A microscale roadway dispersion model based on Lagrangian diffusion theory has been developed and tested. The model incorporates similarity expressions for the mean wind and turbulence energy in the atmospheric boundary layer, through which the effects of wind shear and atmospheric stability are taken into account, and a parameterization of vehicle wake turbulence. Through simple modifications, the model can be structed to treat particle settling, deposition, and resuspension, as well as buoyancy of the effluent material. Calm winds, winds parallel to the roadway, flows around depressed or elevated roadways, shallow mixed layers, and transient of spatially variable meteorological conditions can all be explicity taken into account within the framework of the modeling approach. The model was tested by applying it to the 30-minute experimental periods reported in the General Motors sulfate study. Of the 1040 predicted values of the mean concentration of an inert material (SF6), half were found to be within +30 percent of the measured values. The overall correlation coefficient was 0.91. The computer time (but not core storage) required by the model is directly proportional to the distance between the farthest receptor and the road. For the studies reported the model requires on the average 20 seconds of CPU time on the CDC 7600 to simulate each of the 30-minute General Motors experiments.

EPA 600/4-79-023. April, 1979.