Analysis and assessment of trace element contamination in offshore sediments of the Augusta Bay (SE Sicily): A multivariate statistical approach based on canonical correlation analysis and mixture density estimation approach

An application of multivariate statistical methods is provided to identify anthropogenic contaminants and lithogenic elements in offshore sediments collected near the heavily industrialized Augusta Bay, Sicily. An exploratory statistical technique, based on canonical correlation analysis (CCA) and mixture density estimation approach, is used for distinguishing between natural and anthropogenic contributions of trace elements in the investigated sediments. Following the intensive industrialization of Augusta area, marine sediments reveal the severe impact of local anthropogenic activities for many elements (e.g. As, Cd, Hg, Pb, and Sb), which are considered very dangerous for the environment. In contrast, a prevalent natural/lithogenic origin can be ascribed to Co, Cr, Cu, Mn, Ni, V, and Zn. Themultivariate statistical approach used in this study,which aims to estimate and to classify trace element pre-industrial levels and sources, could be easily applied to other different contaminated areas.