Do stopes contribute to the seismic source?

Parameters such as source location, seismic moment, energy, source size, and stress drop are routinely calculated from mining-induced seismic data. Seismic moment tensors are inverted less routinely because their calculation is more complex and their accuracy depends on the network geometry, among a number of other factors. The models utilized in the source parameter calculations, the most wellknown of which is the Brune model, were developed for the global seismicity problem and assume a solid, homogeneous Earth model. However, the tabular orebodies in South African gold and platinum mines are mined extensively and the excavations (stopes) can extend for many kilometres. The seismic source mechanisms on deep-level gold mines are generally compatible with shear failure (Hoffmann et al., 2013), whereas the source mechanisms of events at intermediate-level bord and pillar mines in the platinum district are more compatible with pillar failure and accompanying stope closure (Spottiswoode, Scheepers, and Ledwaba, 2006; Malovichko, van Aswegen, and Clark, 2012). In this paper we investigate the influence of the stope on seismic inversions for the scalar moment, corner frequency/source radius, stress drop through numerical modelling using WAVE3D. The main objective is to determine whether the source parameters calculated from the recorded waveforms are due to a combination of the stope and shearing sources, rather than being related only to a shear source in the host rock. The modelled source is shear rupture in the footwall of a stope. The results show that the stope appears to have an appreciable effect on the seismic inversions. The seismic moment and source radius of the shear source in the stope are larger for the model with a stope compared to the model with no stope. The stress drop for the case with a stope is less than the applied stress drop, which could be an effect of the apparently larger source. This work provides a possible explanation of the second corner frequency often observed in the spectra of seismograms recorded in South Africa platinum mines. This has implications for the accurate determination of source parameters and the assessment of the intensity of shaking in stopes.