Spatial analysis of compositional data using a spectrum-area multifractal model in Campania region (Italy)

In this study, the spatial distributions of Mn, Fe, Se and Hg in Campania region (Southern Italy) were explored. The study area covers 13,595 km2 on a 16 km2 grid in the suburban and agricultural areas and on a 4 km2 grid in... [ view full abstract ]

In this study, the spatial distributions of Mn, Fe, Se and Hg in Campania region (Southern Italy) were explored. The study area covers 13,595 km2 on a 16 km2 grid in the suburban and agricultural areas and on a 4 km2 grid in urban areas. A total of 3,535 top soil samples were collected and analysed by ICP-MS after an aqua regia digestion. The isometric logratio transformation was applied to deal with the closure effects of geochemical data prior to statistical analyses. A spectrum-area (S-A) multifractal model was used to decompose the interpolated maps and to identify background and anomalous components. The results demonstrate that (1) strong anomalies of Fe mainly occur in the pyroclastic coverings of the Somma-Vesuvius volcanic district and the Apennine margin, and the spatial distribution characteristics of Fe in Phlegrean district is different from the Somma-Vesuvius Volcanic district, probably due to the different eruptive history; (2) sedimentary lithologies host widespread Mn anomalies, and Fe and Mn distribution is mostly controlled by geogenic factors; (3) selenium data show a strong correlation with the annual rainfall, and high Se concentration values (> 2 mg/kg) occur in correspondence of the drainage channels network known as “Regi Lagni” affected by urban and industrial wastewater; (4) Hg spatial distribution highlights anomalies (> 1 mg/kg) where urban areas and highway arterial roads occur. These results suggest that the S-A model is a powerful tool to decompose the mixed geochemical patterns and this approach coherent with composition increases information quality.