Disparities in exposure to traffic-related air pollution have been widely reported. However, little work has been done to simultaneously assess the impact of various vehicle types on populations of different socioeconomic/ethnic backgrounds. In this study, we employed an extreme gradient-boosting approach to spatially distribute light-duty vehicle (LDV) and heavy-duty truck emissions across the city of Toronto from 2006 to 2020. We examined associations between these emissions and different marginalization indices across this time span. Despite a large decrease in traffic emissions, disparities in exposure to traffic-related air pollution persisted over time. Populations with high residential instability, high ethnic concentration, and high material deprivation were found to reside in regions with significantly higher truck and LDV emissions. In fact, the gap in exposure to traffic emissions between the most residentially unstable populations and the least residentially unstable populations worsened over time, with trucks being the larger contributor to these disparities. Our data also indicate that the number of trucks and truck emissions increased substantially between 2019 and 2020 whilst LDVs decreased. Our results suggest that improvements in vehicle emission technologies are not sufficient to tackle disparities in exposure to traffic-related air pollution.
Keywords: air pollution; environmental justice; machine learning; nitrogen oxides; particulate matter; traffic emissions; trucks.