CONICET | Buscador de Institutos y Recursos Humanos

Eleven South American and Caribbean countries participate in an International Atomic Energy Agency (IAEA) project (RLA/1/010 Improvement of contamination management of surface water bodies contaminated with heavy metals) to establish environmental basic procedures related to the management of surface-water resources. The procedures embrace all of the activities, from sampling design to interpretations of data analysis. As part of this general project, an additional bilateral cooperation project (SLO/09/12) has been established between Slovenia (MHEST) and Argentina (MINCyT) in order to apply chemometric methods of analysis to information recovered at the site of study in Argentina: the Reconquista River in Buenos Aires Province. This work is performed for a data base that consists of 17 campaigns carried out along 10 years, 1996-2006. Each campaign involved measurements of 26 physical-chemistry variables on 15 sample points distributed along a sector of the river and its affluent channels. The original data set is a matrix of 290 objects row and 26 columns of variables. In this presentation it is shown conclusions obtained by application of statistical and computational methods of analysis, like Multivariate Curve Resolution (MCR), Canonical Correlations (CC) and Counter Propagation Artificial Neural Networks (CPANN). MCR lets us to discriminate pollutants coming from different sources or group of sources and its impact on different sampled points. By CC it is possible to recognize certain correlation between groups of variables. Furthermore, CPANN could find correlations to predict the environmental impact of the channels along the river, but this analysis needs to be carried out for each variable individually and it is actually in course.