The Source Phenomenology Experiment (SPE) was a series of nine, single-fired chemical explosions within the Morenci Copper mine in Arizona. Its purpose was to design, detonate, record and analyze seismic waveforms from these single-fired, partially and fully contained explosions. Ground motion data from the SPE are analyzed in this study to assess the uniqueness of the source representation of these explosions and its ability to resolve yield and depth when containment and geology or physical parameters of the source region may have a range of possible values. The P-wave velocities (Vp) at the test site are well constrained by seismic refraction surveys, but the accompanying shear wave velocities (Vs) are less constrained. In order to assess the effects of source depth and Vs model on the seismic moment tensors, Green’s functions were computed for different source depths as well as different Vs models, holding the Vp model constant. The Green’s functions for the 16, near-source stations were calculated using a one-dimensional velocity model developed from the SPE employing reflectivity modeling in order to include spherical wave effects, body waves and surface waves, focusing on observations in the 37-680 m range. The compensated linear vector dipole and explosion components of the Green’s functions are compared to quantify the possible effects of source depth and Vs on the source representation on expected explosion contributions. For the forward model, Green’s functions with variable depths of burial (DOB) and Vs are convolved with a time function based on the Mueller-Murphy (1971) isotropic source function produce synthetic seismograms for assessing possible tradeoffs between depth and yield in the source models. Our study suggests that the original SPE model parameter values used are most representative of the geology. Subsequently, observational data inversions are conducted within the frequency domain and moment tensors are decomposed into deviatoric and isotropic components to evaluate the effects of containment and yield on the resulting source representation. Isotropic moments are compared to those for other contained explosions as reported by Denny and Johnson (1991) and are in good agreement with their scaling results. Isotropic and Mzz moment tensor spectra are compared to Mueller-Murphy (1971), Denny-Johnson (1991) and revised Heard-Ackerman (Patton, 2012 b) models and suggest that the larger yield explosions with the most confinement fit the models best. Secondary source effects resulting from free surface interactions, including the effects of spallation, contribute to the resulting moment tensors, which include a CLVD component. Hudson diagrams, using frequency domain moment tensors, are computed as a tool to assess how these containment scenarios affect the source representation. Our analysis suggests that, within our band of interest (2-20 Hz), as the frequency increases, the source representation becomes more explosion like, peaking at around 20 Hz. These results guide additional analysis of the observational data and the practical resolution of physical phenomenology accompanying underground explosions.
Number of Pages
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial 4.0 License
MacPhail, Mason, "Detection, Containment and Scaling Relations of Near Source Explosionsin Granite Through Moment Tensor Representations" (2018). Earth Sciences Theses and Dissertations. 7.