Principal Components Analysis and spatial analysis integration for enhanced assessment of pollution emission sources

Abstract

Principal Component Analysis (PCA) is a statistical technique to reduce the dimension of the original dataset. This method is often used in environmental analysis to detect the pollution sources in ambient air and in soil. However, PCA results can be difficult to read, especially when some variables seem to correlate to any principal component. In this case, it should be helpful to combine PCA with another method to better explain the uncertainty. In this study, the use of PCA coupled to Kriging algorithm has been investigated to better identify the emission sources responsible for soil pollution in an agricultural field near an industrial site in the surroundings of the city of Brescia, in Northern Italy, where organic and inorganic micropollutants concentrations in soils are beyond the legal limits. Three clusters of variables have determined as many as the emission sources in the investigated area. PCs condensed the information from most heavy metals and organic compounds. It has been demonstrated that the spatial assessment can be useful to find emission sources for those elements where it has not been possible by PCA analysis only.