Application of set pair analysis and uncertainty analysis in groundwater pollution assessment and prediction: a case study of a typical molybdenum mining area in central Jilin province, China

Abstract

This study evaluated groundwater pollution from a molybdenum mining site in Jilin province, China. First, groundwater pollution in the study area was evaluated using set pair analysis (SPA), which addresses the uncertainties of the evaluation factors and the complex nonlinear relationship among them. Second, groundwater pollution after 3 years was predicted. In this process, the influence of parameter uncertainty on simulation results was analyzed using Monte Carlo simulation, and the results of the analysis were explained from the perspective of pollution risk assessment. Monte Carlo simulation requires repeated invocation of a simulation model, which creates a large computational load. To solve this problem while ensuring high simulation accuracy, sensitivity analysis was used to select the more sensitive model parameters as random variables, and establish a surrogate model based on the Kriging method, thus facilitating the Monte Carlo simulation. We found that: (1) according to the SPA evaluation results, areas of extremely serious pollution were distributed downstream of the tailings and spoil bank. The groundwater in the study area was mainly affected by the tailings reservoir leakage and the spoil bank leachate. (2) Based on the uncertainty analysis, the serious pollution risks for observation wells 1, 2, 3, 4, 5, 6 and 7 were 100%, 100%, 64%, 79%, 52%, 100% and 23%, respectively. Groundwater in the southeast and northwest of the study area had a higher risk of pollution. In addition, pollution risk in the area near the spoil bank and tailings was higher than that of other areas. (3) The surrogate model established using the Kriging method not only had high accuracy and could fully approximate the input–output relationships of the simulation model, but also significantly reduced computing load and computation time. The results can provide a scientific basis for the prevention and control of groundwater pollution.