Apportionment of Human Exposure to Fine Particulate Matter Among Indoor and Outdoor Emission Sources and Processes

摘要

Human exposure to PM_(2.5) is associated with adverse health effects. Current methods for assessing the linkage between exposure and health effects are based on the total mass of PM_(2.5). PM_(2.5) varies substantially in chemical composition depending on whether the particles are from combustion, mechanical processes, or secondary formation in the atmosphere. However, there is not yet sufficient data to attribute health effects to specific chemical components or source-specific mixtures of such components in PM_(2.5). In order to provide a basis for such assessments, there is a need to develop techniques for apportioning PM_(2.5) exposure concentration to source categories, taking into account chemical composition. Exposure apportionment to source categories is conducted based on separately estimating the ambient and non-ambient contributions using a scenario-based exposure simulation model. The non-ambient exposures are further apportioned to specific indoor emission source categories based on sensitivity analysis of the exposure model. The ambient exposures are apportioned to outdoor emission sources and ambient secondary aerosols using a receptor-oriented modeling approach. The results from these methods are combined into one overall apportionment of total daily average exposure to specific outdoor and indoor emission sources, and to secondary aerosols. The case study results are consistent with the local emission inventory and with ambient source apportionment studies for the same geographic region. Regional differences in smoking prevalence might be associated with differences in underlying health conditions of an exposed population that could lead to higher sensitivity to adverse effects from exposure to ambient PM_(2.5). There are limitations to the methodology, such as the use of ambient monitoring data and source profiles that are subject to measurement error, the possible lack of representativeness of available source profiles with respect to local emission sources in a specific location at a specific time, and uncertainties in input data to the exposure model such as regarding representativeness of diary activity data, indoor exposure factors such as air exchange rates, and others. Nonetheless, the example case study results appear to be reasonable and plausible. Thus, these results indicate that the method explored here merits further development and evaluation.