Local sensitivity analysis of fitting parameters for the water retention curve in unsaturated flow models in filtered tailings

Accurate modelling of unsaturated flow in filtered tailings is critical for understanding their hydrological behaviour and improving resource management in mining operations. This study investigates the sensitivity of fitting parameters in the van Genuchten model, which characterises the water retention curve (WRC), through a comprehensive local sensitivity analysis. Laboratory challenges in determining WRCs, due to time and resource constraints, have led to reliance on estimates derived from material index properties which often fail to accurately represent moisture retention behaviour. 3,000 1D simulations were performed, using HYDRUS-1D, to assess the influence of six key parameters: residual moisture content (qr), saturated moisture content (qs), initial moisture content (qi), fitting parameters (a and n) and saturated hydraulic conductivity (Ks). The study focused on filtered tailings from a mining site in northern Chile, with simulations reflecting the region’s typical environmental conditions. The results show that n, qs and qI are the most sensitive parameters, significantly affecting water flow predictions. The presence of a foundation soil alters hydrological responses, in contrast to scenarios with free drainage. This research provides critical insights into the effects of WRC fitting parameters on unsaturated flow behaviour, enabling the definition of parameter variation ranges and improving the accuracy of flow models. The findings are invaluable for improving the reliability of tailings management and risk assessment strategies in Chile and similar environments worldwide, contributing to sustainable mining practices.