Characterization of the chronic risk and hazard of hazardous air pollutants in the United States using ambient monitoring data.

Ambient measurements of hazardous air pollutants (air toxics) have been used to validate model-predicted concentrations of air toxics but have not been used to perform risk screening at the national level. We used ambient concentrations of routinely measured air toxics to determine the relative importance of individual air toxics for chronic cancer and noncancer exposures. We compiled 3-year averages for ambient measurement of air toxics collected at monitoring locations in the United States from 2003 through 2005. We then used national distributions of risk-weighted concentrations to identify the air toxics of most concern. Concentrations of benzene, carbon tetrachloride, arsenic, 1,3-butadiene, and acetaldehyde were above the 10(-6) cancer risk level at most sites nationally with a high degree of confidence. Concentrations of tetrachloroethylene, ethylene oxide, acrylonitrile, and 1,4-dichlorobenzene were also often greater than the 10(-6) cancer risk level, but we have less confidence in the estimated risk associated with these pollutants. Formaldehyde and chromium VI concentrations were either above or below the 10(-6) cancer risk level, depending on the choice of agency-recommended 10(-6) level. The method detection limits of eight additional pollutants were too high to rule out that concentrations were above the 10(-6) cancer risk level. Concentrations of 52 compounds compared with chronic noncancer benchmarks indicated that only acrolein concentrations were greater than the noncancer reference concentration at most monitoring sites. Most pollutants with national site-level averages greater than health benchmarks were also pollutants of concern identified in modeled national-scale risk assessments. Current monitoring networks need more sensitive ambient measurement techniques to better characterize the air toxics problem in the United States.