Polluted aquifers: identification and characterisation by statistical analysis.

Hierarchical clustering and principal component analysis applied to chemical components and physicochemical properties of well water proved to be a useful tool for identification and characterisation of aquifers. Underground water of Lerma Valley (Salta, Argentina) was examined for its physical and chemical properties by sampling 46 wells located in two adjacent areas separated by hills, one of them polluted with boron since 1991. Hierarchical clustering splits sampled sites into two main clusters, corresponding to the two areas, establishing the fact that the aquifers should be considered as two different entities in spite of their common recharge area. Values of boron concentration in the eastern area decreased in most of the wells since the pollution sources were eradicated, while four of them experienced a substantial increase, proof of the slow self-recovery of the aquifer. The use of principal component analysis provided evidence of the incipient boron pollution of the aquifer of the western area.