Evaluation of plant availability of soil trace metals by chemical fractionation and multiple regression analysis

Soil samples with a range of chemical and physical properties were collected from 10 different rural regions of China. Trace metals (Ni, Co, Cu, and Pb) in the soils were partitioned by a sequential extraction procedure into Mg(NO(3))(2) extractable (F1), CH(3)COONa extractable (F2), NH(2)OH.HCl extractable (F3), HNO(3)?H(2)O(2) extractable (F4), and residual (F5) fractions. Chemical fractionation showed that F1 fraction of the metals was less than 1% and residue was the dominant form for Cu and Ni in all samples, and for Co in most of the samples. Significant interrelationships of the fractions varied considerably with the different metals. Winter wheat (Triticum aestivum L.) and alfalfa (Medicago sativa L.) had been grown on the soils in a pot-culture experiment under greenhouse conditions for 40 days. Metal availability to the plants was evaluated by simple and multiple regression analysis. The Mg(NO(3))(2) extractable Co (F1) was significantly correlated with Co concentrations in different parts of wheat and in the whole of alfalfa. For the other metals, the independent variables of the multiple regression models, chosen by stepwise selection, were given as: F1 and F2 + F3 + F4 for Ni; F1, F2 + F3 and F4 for Cu; and F3 + F4 for Pb. The results of this study demonstrate that the sequential extraction procedure, in conjunction with multiple regression models using a combination of correlated fractions as an independent variable, may be useful for the prediction of plant absorption of trace metals in soils.