Abstract
BACKGROUND: The Hanzhong catchment is the primary agricultural area in the upper reaches of the Danjiangkou Reservoir. The challenge of quantifying natural as well as anthropogenic point and nonpoint inputs of riverine phosphorus (P) in this catchment severely hinders P management for the reservoir's upstream regions. METHODS: For the first time, a Bayesian stable isotope mixing model based on the phosphate oxygen isotope (δ(18)O-PO(4)) data from three hydrological periods was employed to apportion natural and anthropogenic sources and discern the influence of environmental factors. RESULTS: Riverine phosphate mainly derives from natural sources during the normal flow period (approximately 44%). Enhanced leaching due to monsoon precipitation promotes natural inputs in most catchments. During the low flow period, the extensive application of basal fertilizer resulted in a marked rise in agricultural sources (58%). Correlation analysis indicates that wastewater input is affected by built-up land area. Conserving water and improving water efficiency are crucial for minimizing wastewater input in the Hanzhong catchment. The fertilizers applied to farmland are dominated by chemical fertilizers. Reducing chemical fertilizer usage contributes to decreasing agricultural inputs. Increased rainfall and temperature result in greater input from natural sources. This work highlights that controlling agricultural sources, especially during the low flow period, is the primary task for P management in the Hanzhong catchment and further reveals that the primary P pollution source (livestock and poultry breeding) in the Hanzhong catchment, identified by prior research, may enter water bodies from cropland in the form of basal fertilizer.