From a multivariate spatio-temporal array to a multipollutant - multisite Air Quality Index

Abstract

AQIs are computed on air pollution data that are usually collected according to time, space and type of pollutant: in a given town/region, data consisting of hourly levels of K pollutants recorded in S monitoring sites, are usually organized in a three-mode array. A first aggregation step usually concerns time, and allows to pass from hourly data to a daily synthesis: in this paper data will be aggregated by time according to the guidelines provided by the national agencies producing the three mode array X. Here we will propose a new approach to get a Multipollutant-Multisite Air Quality Index time series from a multivariate spatio-temporal array. This implies a two step aggregation, according to space and to pollutant. In order to find a spatial synthesis of pollutant time series a three mode factor analysis will be used, obtaining AQIk; the AQI (I2), proposed by Ruggieri et al. (2009) with respect to a single monitoring site, will be then applied to represent the global air pollution situation on a given day.