Background

Certain metals may play an important role in the adverse health effects of fine particulate air pollution (PM_2.5), but few models are available to predict spatial variations in these pollutants.

Methods

We conducted large-scale air monitoring campaigns during summer 2016 and winter 2017 in Toronto, Canada, to characterize spatial variations in iron (Fe) and copper (Cu) concentrations in PM_2.5. Information on Fe and Cu concentrations at each site was paired with a kinetic multilayer model of surface and bulk chemistry in the lung epithelial lining fluid to estimate the possible impact of these metals on the production of reactive oxygen species (ROS) in exposed populations. Land use data around each monitoring site were used to develop predictive models for Fe, Cu, and their estimated combined impact on ROS generation.

Results

Spatial variations in Fe, Cu, and ROS greatly exceeded that of PM_2.5 mass concentrations. In addition, Fe, Cu, and estimated ROS concentrations were 15, 18, and 9 times higher during summer compared with winter with little difference observed for PM_2.5. In leave-one-out cross-validation procedures, final multivariable models explained the majority of spatial variations in annual mean Fe (R ² = 0.68), Cu (R ² =0.79), and ROS (R ² = 0.65).

Conclusions

The combined use of PM_2.5 metals data with a kinetic multilayer model of surface and bulk chemistry in the human lung epithelial lining fluid may offer a novel means of estimating PM_2.5 health impacts beyond simple mass concentrations.