Correlations of particle number concentrations and metals with nitrogen oxides and other traffic-related air pollutants in Glasgow and London

Particle number concentration (PNC) and transition metal content are implicated in the health effects of airborne particulate matter (PM) but they are difficult to measure so consequently their temporal and spatial variations are not well characterized. Daily concentrations of PNC and particle-bound water-soluble metals (V, Cr, Mn, Fe, Ni, Cu, As, Cd and Pb) were measured at background and kerbside sites in Glasgow and London to examine if other metrics of air pollution such as optical darkness (absorbance) of collected filter samples of PM, gravimetric PM, and NO, NO2 and CO gas concentrations, can be used as surrogates for the temporal and spatial variations of the former. NO2 and NOx exhibited a high degree of within-site correlation and with PNC and water-soluble metals (Fe, Cu, As, Cd, Pb) at background sites in both cities. There is therefore potential to use NO2 and NOx as surrogates for PNC and water-soluble metal at background sites. However, correlation was weaker in complex street canyon environments where pollutant concentrations are strongly affected by local sources and the small-scale variations in pollutant dispersion induced by the wind regimes within street canyons. The corollary of the high correlation between NO2 and PNC and water-soluble metals at the background sites is that the latter pollutants may act as confounders for health effects attributed to NO2 from such sites. Concentrations of CO cannot be used as a surrogate for PNC. Increments in daily NOx and NO2 concentrations between trafficked and background sites were shown to be a simple and novel surrogate for daily spatial variation of PNC; for example, increments in NOx explained 78–79% of the variance in PNC at the paired sites in both Glasgow and London, but relationships were city specific. The increments in NOx also explained 70% of the spatial variation in Cu and Ni in Glasgow but not in London. Weekly NO2 measurements derived from passive diffusion tubes were also shown to correlate well with increments in PNC. A high temporal correlation between PNC and 1,3-butadiene and benzene (which can also be measured by passive sampler) implies that passive sampler measurements may be a straightforward tool for deriving long-term spatial patterns in PNC