A predictive model for arsenic accumulation in rice grains based on bioavailable arsenic and soil characteristics

Arsenic (As) is a well-known human carcinogen, and rice consumption is the main way Chinese people are exposed to As. In this study, 14 kinds of paddy soils were collected from the main rice-producing areas in China. The results showed that rice roots and leaves accumulated more As than stems and grains in the following sequence: As_root> As_leaf> As_stem> As_grain. The accumulation of As by rice grains mainly depends on the total As and bioavailable As (0.43 mol/L HNO₃ extractable As), which explained 32.2% and 22.2% of the variation in the grain As, respectively. In addition, soil pH, organic matter (OM) and clay contents were the major factors affecting grain As, explaining 13.1%, 7.9% and 5.3% of the variation, respectively. An effective prediction model was established via multiple linear regression as As_grain= 0.024 BAs - 0.225 pH+ 0.013 OM+ 0.648 EC - 0.320 TN - 0.088 TP - 0.002 AS+ 2.157 (R² =0.68, P < 0.01). Through the verification of the samples from both pot experiments and paddy fields, the model successfully provided accurate predictions for rice grain As.