Statistical properties of seismic foreshocks and aftershocks associated with longwall mining: Implications from the Epidemic Type Aftershock Sequence model

Abstract

The control and management of mining-induced seismic hazards have attracted ever-rising attention, especially in underground longwall coal mines, where continuous mining activities dynamically alter the stress states and induce seismic events. In this work, the Epidemic Type Aftershock Sequence (ETAS) model was applied to formulate the aftershock catalogue of mining-induced seismicity and investigate the formation of event triggering associated with longwall mining. The conventional Baiesi and Paczuski method (2004) was used to separate longwall mining-induced seismic events into triggered and non-triggered catalogues. The latter catalogue contains both non-triggering (NT)-isolated events that do not trigger subsequent events and NT-parent events of the former catalogue. Statistical properties of triggered events were analysed spatially and temporally. The temporal triggering sequence follows the Omori-Utsu law, where the temporal decay of aftershocks is influenced by the magnitude of NT-parent events in mining-induced seismicity. The spatial distribution of aftershocks follows an inverted U-shaped relationship with distance to their corresponding NT-parent events. The quantitative forecasting of triggered events was performed based on the nonhomogeneous Poisson distribution, which achieved a good consistency with their NT-parent events. Amongst the non-triggered catalogue, NT-isolated events are concentrated ahead of NT-parent events, potentially acting as foreshocks for the latter.