Many fundamental questions in earthquake physics remain unresolved, largely because of difficulties in obtaining high-quality, near-source recordings of small earthquakes and deriving reliable source parameters. The overarching goal of this project is to provide an improved understanding of earthquake source processes for a wide magnitude range (approximately -4 to 4). We are performing detailed seismological investigations of source properties that characterize several distinct classes of seismic events (earthquakes, rockbursts, and blasts) in a 3.6 km deep South African mine. We will also conduct a short field campaign to record seismic events at the extremes of the magnitude range and in close proximity to fault structures. Full source tensor inversions, second moment analysis, and spectral methods will be used to determine comprehensive sets of source properties including scalar potencies/moments, focal mechanisms, corner frequencies, strain/stress drops, radiated energies, rupture velocities, and source time functions. Special effort will be devoted to detecting isotropic components of faulting. Theoretical calculations of expected radiation from different event types will help in the interpretation of results. Applying the above analysis to the small-magnitude, near source seismic data will allow us to obtain constraints on many aspects of earthquake physics including nucleation processes, the prevalence of isotropic components of seismic radiation, changes in the physics of rupture with event magnitude or with proximity to geologic and mining structures, and source parameter differences between earthquakes, explosions, and other seismic sources. The high-resolution results on seismic source properties and scaling relations will significantly impact the earthquake physics and rock mechanics communities by helping to resolve long-term controversies on earthquake nucleation and rupture processes, scaling of laboratory results to natural faults, and earthquake energy budgets, among others. In addition, the findings on spatially dependent seismic properties will be useful for decision-making concerning when and where to suspend mining activities to enhance underground safety
This project provided support for a comprehensive analysis of seismic processes in a deep South African gold mine. Using high-quality, near-source recordings of small earthquakes that span six orders of magnitude we have derived reliable source parameters and obtained constraints on many aspects of earthquake physics. The topics we have addressed include assessing the prevalence of isotropic components of seismic radiation and its implications for earthquake nucleation and damage creation. We developed an improved moment tensor inversion technique to investigate the full source tensor of events using nearfield phase amplitudes. These codes will soon be freely accessible and useful to other seismologists as they can be easily applied to many datasets or used in teaching. Our results show both implosive and deviatoric sources throughout the magnitude range. Specifically, from 100 mining-induced earthquakes between -2.7 ≤ MW ≤ 2.5, we find that 35% involve isotropic source components with primarily implosive components. These results extend the range of observed implosive events by two orders of magnitude. Taking the results of our moment tensor analyses together with those of previous studies, we find that mining-induced events of all sizes begin as shear failure that may intersect a void (stope or tunnel) and cause collapse, while only small events (MW < 0) result in explosive sources. Our results suggest maximum nucleation zone size on decimeter-scale. We find no evidence that damage creation in the source region results in significant tensile components of seismic radiation, either due to brittle deformation or to rupture on a bimaterial interface. We have investigated earthquake triggering and stress transfer processes and found that earthquake nucleation caused by mining-induced seismicity follows the same process as that triggered by stress transfer from larger natural earthquakes. Similarly, we found values of apparent stress from events in our data set are consistent with those of natural tectonic earthquakes. This project provided support, seismological training, and a unique dataset for one PhD student, one postdoctoral researcher, and two undergraduates as well as material for lectures in introductory undergraduate classes. The results are published in three peer-reviewed papers with two more articles in preparation.
