Derivation and validation of thresholds of cadmium, chromium, lead, mercury and arsenic for safe rice production in paddy soil

Cadmium (Cd), chromium (Cr), lead (Pb), mercury (Hg) and arsenic (As) are potent toxicants to human health via dietary intake. It is imperative to establish accurate soil thresholds based on soil-plant transfer models and food safety standards for safe agricultural production. This study takes rice genotypes and soil properties into account to derive soil thresholds for five heavy metal(loid)s using the bioconcentration factors (BCF) and species sensitivity distribution (SSD) based on the food safety standard. The BCF generated from two paddy soils was calculated to investigate the sensitivity of heavy metal accumulation in nine rice cultivars in a greenhouse pot experiment. Then, empirical soil-plant transfer models were developed from a middle-sensitivity rice cultivar (Denong 2000, one selected from nine rice) grown in nineteen paddy soils with various soil properties under a proper exogenously metal(loid)s concentration gradient. After normalization, hazardous concentrations from the fifth percentile (HC5) were calculated from the SSD curves, and the derived soil thresholds were obtained from HC5 prediction models that based on the combination of pH and organic carbon (OC) or cation exchange capacity (CEC). The soil Cd threshold derived based on pH and organic carbon (pH 6, CEC ≥ 20 cmolc kg−1) was 3.1 times lower than the current threshold. The derived thresholds for five elements were validated to be reliable through literature data and field experiments. The results suggested that deriving soil heavy metal(loid)s threshold using SSD method and local food safety standards is feasible and also applicable to other crops as well as other regions with potential health risks of toxic elements contamination in agricultural production. [Display omitted] •The heavy metal(loid) thresholds were derived from rice cultivars and soil properties.•The sensitivity of nine rice cultivars was investigated by the SSD curves.•Soil pH, OC and CEC were used to develop the HC5 prediction models.•A method for deriving soil heavy metal thresholds was proposed on a national scale.