Combining regression analysis and air quality modelling to predict benzene concentration levels

State of the art epidemiological research has found consistent associations between traffic-related air pollution and various outcomes, such as respiratory symptoms and premature mortality. However, many urban areas are characterised by the absence of the necessary monitoring infrastructure, especially for benzene (C6H6), which is a known human carcinogen. The use of environmental statistics combined with air quality modelling can be of vital importance in order to assess air quality levels of traffic-related pollutants in an urban area in the case where there are no available measurements. This paper aims at developing and presenting a reliable approach, in order to forecast C6H6 levels in urban environments, demonstrated for Thessaloniki, Greece. Multiple stepwise regression analysis is used and a strong statistical relationship is detected between C6H6 and CO. The adopted regression model is validated in order to depict its applicability and representativeness. The presented results demonstrate that the adopted approach is capable of capturing C6H6 concentration trends and should be considered as complementary to air quality monitoring.