Multivariate and geostatistical methods in urban soil geochemistry

The aim of this study was to analyse soil concentrations of some potentially toxic elements (PTEs)in the southern Italy area of Cosenza and Rende municipalities in order to (i) define the pollutants baselinein soils, (ii) identify the most heavily polluted areas, (iii) assess the contributions of parent material andanthropogenic activity to geochemical baseline determining their different correlation structure to isolatesources of variation acting at different spatial scales. In addition, baseline data for urban areas of Cosenzaand Rende can be used to assist policy makers and legislators to draw up a better legislation with moreappropriate guideline/intervention values.In the urban and peri-urban area of Cosenza-Rende, 149 topsoil samples werecollected (0-10 cm) and analysed for 36 elements by XRF and ICP-MS. In addition, 18 samples of rocks werecollected on outcrops of the whole area and analysed by ICP-ES and ICP-MS. MultiGaussian approach wasused to map the concentrations of major oxides and several trace elements and principal component analysisand factorial kriging analysis were employed to identify the main factors influencing the PTEs spatialvariability (Wackernagel, 2003).PTEs in soil samples showed a wide range of concentrations, primarily controlled by thegeochemical composition of bedrock, with the notable exceptions of As, Pb, and Zn, whose concentrations areheavily affected by land use and anthropogenic pollution in urban areas (Guagliardi et al., 2012, 2015). Twogroups of PTEs were identified: the first one included As, Pb and Zn; and the second one Al, Co, Cr, Fe, La, Nb,Ni, Ti and V. The first group was related to anthropogenic input, while the second one was more related toparent rock composition. The regionalized factors at different scales of variability allowed to aggregate andsummarize the joint variability in the PTEs and consider the probable causes of soil pollution.The geostatistical methods were the key role, which allowed us to identify PTE hot spots (thoselarge enough to be identified at the sample density of the current survey) and areas that may pose potentialrisks to Cosenza and Rende inhabitants. Analysing and quantifying the sources of variation of PTEs acting atdifferent spatial scale and defining the spatial anomalies based on the correlation structure associated at thedifferent spatial scales, some anthropogenic and geogenic anomalous abundances are found in peri-urbanareas. Finally, these results can be used to propose remedial actions aimed at reducing health risk above allto people.