Multivariate statistical analyses of artificial radionuclides and heavy metals contaminations in deep mud of Keban Dam Lake, Turkey

Abstract

On the basis of the artificial radioactive and heavy metal compositions, factor and cluster analyses are employed to identify the inter-relationship among different variables and their similarity groups. In this paper, 15 physico-chemical variables, including the activities of (CS)-C-137, Sr-90, total alpha, total beta, and concentrations of Fe, Mg, Ca, K, Na, Zn, Cu, Cr, Co, Ni, and Mn, are used for the application of the proposed methodologies. The spatio-temporal samples of these variables are collected from deep mud at about 30-35m mean depth in the reservoir of Keban Dam Lake, which is located in the eastern part of Turkey. Spatially, there are 20 sampling sites at the dam with 150 km(2) lake surface area where samples were taken in 2006. The lake is affected by man-made and industrial influxes. The application of the factor and cluster analysis methods yields that the former method reduces the number of variables into six factors with 77.2% variance explanation whereas the latter yields three distinctive groups of the same variables. (C) 2007 Elsevier Ltd. All rights reserved.