Characterization of background air pollution exposure in urban environments using a metric based on Hidden Markov Models

Abstract

Abstract Urban area air pollution results from local air pollutants (from different sources) and horizontal transport (background pollution). Understanding urban air pollution background (lowest) concentration profiles is key in population exposure assessment and epidemiological studies. To this end, air pollution registered at background monitoring sites is studied, but background pollution levels are given as the average of the air pollutant concentrations measured at these sites over long periods of time. This short communication shows how a metric based on Hidden Markov Models (HMMs) can characterise the air pollutant background concentration profiles. HMMs were applied to daily average concentrations of CO, NO2, PM10 and SO2 at thirteen urban monitoring sites from three cities from 2010 to 2013. Using the proposed metric, the mean values of background and ambient air pollution registered at these sites for these primary pollutants were estimated and the ratio of ambient to background air pollution and the difference between them were studied. The ratio indicator for the studied air pollutants during the four-year study sets the background air pollution at 48%–69% of the ambient air pollution, while the difference between these values ranges from 101 to 193 μg/m3, 7–12 μg/m3, 11–13 μg/m3 and 2–3 μg/m3 for CO, NO2, PM10 and SO2, respectively.