Rupture Propagation Imaging at Microseismic Scale
a Master Thesis written by Jonas T. Folesky
A number of recent publications in global seismology deal with the tracking of the rupture front of large (M>7) to megathrust (M>8) earthquakes. Applications were e.g. the Sumatra-Andaman earthquake 2004 or the Tohoku, Japan earthquake 2011. One technique is to back project the seismograms, recorded at an array or at a seismic network to a grid of possible source locations. This method is called Back Projection Imaging and it can provide information on rupture properties like direction, speed or duration. A different approach is to use P-wave polarity estimates for moving time windows i.e. P-wave Polarization Analysis to map the zone of maximum energy release. This thesis deals with the question, if these two imaging attempts can be applied in a microseismic environment on microseismic events.
The motivation for this work is the occurrence of relatively large, induced seismic events at a number of stimulated geothermal reservoirs or waste disposal sites, having a magnitude M≥3 and yielding rupture lengths of several hundred meters. The configuration of the seismic network and reservoir properties of the Basel Geothermal Site is being used to create a synthetic model space with realistic characteristics. Using finite-difference modeling, a rupture is simulated by implementing a series of spatio-temporally separated single sources. As a next step, the Back Projection Imaging and P-wave Polarization Analysis techniques are adapted to the microseismic scale and realised in Matlab®. The programs are used to test the feasibility of recovering rupture properties like orientation, length and speed at three different synthetic reservoirs. Therefore the processing is explained in detail.
The analysis of two reservoirs with full azimuthal station coverage and unilateral ruptures yield precise results for orientation, direction and length. The application of the same imaging attempts at the synthetic Basel reservoir, using significantly less stations and thus poorer azimuthal coverage, yields respective results. However, both rupture imaging approaches are found to confirm their applicability at reservoir scale. Finally, the general results are compared and discussed, a few shortcomings are evaluated and future enhancements are proposed.
© Jonas T. Folesky