Can exposure surfaces and GPS data predict personal exposures to air pollution and noise? Findings from a panel study

摘要

Exposure surfaces developed from land-use regression (LUR) models are commonly used to quantify the risks of various air pollutants and noise on public health. Typically, exposures are associated with the home location of participants in a cohort. While a few studies have begun to incorporate data on the daily mobility to refine exposure estimates, often, the "mobility-based" exposures are not validated against personal data. During summer 2016, we conducted a panel study in the city of Toronto, Canada, which involved 46 participants who participated on two days. Participants were provided with instruments measuring levels of Ultrafine Particles (UFP), Black Carbon (BC) and noise, as well as a GPS. Over a 6-hours duration, they were free to pursue their daily activities with the constraint of walking for 2 hours outdoors. During the same summer a data collection campaign took place whereby the levels of UFP, BC and noise were measured in Toronto with the same devices via two different protocols. A mobile monitoring campaign using bicycles enabled the coverage of 3,895 unique road segments visited 5 to 6 times. In addition, 92 fixed points were also sampled 5 to 6 times. These simultaneous monitoring campaigns enabled the development of LUR models and associated exposure surfaces. The GPS coordinates of all participants were synchronised with the air pollution and noise data to generate estimates of personal exposures. These estimates were analyzed in terms of spatial variability throughout the city, and differences between indoor and outdoor levels. In addition, GPS data were intersected with the surfaces to derive "mobility-based" exposures. We are currently comparing the personal and "mobility-based" exposure estimates and identifying locations, and trajectories with strong and poor agreement between the two measures. We are also developing models that can correct "mobility-based" exposures using information on daily levels measured at a central location.