Forecasting the Time of Failure of Tailings Dams and Heap Leach Facilities through Satellite InSAR Data: myth or reality?

This paper addresses the complex challenge of forecasting the Time of Failure (ToF) for Tailings Storage Facilities (TSFs) and Heap Leach Facilities (HLFs) using satellite-based Interferometric Synthetic Aperture Radar (InSAR) data. TSF and HLF structures can be susceptible to different failure modes, some of which can result in sudden instability due to the tailings' contractive and brittle behaviour. Moreover, the infrequent satellite passes can limit the ability to detect rapid changes or deformations in these structures, potentially reducing the effectiveness of monitoring efforts. Some recent papers propose predicting brittle failures in such structures using tertiary creep-based forecasting methods. However, this often requires a subjective procedure, and the results yielded are sometimes different and sometimes contradictory. To fully exploit the predictive potential of satellite InSAR data, a statistical multi-model methodology is introduced, with a primary focus on providing both quantitative estimates of the predictions and forecasting reliability. Through this methodology, different levels of warnings are issued. The first level, low-confidence alert, would have been reached for Cadia up to 2 months in advance, whereas the highly confident level would have been reached 12 days prior to the collapse. Moderate confidence alert 39 days in advance Feijão failure. Special emphasis is placed on the evolution of failure risk in a simulated prospective manner, reproducing operational monitoring conditions. Our findings highlight the need for multiple technologies to accurately monitor such structures and justify the use of satellite InSAR for hazard screening. • A statistical multi-method ToF forecasting approach is introduced. • The methodology detects alerts within a simulated prospective approach. • RA–InSAR framework can provide confidence-based early-warning alerts. • Satellite monitoring proved effective for selected cases, such as Cadia. • Complementary monitoring systems are needed for rapidly evolving failures.