CHARACTERISING THE TEMPORAL VARIATIONS OF GROUND-LEVEL OZONE AND ITS RELATIONSHIP WITH TRAFFIC-RELATED AIR POLLUTANTS IN THE UNITED KINGDOM: A QUANTILE REGRESSION APPROACH
Free (open access)
Volume 9 (2014), Issue 1
29 - 41
S. MUNIR, H. CHEN & K. ROPKINS
Ground-level ozone is a secondary air pollutant and is photochemically produced by solar radiation from the reaction of volatile organic compounds (VOCs) and nitrogen oxides (NOx). Ground-level ozone is considered a harmful pollutant due to its adverse impact on human health, agricultural crops and materials. The concerning factor is that in spite of decreasing trends in some other air pollutants (e.g. NOx), ozone concentrations are still increasing. This paper describes the temporal variations of ozone at four air quality monitoring sites (Harwell, Leeds, Marylebone and Strath Vaich) in the United Kingdom for the year of 2008. The association of ozone with some traffic-related air pollutants has been explored applying a quantile regression model (QRM). The traffic-related air pollutants considered as predictors for this study are hydrocarbons (HC), nitric oxides (NO), nitrogen dioxides (NO2), carbon monoxides (CO) and particulate matter (PM2.5). QRM can handle the non-linearities in the relationship of ozone and its predictors and is applicable to non-normal air quality data distribution. The behaviour and interaction of ozone with its predictors vary at different regimes of ozone distributions, which remains hidden when applying an ordinary least square regression model. QRM explains significantly more variations in ozone concentrations (global goodness of fit R1 = 0.88) as compared to ordinary least square regression (coefficient of determination R2 = 0.32) and is therefore better suited for ozone data analysis and prediction.
Air pollution, ground-level ozone, nitrogen oxides, ozone variations, quantile regression