Fiber-optic monitoring of tailings dams in the MOSMIN project

Tailings and waste rocks are residuals of industrial mining operations and their collection amounts to the world's largest human-made structures both in spatial extent and mass. The catastrophic consequences of tailings dam failures have been extensively documented, prompting the establishment of stringent regulatory frameworks to mitigate environmental and societal risks. The United Nations Environment Programme (UNEP) introduced the Global Industry Standard on Tailings Management (GISTM) in 2020, which mandates the implementation of monitoring concepts that manage risks throughout the lifecycle of a tailings facility. Addressing this requirement is a key goal of MOSMIN (Multiscale observation services for mining-related deposits), an EU-funded initiative aimed at establishing comprehensive monitoring solutions for mining-related deposits. The project integrates Earth observation technologies with ground-based geophysical measurements to create unified datasets. These datasets are then analysed using advanced computational techniques, including machine learning algorithms, to characterise spatio-temporal dynamics relevant to the safety and sustainability of mining-related deposits.

We contribute to these efforts with in situ high-resolution passive seismic measurements conducted at the tailings storage facilities of two copper mines: the FQM Sentinel mine in Kalumbila, Zambia, and the CODELCO Chuquicamata mine near Calama, Chile. Both experiments aim to seismically characterize the internal structure of the dams and to monitor subsurface processes at different scales, resolutions and depths of investigation through the creation of shear wave velocity models using ambient noise tomography (ANT). Similar array designs were implemented for both sites. Each site was equipped with a kilometers-long, trenched fiber-optic cable interrogated by a commercial Distributed Acoustic Sensing (DAS) system along with 30 conventional geophones. Both types of instrumentation were installed parallel to targeted tailings dam sectors and recorded during regular mining, disposal, and maintenance activities around the tailings facility for several months. However, the highly variable seismic wavefield generated by the active mining environment poses challenges for the ANT procedure. In order to obtain an overview of the wavefield’s spatio-temporal behaviour, we calculate the strain-rate root-mean-square (RMS) in different frequency bands across the entire recording period and fiber array, which encompasses approximately two months across 4 km and 9 months across 1.5 km of optic fiber for the Chile and Zambia sites, respectively. We present results regarding the spatio-temporal variation and stability of cross-correlations, and discuss the feasibility of performing Multi-channel Analysis of Surface Waves (MASW) to obtain high-resolution profiles of the velocity structure of the dam across space and time.

Ultimately, tracking seismic changes in the dam structure could be used as an additional tool for non-invasive, spatially and temporally continuous geotechnical monitoring of the tailings storage facility and, in joint analysis with InSAR-derived surface deformation, reduce false alarms and enable physically meaningful, surface-subsurface interpretation.