Integrating isotope and fluorescence approaches for joint source tracking of nitrate and CDOM in an agricultural watershed

Effective management of surface water quality requires a thorough knowledge of the characteristics and contributions of various pollution sources. While stable isotope methods are highly effective for nitrate source tracking, their high cost and operational complexity constrain their routine use in watershed management. This study examined and contrasted the dual stable isotopes (δ¹⁵N-NO₃ ^-, δ¹⁸O-NO₃ ^-, and the MixSIAR [Bayesian mixing models in R] model) with fluorescence analysis (EEM-PARAFAC [fluorescence excitation-emission matrices coupled with parallel factor analysis]) to trace the origins of nitrate and dissolved organic matter in the Lujiang River, a coastal agriculture-dominated river in southeastern China. Additionally, the feasibility of EEM-PARAFAC as a low-cost complementary tool for nitrate source tracking was assessed. The MixSIAR model identified soil nitrogen (41.0%-49.7%) and fertilizers (29.5%-37.9%) as dominant nitrate sources, pointing to significant nonpoint source pollution. With increasing urbanization, point source pollution from manure and sewage increased from upstream (11.5%) to midstream (13.1%) and downstream (15.3%). EEM-PARAFAC analysis supported these findings, with humic-like components (C1 + C2: 59.31%, 56.85%, and 46.98% in upstream, midstream, and downstream, respectively) showing a strong correlation (r = 0.97) with soil nitrogen contributions identified by MixSIAR. Protein-like components (C5, r = 0.93; C6, r = 0.97) were associated with fertilizers and sewage, respectively, consistent with their increasing downstream contributions, validating EEM-PARAFAC's capacity for cost-effective source discrimination. This study demonstrates that integrating the two methods enhances the understanding of pollutant source dynamics in complex river systems, offering valuable insights for improving water quality and watershed management.