Although sulfur-bearing minerals are valuable resources, they pose significant environmental risks to river ecosystems by releasing hazardous leachate. Accurately tracing these sources is crucial but challenging due to overlapping chemical signatures and pollutant transport dynamics in river systems. This study investigates seasonal and spatial variations in sulfate (SO42 -) and trace element contributions in mining districts of the upper Nakdong River basin, South Korea. Sulfate isotopes (δ34SSO4 and δ18OSO4) and Monte Carlo simulations enhanced the precision of source estimation. Bayesian stable isotope mixing models revealed acid mine drainage (55.3 %) dominated during the wet season, while smelter tailings and mine leachate dominated during the dry season (43.8 %). Model uncertainty was addressed through probabilistic approaches, with Monte Carlo simulations, ensuring robustness in δ18OSO4 predictions. Additionally, the Positive Matrix Factorization (PMF) model identified dominant emission sources and impact ranges for hazardous elements, mostly validated by high correlation coefficients (r > 0.8). Notably, the contribution of Cd-dominated emissions from smelters increased from 13 % in the wet season to 51.5 % in the dry season, while the contribution of As-dominated emissions from mining increased from 30.3 % to 40.7 %. Despite sorption and coprecipitation processes, As contamination migrated ∼ 50 km downstream during the dry season. Statistical correlation and cluster analyses further confirmed consistency between observed and modeled geochemical patterns. This study underscores the reliability of integrating Bayesian and PMF models with isotopic and hydrogeochemical data, offering valuable insights for tracing anthropogenic sources in complex mining environments.
Read full abstract