Intra-urban correlation and spatial variability of air toxics across an international airshed in Detroit, Michigan (USA) and Windsor, Ontario (Canada)

As part of a larger research project initiated by the Geospatial Determinants of Health Outcomes Consortium (GeoDHOC), an air quality study was conducted in an international airshed encompassing Detroit, Michigan, USA, and Windsor, Ontario, Canada. Active and passive samplers were used to measure nitrogen dioxide (NO2), sulfur dioxide (SO2), 26 volatile organic compounds (VOCs), 23 polycyclic aromatic hydrocarbons (PAHs) and pesticides, and three size fractions of particulate matter (PM) over a two-week period in September, 2008. Measurements of NO2 and 14 VOCs were found to be acceptable at 98 out of 100 passive monitoring sites. PAH and PM measurements were acceptable at 38 out of 50 active sites. Mean concentrations for all analytes except for PM2.5e10 were higher in Detroit than Windsor by a factor of up to 1.8. Strong statistical correlations were found among benzene, toluene, ethylbenzene, and xylene (BTEX), as well as between NO2 andPM in Detroit. In Windsor, the strongest correlations were between NO2 and total VOCs, as well as total PAHs and total VOCs. Differences in the degree of correlation observed in Detroit and Windsor are attributable to differences in the volume and composition of emissions within the two cities. Spatial variability was evaluated using a combination of statistical (coefficient of variation) and geostatistical (standardized variogram slope) metrics together with concentration maps. Greater spatial variabilitywas observed for total VOCs and total BTEX in Detroit, while greater variability of NO2, total PAHs, and PM was found in Windsor. Results of this study suggest that statistical correlations between NO2 and other contaminants may not provide sufficient justification for the indiscriminant use of NO2 as a proxy for those contaminants if smaller scale features are to be reproduced.