Multivariate Model for Predicting Crop Fluorine Content Based on Soil Physicochemical Properties and Fluorine Speciation

Fluoride (F), an essential trace element with dual health implications, poses significant ecological risks when its biogeochemical cycle is disrupted in soil-crop systems. This study addresses the limitations of total soil F content in assessing crop F bioavailability through a systematic investigation of 30 paired agricultural soil-wheat samples from Zhouzhi County, Guanzhong Plain. Sequential extraction revealed residual F (Fre: 456.83-932.12 mg kg⁻¹) as the dominant speciation (79.5-99.04% of total F), significantly restricting its environmental mobility. While surface soils exhibited elevated total F (753.3 mg kg⁻¹ mean; 1.6× national baseline), wheat grains maintained safe F levels (1.2 mg kg⁻¹ mean; <30% of national food safety thresholds). A robust multivariate model (R²=0.72) integrating water-soluble F (Fws), CaO, Mn, Alox-Feox complexes, P, and pH effectively predicted wheat F accumulation, with Fws contributing the biggest to total variance. These findings establish a predictive framework for F transfer mitigation and provide scientific support for green agricultural practices in fluoride-affected regions.