A Multisite-Multipollutant Air Quality Index

In this paper, starting from a multivariate spatio-temporal array, containing air pollution data collected for the main pollutants at different monitoring sites over a 1-year period, a new approach is proposed to get a Multipollutant-Multisite Air Quality Index (AQI) time series. A two steps aggregation, related to space and to pollutants, is considered. For the first aggregation (spatial synthesis) a PCA is performed on data array opportunely rearranged, while the index I2, proposed in Ruggieri and Plaia (2011), is used for the second aggregation (pollutant synthesis), obtaining the new index I2MS. Daily data of four air pollutants from the city of Palermo (Italy) are analyzed to test the performance of the new index. The index I2MS overcomes the main issues presented by other indices, many of which based on AQI computed by US EPA, considering the highest pollutant concentration adequately standardized. The comparison carried out shows that the index here proposed has a better performance than the synthesis by median-maximum (Bruno and Cocchi, 2002) and the synthesis by PCA along both space and pollutant dimensions, if the conjoint effect on air quality of the different pollutants has to be taken into account.

► We propose a Multipollutant-Multisite Air Quality Index. ► We take into account the adverse effects due to the coexistence of all the pollutants. ► The methodology is based on an asymmetrical approach to cope with the three-mode array synthesis. ► The index behavior along the year is free from ambiguity and eclipsicity problems.