Using Phase Amplitude Ratios to Differentiate Blast, Collapse, and Fault Slip Seismic Events from Deep Gold Mines in South Africa

Effective differentiation of subsurface explosions from slip on faults is fundamental to the nuclear explosion monitoring community. Typically, methods that are effective at distinguishing explosion sources (e.g., Ms:mb) have been developed and tested for relatively large (M > 4.75) events observed at regional and teleseismic distances. Using the site and path corrected amplitude ratios of P- and S-waves at > 2-4 Hz has improved source type identification for smaller (M > 3), regionally recorded events. Currently, a large effort is underway to extrapolate the theory and application of these methods to smaller events (M < 3) recorded at local to near-regional distances. To address this challenge, we utilize data from sources related to deep (0-4 km) gold mining operations, along with associated earthquakes, recorded on in-mine and surface networks to test previously developed methods (i.e., phase and spectral ratios). Data consist of hundreds of (0.5 < M < 3.5) mine blasts, pillar collapses, and earthquake aftershocks of the August 5, 2014 M5.5 Orkney earthquake that occurred within and directly beneath the Kloof Gold Mine in Klerksdorp, South Africa. We first develop and apply local site and path corrections using the MDAC approach to waveforms recorded on surface seismometers. From these corrected waveforms, we measure phase, spectral, and cross-spectral amplitude ratios (e.g., P/S for varying frequency bands) to determine the effectiveness of these discriminants for small yield events.