Measuring and modeling particulate air pollution using a mobile, bicycle-based platform

Abstract

Background: Air pollution concentrations exhibit significant within-city variability. Few studies use mobile measurements to estimate land-use regression models. Aim: To develop land-use regression models of particulate air pollution using a bicycle-based, mobile sampling method. Methods: We measured particulate air pollution during morning (7-9am; n=12 days) and afternoon (4-6pm; n=30 days) rush hours on weekdays (8/14 – 10/16/2012). Measurements were performed using a mobile platform. Our modified bicycle trailer contained equipment to measure: (1) particle number concentration, (2) PM2.5 mass concentration, (3) black carbon concentration, and (4) particle size distributions. Three sampling routes (~20 miles each) were cycled repeatedly (10 [4] times per route in the afternoon [morning]) to estimate typical rush hour concentrations in urban environments throughout Minneapolis, MN. Results: Our results are in mainly two categories: (1) correlates of the built environment and particulate air pollution and (2) land-use regression models to estimate concentrations of air pollution during rush hour for street segments in Minneapolis. We found particulate air pollution concentrations are related to aspects of the built environment, for example, mean particle number concentration varies by street classification: 19,009 pt/cc (arterials), 18,374 pt/cc (collectors), 14,059 pt/cc (local), and 13,186 pt/cc (off-street trails). In general, particle number concentrations were 1.9× higher during morning rush hour (mean: 31,043 pt/cc) than during afternoon rush hour (mean: 16,312 pt/cc). Policy implications: Our results demonstrate mobile-monitoring land-use regression, and inform efforts to reduce air pollution exposures and plan for health-promoting non-motorized transportation.