Occurrence, spatiotemporal distribution, and risk assessment of herbicides in a subtropical agricultural watershed

The extensive use of herbicides in agricultural regions has raised increasing concerns regarding their occurrence, spatiotemporal distribution, and ecological risks in watershed systems. This study investigated the occurrence, spatiotemporal distribution, and aquatic ecological risks of 16 current-use herbicides in a representative subtropical agricultural watershed in South China. Soil, sediment, and surface water samples were collected monthly throughout 2024 in the Nala watershed, Fusui County, Guangxi, China, covering wet, dry, and transitional seasons. Herbicide residues were determined using a modified QuEChERS extraction procedure coupled with UHPLC-MS/MS. All 16 target herbicides were detected in at least one environmental matrix. Diuron was frequently detected, with concentrations ranging from 0.309 to 175.406 µg kg^-1 in soil and reaching 91.292 µg kg^-1 in sediment. In surface water, maximum concentrations of ametryn and mesotrione reached 30.294 and 20.264 µg L^-1, respectively. Herbicide concentrations exhibited distinct matrix- and compound-specific seasonal patterns. Predominant herbicides in soil, particularly diuron and ametryn, showed significantly elevated concentrations in spring, whereas no significant seasonal differences were observed for the predominant herbicides in sediment. Surface-water concentrations varied significantly among seasons for several compounds but followed compound-specific patterns, with butachlor reaching its highest median concentration in summer, while ametryn and mesotrione showed numerical maxima in spring. These contrasting temporal patterns suggest that agricultural application timing influences initial herbicide inputs, whereas rainfall and hydrological processes subsequently regulate their transport, redistribution, and dilution. Spatially, higher residue levels were primarily observed in southwestern agricultural areas, while downstream redistribution was evident during the rainy season. Aquatic ecological risk assessment using the risk quotient (RQ) approach identified diuron, ametryn, and imazapic as the principal risk drivers in surface water, with maximum RQ values of 1251.23, 841.50, and 789.20, respectively. Their median RQ values remained above 1, indicating sustained ecological concerns. Several additional herbicides also exhibited RQmedian values > 1, although their risk magnitudes were lower than those of the three principal risk drivers; other compounds showed high RQmax but lower RQmedian values, indicating more episodic risk patterns. These findings demonstrate that herbicide occurrence, transport, and aquatic ecological risks in subtropical agricultural watersheds are jointly influenced by agricultural practices, compound-specific behavior, and hydrological processes. This study provides a scientific basis for targeted herbicide monitoring, runoff control, and watershed-based pollution management in subtropical agricultural regions.