Use of multivariate statistical techniques to determine the source apportionment of heavy metals in soils and sediments

Abstract Soil quality can be classified as the multiphase system of the Earth’s surface including the function of living organisms. Some of the human activities correlated to the exploitation of minerals can lead to soil pollution and degradation. Furthermore, an extensive approach for soil and sediment monitoring can be evaluated using various statistical approaches. Over the past decade, different multivariate approaches in different affected areas have been evaluated continuously, in order to characterize any abnormal changes in elemental distribution in automorphic and alluvial soil samples. Therefore monitoring activities have been implemented in a very specifically affected area (with mining and ore processing activities). This area has been monitored for the soil contents of 69 elements: Ag, As, Al, Au, B, Ba, Be, Bi, Br, Ca, Cd, Ce, Co, Cr, Cs, Cu, Dy, Er, Eu, Fe, Ga, Gd, Ge, Hf, Hg, Ho, I, In, Ir, K, La, Li, Lu, Mg, Mn, Mo, Na, Nb, Nd, Ni, Os, P, Pb, Pd, Pr, Pt, Rb, Re, Rh, Ru, Sb, Sc, Se, Sm, Sn, Sr, Ta, Tb, Te, Ti, Th, Tl, Tm, V, W, Y, Yb, Zn, and Zr. Multielement contents were determined in top/bottom samples of automorphic and alluvial soil. Basic descriptive statistics of the data set was applied for primary data processing. A bivariate statistics methodology was applied for elements content correlations. Scatter plots were applied in order to examine the relations between two data sets. Multivariate statistics was also used in order to extract the geochemical associations of the dominant chemical elements distributed in the surface soil. Factor analysis was applied to the data sets, which were standardized to a zero mean. The main goal of the applied processing techniques was to extract the dominant marker groups. Box-Cox data transformation was applied for the total data set. Statistical data processing software Statistica 11.0 was applied for all of the mentioned processing methods.