AbstractView references

Among the contaminants of greatest concern, it is still possible to detect in aquatic systems "old" classics such as polychlorinated biphenyls (PCBs). Since PCBs are detected in all environmental matrices and have been identified as harmful substances due to their toxicity, persistence, and bioaccumulation in humans and wildlife, they are still one of the important groups of POPs. For this reason, this original approach studies the toxicological influence of PCBs, quantified in sediment samples collected at ten sites along the river Danube, by an application of advanced classification and clustering methods such as hierarchical cluster analysis (HCA) and Kohonen's self-organising maps (SOMs). Selected multivariate techniques were applied to the monitoring dataset in order to obtain visual images of the components distributed at each sampling site when all components are included in the classification and data projection procedure. After analyzing the data set using both techniques were isolated groups that exhibit similar behavior. In the hexagon and dendogram of variables three main clusters were distinguished. Towards the identification of pollutant spatial patterns, the SOM did not isolate a clear phenomenon probably due to the absence of local pollution sources contributing to the elevated concentrations of these compounds. The presented assumptions indicated that the supplemental application of SOM and HCA offers advantageous features over the usually rough interpretation of PCBs pattern and over the single use of the methods. © 2019 University of Novi Sad, Faculty of Technology. All rights reserved.