Prediction of the distribution coefficients of metals in Amazonian soils.

Soil pollution with metals alters ecosystem structure and function, as well as poses risks to human health. In this study the distribution coefficients (Kd) were determined for Cd, Co, Cr, Cu, Hg, Ni, Pb, and Zn in 21 soils from eastern Amazonian via Freundlich adsorption isotherm. The stepwise multiple regression models were used to estimate the Kd values through the combined effect of soil attributes. Multiple regressions showed that Kd values were strongly correlated with the soil pH, CEC and Fe oxide contents. A close association between Kd of metals and soil properties was observed using three principal components (PCA) and Pearson's correlation. In the principal components analysis showed that concentration of metals was related to pH, Fe and Mn oxides contents, which in combination explained 90 percent of the variation. This probably happens because of greater affinity of those metals to Fe and Mn oxides formed predominantly under tropical conditions.