| Abstract: | Soil‐plant transfer of metals is a nonlinear process. We therefore aimed at evaluating the potential of Freundlich‐type functions (cPlant = b × cSoila) to predict Cd, Cu, Pb, and Zn concentrations in wheat (Triticum aestivum L.) grain and leaf (cPlant) from soil concentrations (cSoil). Wheat plants and soil A horizons, mainly developed from Holocene sediments, were sampled at 54 agricultural sites in Slovakia. Metals were extracted from soils with 0.025 M EDTA at pH 4.6 and concentrated HNO3/HClO4 (3:1); plant samples were digested with concentrated HNO3. Total metal concentrations of soil samples were 0.07—25 mg Cd kg—1, 9.3—220 mg Cu kg—1, 14—1827 mg Pb kg—1, and 34—1454 mg Zn kg—1. On average, between 20 % (Zn) and 80 % (Cd) of the total concentrations were EDTA‐extractable. The total metal concentrations of grain samples were < 0.01—1.3 mg Cd kg—1, 1.3—6.6 mg Cu kg—1, < 0.05—0.30 mg Pb kg—1, and 8—104 mg Zn kg—1. The leaves contained up to 3.2 mg Cd kg—1, 111 mg Cu kg—1, 4.3 mg Pb kg—1, and 177 mg Zn kg—1. Linear regression without data transformation was precluded because of the nonnormal data distribution. The Freundlich‐type function was suitable to predict Cd (grain: r = 0.71, leaf: 0.86 for the log‐transformed data) and Zn concentrations (grain: 0.69, leaf: 0.68) in wheat grain and leaf from the EDTA‐extractable metal concentrations. The prediction of Cu and Pb concentrations in grain (Cu: r = 0.44, Pb: 0.41) was poorer and in leaf only possible for Pb (0.50). We suggest to use the Freundlich‐type function for defining threshold values instead of linear regression because it is more appropriate to simulate the nonlinear uptake processes and because it offers interpretation potential. The results suggest that the coefficient b of the Freundlich‐type function depends on the intensity of metal uptake, while the coefficient a reflects the plants' capability to control the heavy metal uptake. The latter is also sensitive to metal translocation in plants and atmospheric deposition.<?show $6#> |
