Unveiling predominant and latent factors influencing the geochemical variability in a river catchment: A statistical and visual approach

Abstract

The Sabato River catchment basin in Southern Italy was investigated to assess and understand geochemical variations affecting potentially toxic elements along its course. For this purpose, thirty-five stream sediment samples were collected in the fall of 2018 from the river's main course. The individual composition found at each collection point was associated with the relative sample catchment basin (SCB) before undergoing a compositional data analysis (CoDA). Both raw and centred log-ratio (clr) transformed data were plotted against sediment granulometric and basin morphological features to reveal any dependence of geochemistry on landforms and catchment hydrodynamic settings. Geochemical Upper Baseline Limits (UBLs) were estimated for the basin, using raw data to discriminate local anomalies. In contrast, clr-data were compared with a proposed anomaly threshold (AT) for the same purpose.

A Robust Principal Factor Analysis (RPFA) was conducted on the clr-transformed data to identify natural and anthropogenic processes influencing the sediments of the Sabato River. The RPFA revealed that the predominant geochemical processes in the studied catchment basin were natural, with the weathering of pyroclastic covers and variations in grain size as key factors. The compositional balance technique, based on isometric log-ratio (ilr), was used to explore the results and emphasise one factor's predominance over another. A new thematic map, “Rainbow map”, was finally proposed to show the spatial dominance of one process over the others or their concurrent co-dominance.