The Use of AERMOD Air Pollution Dispersion Models to Estimate Residential Ambient Concentrations of Elemental Mercury

Abstract

Source-oriented models are ideally suited to examine the impact of terrain and meteorology and source factors such as stack height when evaluating exposures to air pollutants. A source-oriented, Gaussian plume air pollution dispersion model AERMOD was used to estimate the spatial distribution of elemental mercury (Hg0) from a typical coal-fired boiler emitting 0.001 g Hg0/s. Hg0 was chosen because of its health impact related to potential neurological and reproductive effects which may be especially important for high-risk populations. Results from four simulations using meteorological data from 2004 were compared for flat and hilly terrain from 20- and 55-m stacks at a distance of 1,350 m from the source. Variations within a quadrant were affected primarily by topography. For the 20-m stack, the average annual ambient concentration for individuals living within the northeast (NE) quadrant was significantly lower at 2.5 ng Hg0/m3 (P < 0.001; confidence interval (CI), 2.4–2.6) in flat terrain versus 3.3 ng Hg0/m3 in hilly (P < 0.001; CI, 1.2–1.3). NE concentrations of the source showed high spatial variability attributed to topography with 1-h maximums of 4.0 ng Hg0/m3 flat versus 7.1 ng Hg0/m3 hilly. Not unexpectedly, average annual concentrations were considerably lower for the 55-m stack although topography remained a significant variable with 0.1 ng Hg0/m3 in flat terrain (p < 0.001; CI, 0.11–0.13) and double that exposure at 0.2 ng Hg0/m3 in hilly terrain (p < 0.001; CI, 0.16–0.18). Annual average mercury concentrations due to emissions from the 20-m stack were ~20 times higher than ambient concentrations associated with the 55-m stack. A sensitivity analysis was performed for meteorological effects, using meteorological data from years 2001–2005. Varying the roughness factor had no significant effect on the results. For all simulations, the highest concentrations were located in the NE quadrant. During 2001–2005, the highest average annual ambient Hg concentration ranged from 6.2 to 7.0 ng Hg0/m3 for the 20-m stack and 0.3–0.5 ng Hg0/m3 for the 55-m stack. Thus, this model is robust. These results demonstrate the usefulness of a source-oriented model such as AERMOD for incorporating multiple factors for estimating air pollution exposures for communities near point sources. The importance of considering topography, meteorology, and source characteristics when placing air samplers to measure air quality and when using buffer zones to estimate ambient residential exposures is also illustrated. Residential communities in hilly terrain near industrial point sources may have between two to three times the exposures as those in flat terrain. Exposures will vary depending on the stack height of the point source.