Quantifying the ozone "weekend effect" at various locations in Phoenix, Arizona

Analysis of pollution data from a network of monitors in Maricopa County, Arizona, reveals considerable variation in the magnitude of the ozone "weekend effect" depending on how and where it is measured. We used four separate methods to calculate the weekend effect, all of which showed that the phenomenon is stronger in the urban core, where ozone is produced. Spatial linear regressions show that the magnitude of the weekend effect and the goodness of fit of weekly harmonic cycles in ozone is a function of urbanization, described quantitatively using an index of traffic counts, population, and employment within a 4 km buffer zone of monitoring sites. Analysis of diurnal patterns of ozone as well as oxides of nitrogen (NO,) at a representative site in the urban core supports the hypothesis that lower levels of NO, on Sundays reduce the degree to which ozone is titrated, resulting in a higher minimum and hence mean for that day of the week (DOW). Fringe sites, where ozone concentrations are higher in absolute terms than in the urban core, show almost no "weekend effect," regardless of which of the four individual methods we used. Alternative quantification methods show statistically significant DOW differences in ozone levels in urban fringe locations, albeit out of phase with the weekly cycling of ozone in the urban core. Our findings suggest that multiple metrics need to be used to test for the weekend effect and that the causes of DOW differences in ozone concentrations may be location specific. (c) 2006 Elsevier Ltd. All rights reserved.