Past lead ore processing conducted in the Southeast Missouri Old Lead Belt since the 1700s has left numerous large areas of lead contamination in elevated piles of fine gravel waste called "chat" and dried-out tailings ponds. Wind suspension and atmospheric dispersion are known to transport these materials to the surrounding communities where the lead could pose a human health threat through inhalation or ingestion of the deposited contamination. The purpose of this study was to estimate potential wind suspension rates, perform dispersion modeling of the tailings and chat sources, and determine ground surface deposition rates and potential soil concentrations of lead in the surrounding areas. The results can be used to prioritize soil sampling locations, site air monitors, help identify the source of soil lead contamination, and to help develop remediation plans. Numerous, large complex sources in the region were parameterized into 33 area sources with the aid of digital aerial photos, digitized typographic maps, Geographic Information Systems (GIS) software, and site inspections. An AP-42 particulate emission model was used to estimate lower-and upper-bound hourly emission rates using six years of hourly wind speed data obtained from the St. Louis Airport. The emissions model accounted for wind speed, precipitation, source-specific aggregate size, fraction of vegetation cover, and site-specific lead concentrations. An alternative simplified method to calculate emissions from elevated chat piles was developed. The Fugitive Dust Model (FDM) was then used to calculate long-term average and maximum 24-hour deposition rates of lead over a 200 km 2 region. Soil concentrations were estimated from modeled deposition rates, time of deposition (80 y) and an assumed surface (0-5.08 cm) mixing depth. Model performance was evaluated by comparing lower-and upper-bound modeled predictions to both air and soil sampling data obtained at two sites. The predicted-to-observed (P/O) ratios calculated using the geometric mean of the lower-and upper-bound modeled concentrations ranged from a low of 0.84 at 300 m to a high of 1.4 at 1800 m, with an average of 1.1. The P/O ratios tended to increase slightly at distances beyond about 1 km. These P/O ratios demonstrate exceptional model performance for the particular sampling transect (location and direction) investigated.
CONTENTS