RESUMO
Accurate data of cadmium (Cd) and uranium (U) leaching are needed in the context of identifying their mass balances in agricultural soils. There is some controversy related to sampling methods and the contribution of colloid facilitated transport. Here, leaching was measured in undisturbed unsaturated soils and the impact of colloids was measured with due attention to solution sampling protocols. Soils were sampled in an arable, pH neutral silty loam soil. The columns (n = 8) were irrigated and PTFE suction plates (1 µm pores) at the bottom ensured unsaturated flow. New here is that both percolates and associated suction plates were collected, the elements in the plates were recovered with acid digestion and used as a lower estimate of colloidal forms. The fraction of elements collected in the plates were 33 % (Cd) and 80 % (U) of the total mobility (=percolates + plates), illustrating colloidal transport. Composition of pore water extracted by soil centrifugation varied largely between initial and final samples and showed that colloids increased as a result of reduced solution calcium after leaching two pore volumes with low calcium water. Flow Field-Flow Fractionation (FIFFF) of the pore water and percolates revealed co-elution of U with colloidal organic matter, oxyhydroxides and clay, illustrating colloidal transport of U by these vectors. Colloidal transport of Cd was less pronounced and was dominated by organic matter. Soil extracts with 0.01 M CaCl2 have lower colloid concentration and consequently underestimate mobile U. In contrast, Cd concentrations in 0.01 M CaCl2 extracts exceed that of percolates due to chloride complexation and higher calcium, mobilizing Cd. Soil leaching experiments better indicate potential leaching losses than a single pore water composition because the former yields the time integrated data. Suction plates and/or bottom filters need to be analysed in leaching studies to account for metal transport by colloids.