Monitoring, mapping, and modeling spatial–temporal patterns of PM2.5 for improved understanding of air pollution dynamics using portable sensing technologies

Abstract

Fine particulate matter with an aerodynamic diameter of less than 2.5 µm (PM2.5) is highly variable in space and time. In this study, the dynamics of PM2.5 concentrations were mapped at high spatio-temporal resolutions using bicycle-based, mobile measures on a university campus. Significant diurnal and daily variations were revealed over the two-week survey, with the PM2.5 concentration peaking during the evening rush hours. A range of predictor variables that have been proven useful in estimating the pollution level was derived from Geographic Information System, high-resolution airborne images, and Light Detection and Ranging (LiDAR) datasets. Considering the complex interplay among landscape, wind, and air pollution, variables influencing the PM2.5 dynamics were quantified under a new wind wedge-based system that incorporates wind effects. Panel data analysis models identified eight natural and built environment variables as the most significant determinants of local-scale air quality (including four meteorological factors, distance to major roads, vegetation footprint, and building and vegetation height). The higher significance level of variables calculated using the wind wedge system as compared to the conventional circular buffer highlights the importance of incorporating the relative position of emission sources and receptors in modeling.