensional Lagrangian stochastic dispersion model (LSDM), which has been formulated to predict dispersion in three-dimensional nocturnal flows over complex terrain, is described. The model uses a Langevin equation in the lateral direction and a pure random-walk equation in the vertical direction. The performance of the model has been evaluated using the U.S. Department of Energy's ASCOT tracer databases collected during the 1984 field study in the Brush Creek valley in western Colorado, and the 1991 field study near the Rocky Flats Plant northwest of Denver. Both field experiments were performed in complex terrain regions where three-dimensional drainage flows were observed at night. The input wind fields and turbulence parameters required by the Lagrangian stochastic dispersion model are derived from meteorological data collected during the tracer experiments. The model-predicted concentrations are extensively compared with the corresponding observed concentrations. Several statistical measures calculated to quantitatively assess the performance of the model are presented. The agreement between the predicted and observed concentrations is satisfactory.