Insights into induced earthquakes and aftershock activity with in-situ measurements of seismic velocity variations in an active underground mine

Abstract:

The behaviour of the crust shortly after large earthquakes has been the subject of numerous studies, but many co- and post-seismic processes remain poorly understood. Damage and healing of the bulk rock mass, post-seismic deformation and the mechanisms of earthquake triggering are still not well understood. These processes are important to properly model and understand the behaviour of faults and earthquake cycles.

In this presentation, we will show how in-situ measurements of seismic velocity variations have given new insights into these co- and post-seismic processes. An experiment was performed where a blast was detonated in a tunnel in an underground mine, while seismic velocity variations were accurately (0.005 %) measured with ambient seismic noise correlations.

Additionally, aftershock activity was examined and the influence of the removal of a piece of solid rock was estimated with elastic static stress modelling. The majority of the aftershocks were delayed with respect to the passing of the dynamic waves from the blast, while the locations of the aftershocks appeared clustered and not homogeneously spread around the blast location.

A significant velocity drop is visible during the time of the blast, which is interpreted as co-seismic damage and plastic deformation. These non-elastic effects are healed by the confining stresses over a period of 5 days until the seismic velocity converges to a new baseline level. The instantaneous weakening and gradual healing observed from the velocity variations are qualitatively similar to results reported in laboratory studies.

The change in the baseline level of the seismic velocity before and after the blast indicate a change in the static stress that is comparable to the results of elastic static stress modelling. The differences between the elastic model predictions and the seismic velocity variations could be due to zones of fractured rock, indicated by the spatial clustering of the aftershocks, that are not represented by the homogeneous isotropic elastic model.

The delayed timing of the aftershocks, and the duration to velocity evolution, suggest that the same processes (stress change and damage) could control the timing of the two phenomena.