Evaluation of Methodologies for Exposure Assessment to Atmospheric Pollutants from a Landfill Site

Epidemiological studies around landfill sites are limited by several factors, particularly a lack of accurate exposure assessment. Traditionally, exposure estimates are based on distance between place of residence and the landfill site. However, this measure of exposure ignores the effects that environmental factors may have upon exposure. A previous epidemiological study at a landfill site in the United Kingdom provided the basis for a case study to investigate exposure assessment methodologies that could support ongoing and future epidemiological work. Estimation of relative exposure to atmospheric pollutants near the site was refined using the Atmospheric Dispersion Modeling System (ADMS) 3.1. Annual average concentrations were calculated around the landfill site, which was modeled as an area source with a steady release rate over its entire active surface. Local meteorological and terrain data were used in the assessment. A geographical information system (GIS) was then used to link the results of the modeling to population and other data. Sensitivity studies were included to examine the variation of predicted exposure with several modeling assumptions and hence set other uncertainties in context. No simple relationship existed between the relative individual exposure measured by distance from the site and by dispersion modeling. A reassessment of exposure assessment in epidemiological studies around landfill sites was then undertaken with the refined estimates of exposure. This concluded that use of distance from the site as a proxy for exposure could lead to significant exposure misclassification in comparison with exposure assessment using atmospheric dispersion modeling and GIS. The study also indicated that assessment of peak exposure rates (i.e., extreme concentration levels) might be necessary in some epidemiological work. Optimum strategies for increasing the probability of observing effects in the more highly exposed population can be derived by combining the results of dispersion modeling with population data and, where feasible, knowledge of the toxicology of the substances of interest.