Stricter cadmium and lead standards needed for organic fertilizers in China

This study aims to evaluate the adequacy of China's national standards for heavy metals in organic fertilizers by predicting their concentrations in grains using machine leaning. A comprehensive dataset was collected from literature, including soil properties, organic fertilizer application rates, and heavy metal concentrations in soils, fertilizers, and grains. Ten key variables were used as inputs for machine learning models developed via H2O AutoML, to predict heavy metal concentrations in maize, wheat, and rice grains across China. The models demonstrated strong predictive performance, with R² values ranging from 0.74 to 0.95. Predictions were generated using the national standard limits for heavy metals in organic fertilizers, along with geographical feature maps of China and the pre-trained machine learning models, to assess compliance with food safety standards. The results indicated that the current national standards for As, Cr, and Hg in organic fertilizers are adequate, as the heavy metal concentrations in grains remained within safe limits nationwide. However, for Cd and Pb, even when organic fertilizers met the national standard limits, the predicted concentrations in grains exceeded safety standards in 39.1-63.6 % and 81.2-95.5 % of agricultural areas, respectively. These findings suggest that the existing national standards for Cd and Pb in organic fertilizers are too lenient to ensure food safety in China. Stricter regulations are therefore necessary to reduce the risk of heavy metal contamination in staple crops and to safeguard public health.