T23D-2980
Anisotropic Velocity Structure near the Alpine Fault, New Zealand, from Shear Wave Splitting of Local Earthquakes

Tuesday, 15 December 2015
Poster Hall (Moscone South)
Martha K Savage1, Carolin Morag Boese2, John Townend1 and the DFDP Alpine Fault drilling team, (1)Victoria University of Wellington, Wellington, New Zealand, (2)International Earth Sciences IESE Lt., Auckland, New Zealand
Abstract:
We use shear wave splitting of local earthquakes to characterize seismic anisotropy along the central portion of the transpressive, plate-bounding Alpine Fault and surrounding the Deep Fault Drilling Project (DFDP) boreholes. We combine data from the SAMBA network of 2 Hz borehole seismometers, the DFDP10 short-period surface network, and the GeoNet broadband network. Using the MFAST automatic shear wave splitting program applied to hand-picked S arrivals, 1.5 years of data yielded 15,902 measurements. Of these, 1291 measurements from 23 stations were of high quality (grades of A or B); their fast azimuths exhibit three predominant trends: one NE/SW, subparallel to the Alpine Fault, one perpendicular to it and the third E-W. The last of these, observed at the six stations closest to the fault, is roughly parallel to the maximum principal stress direction. Some stations show a single population of fast directions and others show two or more populations. We consider that the orientations are likely caused by a combination of stress-controlled, crack-induced anisotropy and structure-controlled anisotropy associated with the fault fabric and its interaction with the dominantly schistose rocks in the region. Delay times for the highest-quality measurements average 0.067±0.002 s, consistent with small splitting of the high-frequency phases (dominant frequencies 6.9±0.07 Hz) with short path lengths (4.5±0.09 s travel-time). If the anisotropy is present along the entire path, the percentage anisotropy is 1. 5%. However, delay times do not increase much with hypocentral distance, suggesting near-surface effects are dominant and hence that anisotropy is larger near the surface. To ensure the shear-wave splitting results do not depend on the particular algorithm used, a second semi-automated processing routine with different shear-wave splitting criteria is applied. A comparison of the results is anticipated to help to eliminate any possible artefacts.