Source mechanics of fluid driven volcanic earthquakes

Collaboration between UofP and the Universities of Mainz and Gottingen: Self funded student Thomas Grant ([email protected])

Seismic data is by far the most useful signal for volcanic hazards, exemplified by its use for forecasting volcanic eruptions and monitoring aftershock sequences. This project aims to generate a new geophysical (rock-physics) image as a quantified metric at elevated but modest temperatures. The aim of this project is to take existing laboratory data and to calculate likely source mechanisms for two key types of seismicity. The types of seismicity to be investigated would be; Volcano-Tectonic [VT] (brittle fracture), and Low-Frequency [LF] (primarily driven by rapid fluid flow), as a function of stress and pore fluid decompression/stress, and would be backed up with physical data using X-Ray Computed Tomography [XCT]. In addition, two models will be applied to the data: the crack resonance models of Kumagai and Chouet (2000), and that of Neuberg (Thomas & Neuberg, 2013). In both cases, it is postulated that the initial fracturing of country rock provides energy (as a VT event), that subsequently becomes trapped by cracks and/or the conduit. The resonance of the cracks/conduits generate resonance that is manifested at LF energy. The project will use the directly measured geometry of the crack and conduit network, as measured via Acoustic Emission [AE] signals, and XCT, which will be linked to the measured energy from a suite of calibrated AE sensors.

To test the Kumagai and Chouet model, the modulus, alongside known parameters of fluid and rock densities, will be used to calculate a complex frequency via the crack aspect ratios that generate that event. This data shall be used to calculate a far field seismic signature, which, in turn, will then be compared to the calibrated AE. Doing this will; test the limit(s) of the model, test sensitivity to different parameters, and forward model the likely cracks based on AE alone. The second model (Neuberg), focuses on the trigger mechanisms itself, and specifically considers a constriction, or ‘bottleneck’, on the conduit as a trigger site. This has also been verified in laboratory data in previous work (Benson et al., 2008). By combining new calibrated AE signals with these model outputs, the aim of the project is to better understand the role of fluids in generating volcano seismicity across a range of simulated pressure (depth) conditions.