Determination of full moment tensors of microseismic events in a very heterogeneous mining environment

The determination of source parameters and full moment tensors in mines is a difficult task, especially, when the velocity model of the mining environment is complex and strongly heterogeneous. The heterogeneities in the velocity model are usually caused by the presence of ore bodies, host rocks, tunnel systems and large excavations due to mining activity. The mined-out cavities introduce strong velocity contrasts in the model, cause multiple scattering of waves and result in a complex wave field with long coda waves. We have analysed five blasts and five induced microseismic events recorded at the Pyhäsalmi ore mine, Finland, and suggest a strategy of successfully inverting for the seismic moment tensors. We compute accurate locations using an eikonal solver and perform the time-domain moment tensor inversion from full waveforms using a generalized linear inversion. Green's functions are computed using a 3-D finite difference visco-elastic code capable of reproducing complex interactions of waves and structures. To suppress the sensitivity of the inversion to inaccuracies of locations and the velocity model, we analyse the data in the frequency range from 30 to 80 Hz. The analysis of blasts and microseismic events proves that the moment tensor inversion is successful. The moment tensors of blasts display a high percentage of positive isotropic components. However, the presence of minor shear faulting triggered during blasting cannot be excluded. On the other hand, the moment tensors of microseismic events display significant negative isotropic and compensated linear vector dipole components. This indicates that the predominant mechanism of the events is probably related to the collapse of rock due to mining activity.

► Successful inversion for moment tensors in very complex mining environment
► Modelled polarity switch on focal sphere by diffraction of energy around cavities
► Computed Green's functions using 3-D velocity model to account for complex structure
► Time-domain full-waveform inversion to account for complex wave field (coda waves)
► Event mechanisms display significant negative ISO and CLVD components.