Estimation of anthropogenic pollution using a Bayesian contamination model: An application to fractured bedrock groundwater from Han River Watershed, South Korea

It is well known that groundwater is a valuable but vulnerable natural resource. To set forth a proper strategy for conservation and sustainable use of groundwater resources, we need precise evaluations of the human impact on groundwater quality. In this paper, we develop a Bayesian contamination model that clusters the sampling locations of groundwater into polluted and unpolluted groups and simultaneously estimates the average amount of human impact. Among major dissolved ions in groundwater, NO₃^- , Ca ²⁺, SO₄^2-, Cl^-, and Na⁺ were documented as useful variables describing the hydrochemical characteristics of anthropogenically polluted groundwater. Increased concentrations of these ions indicate that overused agrochemicals (particularly nitrogen fertilizers) and domestic sewage are the most important causes of groundwater pollution to considerable depths in the studied region. Our proposed model can be used to identify effective measures for groundwater quality management such as source control.