Probabilistic quality standards for heavy metals in soil derived from quality standards in crops

This paper contributes to the development of critical levels of heavy metals in soil at which functions of the soil are safeguarded. For agriculture production, one of the key aspects is food safety. In this study, a quality standard for Cd in soil is derived from the EC-food quality standard for Cd in wheat. The transfer of the Cd in the soil to the crop is modelled by a regression model. Such models offer the possibility of calculating probabilistic soil quality standards from food quality standards. Given the food quality standard and a maximum acceptable probability of exceeding this standard, the Cd concentrations in soil at which the probability of excess Cd in wheat equals this maximum is calculated by the inverse use of the regression model. It is important to account for errors in the Cd measurements in wheat used to calibrate the model. Neglecting these errors gives conservative estimates of the p% critical threshold for p smaller than 50. For representative arable soils, the 5% critical threshold varies from 0.5 mg kg(-1) (non-calcareous sand soils) to 3 mg kg-1 (calcareous clay soils). Even at low probability levels, the model must be extrapolated for many arable soils, and therefore, more calibration data must be collected, especially from sand soils with relatively low pH values. Validation of the model with data from contaminated bed sediments in the floodplain of the Meuse showed that the model seriously overestimated the critical threshold, which can be explained by the higher availability of Cd in contaminated soils. (c) 2005 Published by Elsevier B.V