Geophysical Imaging and Monitoring of Tailings Storage Facilities: A Case Study at the Trident Copper Mine, Zambia, within the MOSMIN Project

Industrial mining operations create substantial amounts of residuals known as tailings or waste rocks. Their deposition creates amongst the largest human-made structures in the world both in spatial extent and mass. Their physical stability is not self-evident and tailings dam failures have been regularly documented throughout the world. In order to mitigate environmental and societal risks, stringent regulatory frameworks have been established, e.g. through the Global Industry Standard on Tailings Management (GISTM) in 2020 introduced by The United Nations Environment Programme (UNEP) , which mandates the implementation of monitoring concepts that manage risks throughout the lifecycle of a tailings facility. The EU-funded initiative MOSMIN (Multiscale observation services for mining-related deposits) addresses this key goal by establishing comprehensive monitoring solutions of mining-related deposits. The project combines (remote) Earth observation technologies with ground-based geophysical measurements to produce vertically integrated datasets which are subsequently analysed through advanced computational techniques, including machine learning algorithms, to characterise the spatio-temporal dynamics governing deposition and containment of tailings.

 

This contribution focuses on fiber-optics-based passive seismic techniques used in MOSMIN to analyse ambient ground vibrations and generate spatially resolved shear-wave velocity models of tailings dams. The ambient noise tomography (ANT) aims at characterizing and monitoring their internal material properties across time and at different scales, resolutions and depths of investigation. We present results from a field campaign at the First Quantum Sentinel copper mine in Kalumbila, Zambia, during which passive seismic data was semi-continuously recorded with a network of passive sensors and a fibre-optic cable for almost a year. The mine’s tailings dam was equipped with 30 autonomous seismic sensors and a 7 km-long, trenched fiber-optic cable installed parallel to the dam structure and interrogated by a commercial Distributed Acoustic Sensing (DAS) system, recording continuous strain-rate data along the cable. The collected seismic data analysed here comprises 9 months of strain-rate recordings across 1.5 km of optic fiber.

 

To investigate the temporal stability of the ambient seismic wavefield as required for subsurface monitoring purposes, we characterize how it is influenced by ongoing mining activity over weeks to months. We do so by calculating the strain-rate root-mean-square (RMS) in different frequency bands across the entire recording period of the DAS campaign. Further, we discuss the feasibility of retrieving high-resolution velocity profiles of the dam across space and time using Multi-channel Analysis of Surface Waves (MASW). One important aspect is the selection of recording periods with favourable noise conditions. We investigate different strategies for the selective stacking of several thousands of virtual-shot gathers obtained from cross-correlations of 30 s-long time windows. Results show that the MASW workflow is strongly influenced by the influx of anthropogenic noise created by the mining activity during the day time.

 

Our case study works towards establishing and implementing fibre-optics-based passive seismic surveys, adapted to site-specific requirements, as a scalable, non-invasive framework for geotechnical monitoring of active TSFs. Joint analysis of passive seismic models with satellite-derived surface deformation or spectral information offers potential for improved understanding of surface–subsurface interactions in and along tailings dams.