Multi-season, multi-year concentrations and correlations amongst the BTEX group of VOCs in an urbanized industrial city

An air quality monitoring study focusing on spatial patterns was carried out in the urban industrial city of Windsor, Ontario, Canada (42.267°N, 82.95°W). This study took place over a three-year period (2004–2006), during all four seasons for a total of 12 two-week sampling periods (each with approximately 50 sites) at 162 sites across the city. Benzene, toluene, ethylbenzene, (m + p)-xylene, and o-xylene (BTEX) were measured using 3M #3500 Organic Vapor Samples (Guillevan, Montreal). Results from this investigation indicate that significant variability is present temporally (seasonally and annually) and spatially. The three-year mean concentrations in μg m−3 were: benzene (0.76), toluene (2.75), ethylbenzene (0.45), o-xylene (0.47), (m + p)-xylene (1.36), and total BTEX (5.64), with greater variability within each year compared with variability between the three years. Concentrations were highly correlated between most BTEX species, consistent with previous studies in urban areas. Toluene to benzene ratios were consistent between years, with the highest observed ratios occurring in summer, the lowest in winter, and fall and spring values falling in between. The range of ratios suggests that the majority of these two compounds originate from mobile emissions. (m + p)-Xylene to ethylbenzene ratios, used as indicators of photochemical age, showed a three-year mean of 3.0 with little spatial and temporal variability suggesting that relatively fresh and homogenous sources of these species are present in this area. The seasonal trends across the spatial network were representative of the seasonal patterns obtained at a long term monitoring station, with both methods indicating that fall and spring concentrations were preferred proxies of annual means. However, significant spatial variability of concentrations was observed. The upper range of concentrations compared with the values obtained at the central monitoring station differed by up to a factor of six, highlighting the importance of multiple sampling sites to reduce exposure misclassification. Significant correlations between seasons were observed in all three years, revealing a consistent rank order of high to low concentrations among the monitoring sites within a given year. Consideration of all analyses suggests that when resources are limited to a single sampling season, fall and spring mean values for concentrations and ratios seem to best represent the annual average in this city. Other study design consideration suggestions are also provided for future air quality research.