Soil-sediment linkage and trace element contamination in forested/deforested areas of the Itacaiúnas River Watershed, Brazil: To what extent land-use change plays a role?

Abstract

Trace elements (TE) contamination in forested areas of the Itacaiúnas River Watershed (IRW), Brazilian Amazon, arouses growing interest owing to the rapid deforestation and mining activities. In this study, soils (surface, SS; bottom, BS) and stream sediments (SD) from forested/deforested areas of IRW were analyzed with the aim of (1) evaluating the major sources of TE (mainly As, Ba, Cd, Cu, Co, Cr, Hg, Mo, Mn, Ni, Pb, V, and Zn), and (2) examining the soil-sediment TE link related to land-use change and/or geologic factors. Compositional data analysis (CoDA) was used to eliminate data closure issues and the centred log-ratio (clr) transformation yielded better results in Principal Component Analysis (PCA). The TE distribution pattern was significantly different (p < 0.05) between forested and deforested areas, but in both areas the TE distribution pattern is significantly correlated between SS, BS, and SD, indicating a strong lithogenic control. PCA (clr-transformed) identified the major geochemical bedrock signature as Fe-Ti-V-Cu-Cr-Ni, which is nearly similar in soil and sediments. The more accentuated enrichment and the maximum number of anomalies of these elements were found in the Carajás Basin and are highly coincident with mineral deposits/local lithologies without clear indication of anthropogenic contamination from point sources. Besides geogenic factors, deforestation is also affecting TE distribution in the basin. In deforested areas, Mn was significantly enriched in the surface horizon. Furthermore, linear regression analysis shows stronger TE relationships between soils and sediments in deforested areas than in forested ones, reflecting higher erosion in the former. This could be the reason for the relatively higher enrichment of TE (e.g., Fe, Mn, Cu, Cr, Ni) in deforested sediments. The TE contamination using regional background values provides more accurate results than worldwide reference values. Thus, the former should be considered for a more realistic environmental risk assessment in IRW and other forest ecosystems in the Brazilian Amazon.