No abstract was published. A summary of the Introduction follows.

For water-related environmental problems, there are numerous existing methodologies which are based on rigorous application of the fundamental theory of turbulent diffusion. Turbulent diffusion enables the prediction of concentrations in which pollutants may be found in the vicinity of a point of release, such as a chimney or sewage outfall. Approaches currently used for air-quality problems, however, are yet to be improved. EPA-designed air-quality models are based on the application of Gaussian plume (or puff) dispersion in an unbounded domain with a variety of statistical and empirical adjustments which are often difficult to justify. These models produce reasonable results sometimes, like the far-field concentration of an unbounded plume in a unidirectional wind, but usually predicted values are far from reality due to excessive empiricism and simplifications in the approach. As such, these models can only serve as crude guidelines for comparative purposes.

This report describes the model development stage, which has attempted to compromise between the practical applicability and the theoretical conformity in prediction of three-dimensional, steady-state air quality over a complex terrain. A unique feature of this model is the method of optimizing the 3-D wind field using the variational principle and sparsely measured field data. The optimized wind field fully satisfies the continuity requirement, so the mass conservation for the subsequent pollutant dispersion computation is guaranteed. This study is intended to improve some of the shortcomings of the EPA models.