Fiber bundle model applied to slope stability assessment: co-detection multi-threshold analysis for early warning

Abstract

Forecasting the imminent failure of natural slopes is crucial for effective Disaster Risk Reduction. However, the nonlinear nature of geological material failure makes predictability challenging. Recent advancements in seismic wave monitoring and analysis offer promising solutions. In this study, we investigated the co-detection method, which involves real-time processing of micro-seismic events detected concurrently by multiple sensors, to provide easy access to their initial magnitude and approximate location. By studying the Fiber Bundle Model and considering the attenuation of seismic waves, we demonstrated disparities in the statistical behavior of various rupture types before global catastrophic failure. Comparing avalanches with attenuated seismic wave amplitudes directly measured at sensor locations, we observed differences in their evolution towards catastrophic rupture. Leveraging a network of seismic wave sensors, we showed that the co-detection method was effective in detecting precursory seismic events, even with weak signals, making it a valuable tool for monitoring and predicting unstable slopes. Additionally, we demonstrated that a multi-threshold analysis of co-detection activity allowed for instantaneous capture of the seismic activity structure on unstable slopes. These findings contribute to our understanding of slope stability and offer insights for improved hazard assessment and risk management.