Prediction of organic contaminant uptake by plants: Modified partition-limited model based on a sequential ultrasonic...

Predicting the translocation of organic contaminants to plants is crucial to ensure the quality of agricultural goods and assess the risk of human exposure through the food web. In this study, the performance of a modified plant uptake model was evaluated considering a number of chemicals, such as polycyclic aromatic hydrocarbons (PAHs), organochlorine pesticides (OCPs) and polybrominated diphenyl ethers (PBDEs), with a range of physicochemical properties; different plant species (Ipomoea aquatica Forsk (swamp morning glory), Chrysanthemum coronarium L. (crown daisy), Zea mays L. (corn), Brassica rapa pekinensis (Chinese cabbage), Cucurbita moschata (pumpkin), Raphanus sativus L. (radish), Spinacia oleracea L. (spinach) and Capsicum annuum L. (pepper)); and different types of soil (paddy soil, laterite soil and black soil). The biases of predictions from a previously used partition-limited model were -76.4% to -99.9% relative to the measured concentrations. An overall transmission factor (α_tf=0.39), calculated from a linear regression of the measured bioavailable fraction (C_bio) and the total concentration in plants, was considered a crucial modification and was included in the modified model. C_bio was found to better represent the chemical content available in soil for root uptake. The results from this study improve the accuracy of predictions for vegetation-uptake assessments by modifying the partition-limited model and then validating the modified model using comparisons between predicted data and measured values. The accuracy of the concentrations of organic contaminants in plants improved: when using the modified model, 89.5% of the predictions were within 40% of the actual value. The average bias was limited to 1.5%-30.5%. The model showed great potential to predict plant uptake using the bioavailable fraction concentration in soil.