Estimating Exceedance Probability in Air Pollution Time Series

In the last years, increasing attention has been paid to air pollution, due to its impact on human health and on the environment. Current EU legislation establishes fixed limits for some air components that have been shown to have adverse effects on human health. It is therefore important to identify regions where the probability of exceeding those limits is high. In this paper, we propose a bootstrap scheme to obtain the distribution of the considered air pollutant at a given time point. In particular, the proposed resampling scheme is based on the residuals of a semiparametric model which is able to incorporate some stylized facts usually observed in such kind of data, such as missing data, trends and conditional heteroscedasticity. The estimated bootstrap distribution is then used to estimate the probability that the air pollutant exceeds the fixed legal limits. An application to (PM10) in Torino area in the North-Italian region Piemonte is shown.