Tracing the source and behaviour of sulphate in karst reservoirs, using stable isotopes and Bayesian isotopic-mixing models.

Increases in sulphate concentrations in natural water bodies can lead to the deterioration of water quality. Human activities, such as coal mining and agricultural fertilisation, can generate sulphate, which can enter water bodies through surface runoff or underground pipelines. Owing to the widespread distribution of coal-bearing strata and an intensification of industrial and agricultural activities, the Pingzhai Reservoir is increasingly at risk of sulphate pollution. In this study, 42 water samples were collected from the Pingzhai Reservoir in April (normal season), July (wet season), and December (dry season) of 2022. Additionally, two precipitation samples, two sewage samples, and two acidic mine drainage samples were collected. Using hydrochemistry, multiple isotopes (δS, δO, δC, and δO) and Bayesian isotopic-mixing model methods, we qualitatively and quantitatively determined the source, contribution proportion, and behaviour of SO in the Pingzhai Reservoir watershed and evaluated the uncertainty of the estimated results. Isotope analysis and the Bayesian isotope-mixing model results indicated that the sources of SO in the Pingzhai Reservoir were coal sulphides and organic sulphur oxidation (64.2 %), soil organic sulphur (18.7 %), sewage (9.9 %), and agricultural sulphur fertiliser (7.2 %). The characteristics of karst landforms (thin soil that is easily eroded), combined with periodic fluctuations in water level (hydrofluctuation belts) in reservoirs, resulted in the release of organic matter from soil to water. The proportion of SO sources of coal sulphide and organic sulphur oxidation in the river was lower than that in the reservoir area, whereas the proportion of the SO sources of soil organic sulphur, sewage, and agricultural sulphur fertiliser was greater than that in the reservoir area. Isotope evidence and the aerobic conditions in water indicated that bacterial sulphate reduction processes did not play a major role. The uncertainty index (UI90) indicated that the contributions of agricultural sulphur fertiliser and sewage manure to SO were relatively constant. This study provides a reference for the protection of water environments and for the development of water pollution control strategies in the karst areas of southwestern China.