Cluster Analysis as a Tool in the Study of Groundwater Quality

Abstract

A study of the vulnerability of the river Tenes aquifer has been carried out by means of cluster analysis. On the basis of eleven features measured for forty-seven ground-water samples, pattern recognition techniques allow the visualization of several types of waters. Moreover, the natural differences occurring between manual and pump sampling procedures have arisen. Alkalinity and pH and to a lesser extent sulphates and conductivity have been identified as the features which contribute to the differentiation of the water samples according to the sampling procedure. The withdrawal from the chemometric analysis of these variables leads to the presence of a sole type of water quality in which those samples that show the influence of the surface pollution can be distinguished.