An Integrated Approach for Conducting long-term PM_2.5 Exposure and Health Impact Assessments for Residents of a City-scale

This study aims to show the benefits of integrating the mobile monitoring system (MMS), hybrid land use regression (h_LUR), and air quality monitoring station (AQMS) data in both conducting long-term PM2.5 exposure assessment (EA) and health impact assessment (HIA) for residents in a city-scale. A city installed with a government-operated AQMS was selected. A 1-year PM2.5 dataset was collected from AQMS (AQMS(1yr), reflecting the temporal variation of the rooftop level), and was served as a basis for characterizing the spatiotemporal heterogeneity of the rooftop level (h_LUR1yr) using the h_LUR model. A 1-year dataset was simultaneously collected from an established MMS for characterizing the spatiotemporal heterogeneity of the ground level (MMS1yr). A ground-level PM2.5 concentration predictive model was established by relating hourly MMS1yr to h_LUR1yr data and significant environmental covariables using the multivariate linear regression analysis. To establish long-term exposure datasets, 9-year AQMS data (AQMS(9yr)) were collected, and h_LUR9yr and MMS9yr were established through the application of the h_LUR and the obtained predictive model, respectively. Results show MMS1yr (24-26 mu g m(-3)) > h_LUR1yr (17-19 mu g m(-3)) > AQMS(1yr) (13-15 mu g m(-3)). An R-2 = 0.61 was obtained for the established ground predictive model. PM2.5 concentrations consistently decrease by year for MMS9yr (29-17 mu g m(-3)), h_LUR9yr (25-12 mu g m(-3)), and AQMS(9yr) (22-11 mu g m(-3)), respectively. The result MMS9yr > h_LUR9yr > AQMS(9yr) indicates both the use of h_LUR9yr and AQMS(9yr) would result in underestimating residents' exposures. By reference to the results obtained from MMS9yr, using AQMS(9yr) and h_LUR9yr would respectively lead to the underestimation of the attributed fraction (AFs) 21%-36% and 18%-26% for the 5 disease burdens, including ischemic heart disease, stroke, chronic obstructive pulmonary disease, lung cancer, and lower respiratory tract infection. The above results clearly indicate the importance of using the integrating approach on conducting EA and HIA.