A probabilistic air quality indicator for large cities

Air quality is a factor increasingly relevant and considered in the society. And, this is not only because of its effects in human health, vegetation, material elements, etc, but also because it is becoming a core factor in a variety of people decisions, including the choice of the place of residence or of the city where to enjoy some days of leisure. It is not difficult to find literature about air quality indexes, but the currently used indexes neither include the mixed effects on air quality of most of pollutants nor the combined effects of pollution and atmosphere characteristics. In this paper we use sixteen environmental variables divided in two groups: a) pollution variables: NOx, NO2, SO2, PM10, O3, CO, and b) atmospheric indicators related to te velocity and direction of wind, number of precipitation days, precipitation volume, maximum and minimum temperature, number of sun hours each day, etc.This is why this article prosecutes the following aims: (i) the elaboration of a Global Index of Air Deterioration for big cities, based on factorial analysis, that takes into account some atmospheric variables and traffic; (ii) the construction of a quantitative Air Quality Index obtained by using probabilistic methods. A practical application has been carried out for one of the largest cities in Europe: Madrid, which has a high quality monitoring net. In our empirical case study, we use a massive dataset, including registers corresponding to the temporal range 2000-2010 on a monthly basis. These registers refers to the most dangerous pollutants as well as the main important atmospheric variables.