Source impact modeling of spatiotemporal trends in PM_(2.5) oxidative potential across the eastern United States

摘要

Oxidative potential (OP) of particulate matter measures the ability of particles to catalytically generate reactive oxygen species while simultaneously depleting antioxidants, leading to oxidative stress and, in turn, inflammation in the respiratory tract and cardiovascular system. OP measurements have been linked with adverse cardiorespiratory endpoints, such as asthma/wheezing, lung cancer, and ischemic heart disease. However, measurements of OP are limited, restricting the area over which epidemiologic analyses can be performed. In this work, a modeling approach is developed and evaluated that uses limited measurements of water-soluble OP and PM2.5 source impact analysis to estimate OP over a large spatial domain (eastern United States). The dithiothreitol (DTT) assay was used to measure daily OP of water-soluble PM2.5 from June 2012 to July 2013 across four sites in the southeastern United States. Daily PM2.5 source impacts were estimated using CMAQ-DDM during the same time period and related to OPDTT measurements via multivariate linear regression. This regression was then applied to spatial fields of daily CMAQ-DDM source impacts across the eastern United States to provide daily spatially-varying OPDTT estimates. Backward selection during regression development showed vehicle and biomass burning emissions to be significantly predictive of OPDTT as observed in previous studies. The fire source impact was the largest contributor to OPDTT (29%) across the study domain during the study time period, and both spatial and seasonal variations were largely driven by fires. Vehicular impacts, especially diesel impacts, were more significant in urban areas. This CMAQ-DDM modeling approach provides a powerful tool for integrating OP measurements from multiple locations and times into a model that can provide spatio-temporal exposure fields of OPDTT across a wide spatial domain for use in health analyses, and the results offer insight into the large-scale spatial distribution of OPurr driven by emission source impacts.