Potentially toxic elements (PTEs) distribution and main geochemical processes in Sabato river catchment basin (southern Italy)

The catchment basin of Sabato River (with an area of ca. 459 km²) in Southern Italy has been historically characterized by the presence of urban settlements and industrial development areas along its course (with a total length of 50 km). A total of 35 stream sediment samples was collected in 2019 from the main course of the river right after any tributaries inlet and upstream and downstream of the main industrial and urbanized areas within the basin. After their collection, samples were oven-dried (temperature was kept below 37°C to preserve Hg content) and sieved (< 200 mesh) at the Environmental Geochemistry Laboratory (EGL) of University of Napoli Federico II; an aliquot of each sample was sent to Bureau Veritas Laboratories (Vancouver, Canada) for chemical analysis. The concentrations of 53 elements were determined through a combination of Inductively Coupled Plasma Mass Spectrometry (ICP-MS) and Inductively Coupled Plasma Emission Spectroscopy (ICP-ES) following a modified aqua regia digestion. Grain size analysis was also performed at Nanjing University (China). For the purposes of this study, geochemical data have been processed according to Compositional Data Analysis fundamentals and the dataset has undergone a centred log-ratio (clr) transformation due to its capability of preserving both geometric properties and one-to-one relationship within the variables. Both raw chemical data and clr values of a subset of potentially toxic elements (PTEs – As, Cd, Cr, Cu, Hg, Pb, Zn) has been plotted, from source to sink (S2S). Geochemical data variations along the river course were also compared with clay/silt ratio and sedimentation transport index (STI) in order to recognize any geomorphological or hydrodynamic influence on them. For each element of interest, raw concentrations have been compared with the Upper Baseline Limit (UBL) estimated by Albanese et al. (2007) for Campania region to determine potential anomalies. Furthermore, since in literature there are no reference values for log-ratio transformed data, the anomaly threshold for clr-transformed data have been established applying the approach proposed by Reimann et al. (2005), which replaces the arithmetic mean with the median and the standard deviations with the median absolute deviation (MAD). These estimators are strongly suggested as they are robust against extreme values. Among PTEs, As, Cd, Cu and Zn showed some scattered enrichments along the river course suggesting the influence of some anthropic sources (urbanized and agricultural areas) on chemical composition of sediments. To identify the main natural and anthropogenic processes active in the Sabato catchment, a Principal Factor Analysis (PFA) on clr-transformed data, according to the procedure proposed by Filzmoser et al. (2009), has been performed on a selection of 11 variables (i.e. elements) characterized by a larger count of high correlations (> .80) within the dataset. Two factors (as associations of elements) have been determined: the first factor (including Be, Ce, Al, As, La, Cs and Tl) has been related to the presence of pyroclastic materials spread all along the catchment; the second factor (including Na and K with Ni and Cr in antithesis), has been assumed as a proxy for the grain size variations: positive scores are indicative of a mayor abundance of the coarser sediments and negative scores locate samples with a high content of finest fraction.