Source apportionment for contaminated soils using multivariate statistical methods

The application of statistical techniques for the recognition and identification of contamination sources has become an increasingly important tool. The chemical compositions of soil samples collected in the Puchuncaví Valley (Chile) provide a dataset suitable for the application of source apportionment techniques such as positive matrix factorization (PMF) and principal component analysis (PCA) with varimax rotation. PMF allowed the identification of the chemical profile and the relative contribution of three interpretable factors related to three contamination sources. Combining these results with a PCA analysis successfully showed that the main source of pollution emits Cu, Zn, As, Se, Mo, Sn, Sb and Pb. Therefore, the use of source profiles for contaminated soils shows much promise both for incorporating well-established knowledge about pollution sources and as a tool for incremental, exploratory data analysis.