Iron isotope constraints on particulate iron dynamics in a human-impacted subtropical watershed.

Riverine particulate iron (Fe) mediates nutrient transport and redox-sensitive biogeochemical processes from watersheds to coastal waters, yet quantitative separation of natural and anthropogenic Fe inputs remains challenging in human-impacted river basins. Here, we combined suspended particulate Fe contents, Fe isotopes (δ56Fe), geochemical indicators, land-use information and Bayesian mixing models to constrain the sources and cycling of particulate Fe in the subtropical Jiulong River Basin, southeast China. Particulate Fe concentration (FeSPM, g/kg) ranged from 26.27 to 168.79 g/kg, and δ56Fe values varied from -0.94‰ to 0.17‰, with basin-wide means of 49.65 g/kg and 0.01‰, respectively. Marked spatial heterogeneity was observed among sub-basins. The tributaries of Beixi River had the highest particulate Fe concentration (mean: 64.48 g/kg) and the most negative δ56Fe signatures, indicating intensified inputs of isotopically light Fe. Correlation analysis and Bayesian source apportionment suggested that agricultural land use was a major basin-scale control on FeSPM, accounting for approximately 45% of FeSPM in the Beixi River under wet-season conditions, whereas mining inputs were locally important in the Beixi headwaters and may have contributed to the depleted δ56Fe signatures. Comparison with subtropical and tropical rivers worldwide further implies that agricultural disturbance can enhance riverine particulate Fe mobilization, while mining-derived Fe represents a potentially significant non-point source in mineralized catchments. These results indicate that the coupled isotope-Bayesian framework offers a promising approach for distinguishing overlapping natural weathering and anthropogenic Fe inputs, providing a quantitative basis for assessing particulate Fe transport from human-impacted rivers to coastal waters, pending flux quantification in future work.