Although the health effects of ambient ozone have been widely assessed, their tempo-spatial variations remain unclear. We selected 20 communities (ten each from southern and northern USA) based on the US National Morbidity, Mortality, and Air Pollution Study (NMMAPS) dataset. A generalized linear model (GLM) was used to estimate the season-specific association between each 10 ppb (lag0-2 day average) increment in daily 8 h maximum ozone concentration and mortality in every community. The results showed that in the southern communities, a 10 ppb increment in ozone was linked to an increment of mortality of -0.07%, -0.17%, 0.40% and 0.27% in spring, summer, autumn and winter, respectively. For the northern communities, the excess risks (ERs) were 0.74%, 1.21%, 0.52% and -0.65% in the spring, summer, autumn and winter seasons, respectively. City-specific ozone-related mortality effects were positively related with latitude, but negatively related with seasonal average temperature in the spring, summer and autumn seasons. However, a reverse relationship was found in the winter. We concluded that there were different seasonal patterns of ozone effects on mortality between southern and northern US communities. Latitude and seasonal average temperature were identified as modifiers of the ambient ozone-related mortality risks.
Keywords: USA; air pollution; ambient ozone; mortality; seasonal variation; time-series study.