RAMS with the Land Ecosystem Atmosphere Feedback (LEAF) module is an ideal tool for studying the surface heterogeneity effect on atmospheric processes. In the present study, a Largrangian particle model is embedded in RAMS to study the effect on atmospheric dispersion in short-range and longer-range transport of air toxics or pollutants. Two different parameterizations for subgrid statistics are installed in the particle model for small scale turbulence in short-range transport and sub mesoscale turbulence in longer-range transport, respectively. For the study, the RAMS model will be performed with and without realistic land use data, and with artificial surface heterogeneity, for comparison. The primary result from the model simulations for a case in Mississippi shows that the surface heterogeneity could significantly enhance particle dispersion.