Tracing the dominant sources of sediment flowing towards Lake Victoria using geochemical tracers and a Bayesian mixing model

Abstract Purpose Lake Victoria has been increasingly silting over the past decades, impacting water quality and loss of biodiversity. Sediment control strategies require information on relative absolute contributions sediment from different sources. However, to date, there is no continuous monitoring flux or in any tributaries, prohibiting an assessment scale problem. The aim this study was trace dominant sources riverine using geochemical fingerprinting, thereby generating a knowledge base for improving land management reducing yields Simiyu River catchment, one main contributing rivers Victoria. Materials methods Geochemical tracer concentrations were analyzed transported river two tributaries (riverbed sediments) soils five use types (agricultural land, bush forest channel banks, banks). Dominant attributed Bayesian MixSIAR model. Results discussion mixing model outputs showed that tributary source with 63.2%, while Duma accounted 36.8%. Cultivated shown be sediment, accounting 80.0% 86.4% sub-tributaries, respectively, followed by banks 9.0% both sub-tributaries. Direct unmixing yielded 64.7% contribution cultivated 26.5% banks. Conclusion demonstrated application tracing provides important pathway quantifying flowing towards Eroded soil agricultural areas biggest contributor River. This vital design catchment wide plans should focus erosion delivery farming networks, ultimately supporting food security Basin.