Monitoring data analysis and groundwater flow modelling at a former uranium mine in France

Mining companies are responsible for after-mining and environmental monitoring to ensure that mining waters released in the environment meet environmental quality standards. Water treatment plants can be used to mitigate surface waters when impacts related to past mining activities is evidenced. Indeed, meteoric recharge through waste rocks and tailings stored on sites often creates Acid Mining Drainage (AMD) or metallic signature that can be transferred to rivers through groundwater flow.

We studied a former uranium mine in Bertholène, Aveyron (France), where tailings, covered by waste rock, are stored in a valley behind a waste rock dyke. Mining waters, coming from both tailing drainage and mine facilities (galleries and open pit mine), are collected to a water treatment plant before release in the environment, meeting the environmental standards. Groundwater flow modelling is required to understand and quantify the different sources of AMD and their fate. The objective of this study is to give new insights on flows to guide the potential additional remediation of the site by testing different management solutions. For this purpose, we have developed a 3D hydrogeological model (MODFLOW) for the entire watershed.

This work is divided into two parts. We first analysed all available climatic and hydrogeological data (precipitation, water level, surface water flow, electric conductivity) using auto-correlations, cross-correlations and water balance calculations. These data come from long-term monitoring (14 years) on 13 piezometers and 4 discharge points. Recent two-year daily monitoring of groundwater levels completes the data set and provides a better understanding of the dynamic of the hydrosystem after precipitation. There is a 5 to 10-day time lag between rainfall and increases in water level and flow rate. The analysis also concludes that flows in the tailings occur under unsaturated conditions and that the water level in the gneiss aquifer never reaches the tailings.

Steady-state modelling developed at the watershed scale confirms that water level does not reach the tailings and allows simulating the impact of different management scenarios. Particle tracking has also been used to identify hydrogeological sub-watersheds of interest, such as those of the former open pit mine or the tailings to compute their water balance.