Bioaccessibility of heavy metals in soils cannot be predicted by a single model in two adjacent areas.

Abstract

The objective of this study was to examine whether a single model could be used to predict the bioaccessibility of heavy metals in soils in two adjacent areas and to determine the feasibility of using existing data sets of total metal concentrations and soil property parameters (e.g., pH, total organic carbon, and soil texture) when predicting heavy metal bioaccessibility. A total of 103 topsoil samples were collected from two adjacent areas (Baotou and Bayan Obo). A total of 76 samples were collected from Baotou, and 27 were collected from Bayan Obo. The total and bioaccessible concentrations of arsenic (As), copper (Cu), lead (Pb), and zinc (Zn) were measured following complete composite acid digestion and a simple bioaccessibility extraction test. The average total concentrations of As, Cu, Pb, and Zn were 8.95, 27.53, 28.40, and 79.50mg/kg, respectively, in Baotou and 18.12, 30.75, 38.09, and 87.62mg/kg in Bayan Obo. Except for As, these values were similar in both areas. The average bioaccessible heavy metal concentrations (Bio-HMs) for each target HM were also similar. In Baotou, the average Bio-HM values for As, Cu, Pb, and Zn were 1.16, 3.76, 16.31, and 16.10mg/kg, respectively, and 1.26, 2.51, 14.31, and 8.68mg/kg in Bayan Obo. However, the relative bioaccessibilities for each HM in Baotou were greater than those in Bayan Obo, with mean values for Pb, Zn, Cu, and As of 57, 20, 17, and 12%, respectively, in Baotou and 40, 11, 9, and 8% in Bayan Obo. In both areas, prediction models were successfully created using heavy metal concentrations and soil physicochemical parameters; however, models of the same target element differed between the areas, which indicated that a common model for both sites does not exist. Bio-HMs were highly affected by soil properties, which were found to differ between the adjacent areas. In addition, soil properties with large variations played major roles in the predictive models. This study highlights the importance of incorporating physical and chemical parameters that vary greatly when building predictive models of heavy metal bioaccessibility in soil. A similarity in soil properties between areas might be a prerequisite for the creation of a common predictive model for soil Bio-HMs.