The estimation of risk associated with releases of toxic pollutants in urban areas requires a dispersion model that provides concentration estimates in complex flow situations. Existing dispersion models, such as ISC, are not designed for use in the urban canopy in which the source of pollutant is typically surrounded by other buildings. Additions to ISC, such as PRIME, are designed primarily for the effects of isolated buildings on elevated sources. ISC-PRIME is not applicable to air toxic releases in urban areas.

The literature describes several models that apply to dispersion in the vicinity of isolated buildings. These cannot be readily adapted to situations in which the source is in the midst of several buildings. Several studies have been conducted to understand dispersion in urban areas, but there are no acceptable models for estimating concentrations in such areas.

The paper presents the development and evaluation of a dispersion model for toxic releases in urban areas. This model is designed to use modeled or measured turbulence and flow fields to estimate concentrations without explicitly considering details of buildings in the urban area. The model accounts for the fact that in the urban canopy, turbulent velocities are likely to be as large or larger than the mean winds.

This model will be first evaluated under relatively idealized conditions at the Dugway Proving Grounds, Utah. The experiment involved the release of propylene (C3H6) in a model urban area constructed with 55-gallon drums. The tracer was sampled used photoionization detectors, and the turbulence and mean field was measured using an array of sonic anemometers. The results from this study will provide the understanding of the governing processes required to modify the model for application to real urban situations. In the second phase of the study, the dispersion model will be evaluated using data from field studies conducted in Barrio Logan, San Diego. In these studies, SF6 was released in urban canopies and sampled at 50 locations at distances up to 2 km from the source. Mean flow velocities and turbulence measurements were acquired with an array of six, three component, sonic anemometers. A minisodar was deployed in the vicinity of the sonic arrays to determine the properties of the atmospheric surface layer above the roughness sublayer.