Seismic anisotropy and velocity structure in North Island, New Zealand
This thesis investigates crustal and upper mantle seismic anisotropy, via shear wave splitting (SWS) analysis, across the Hikurangi subduction zone of the North Island, New Zealand. Seismic anisotropy is defined as a directionally dependent elastic response of seismic waves to an anisotropic material. Seismic anisotropy in the Earth can arise from a number of causes, including stress-induced alignment of cracks, alignment of anisotropic crystals, structures (e.g., fault fabric), and lithology. SWS is a powerful technique used to examine anisotropy beneath the surface of the Earth and is one of the few methods available to measure stress in the crust and lithosphere as well as strain and flow in the mantle. The North Island of New Zealand provides an excellent laboratory to study plate boundary deformation processes. The abundance of seismic data for North Island makes it possible to perform SWS analyses across the Hikurangi subduction zone. Several SWS studies have been conducted in North Island, but few have covered the region in high detail, leaving large swathes of North Island unstudied. Here, high-resolution SWS analysis is performed in three distinct areas covered by seismic array experiments. Of these three areas, one is offshore and two are on-land, and each presents a distinct geodynamic context and associated challenges. The first study investigated an area offshore the east coast of North Island, which is a region of frequent slow slip. This study builds on the success of the Hikurangi Ocean Bottom Investigation of Tremor and Slow Slip (HOBITSS) experiment, whose array of ocean bottom instruments captured, for the first time, a large shallow slow slip event (SSE). Understanding the physical processes occurring during SSEs is particularly important from an earthquake hazard perspective, as SSEs may influence the timing of nearby large earthquakes, or alternatively be triggered by earthquakes. The data collected by the HOBITSS instruments are analyzed to extend crustal splitting results offshore and to examine temporal and spatial variations in crustal anisotropy during an SSE. The complementary sensitivity of Vp/Vs and SWS measurements to cracks in the upper crust allowed us to evaluate the anisotropy before, during and after the September-October 2014 SSE. Temporal variations in Vp/Vs and delay time are observed during the SSE and are consistent with fluid pressurization below a permeability barrier and movement of fluids during the build-up to and rupture of a slow-slip patch. This study demonstrates that SWS and Vp/Vs are effective tools for investigating stress changes and fluid migration during SSEs. The second study focused on the central North Island, which is covered by an array of permanent GeoNet stations. I produced one of the largest datasets of crustal anisotropy measurements for this region. Using this dataset, we investigated the relationship between seismic anisotropy and the stress state by examining 42,423 high-quality SWS measurements across 24 GeoNet land-based seismic stations. The initial aim was to search for temporal changes in the state of the crust during one of the longest and deepest SSEs recorded in New Zealand; however, we did not find any significant temporal variations in our SWS results. We compared our SWS fast polarization azimuths to stress orientations derived from continuous campaign GPS (Global Positioning System) and gravitational stress calculations, as well as orientations of active faults. The spatial averaging of SWS fast polarizations azimuths showed significant spatial variations across central North Island. Comparisons with other measurements helped to reveal dominant influences. The fast azimuths at many stations across the North Island Dextral Fault Belt (NIDFB) were consistent with the regional NE-SW fault orientations, suggesting a strong structural control on anisotropy. Two regions showed clear deviations from the structural trend and were more similar to the regional maximum compressional stress SHmax. These included fast azimuths within the Wanganui Basin, as well as in a small area along the NIDFB. Additionally, fast azimuths around Mt. Ruapehu showed complex variations and resemblance to both SHmax and fault orientations, suggesting a combination of both stress-induced and structural control. With knowledge of the local stress field, as well as structural elements in this region we were able to determine whether crustal anisotropy is caused by stress, structure, or a combination of both mechanisms. The third study investigated the relationship between seismic anisotropy and mantle deformation across a deep boundary in the mantle, marked by the Taranaki-Ruapehu Line (TRL), in the back-arc region of western North Island. The TRL marks the boundary between a deeper crust (∼32 km thick) to the south and a shallower crust (∼25 km thick) in the north, interpreted as an abrupt step in the Moho. Numerous geophysical studies have examined changes across the TRL, such as in electrical conductivity, gravity, seismic attenuation, and crustal thickness, but few studies have provided constraints on the upper mantle structure from seismic anisotropy. We investigated upper mantle anisotropic properties across the TRL by analyzing SWS measurements from teleseismic earthquakes recorded on a temporary seismic array, the Ruapehu And Taranaki Teleseismic Imaging Line (RATTIL) network between 2012-2014. SWS measurements revealed a strong NE-SW (42 ◦) oriented anisotropy across the TRL. The similarity of our fast azimuths to previous studies in eastern North Island where there is no wedge under the stations suggests that similar fast azimuths are found in both the mantle wedge and subslab mantle. The dominant trench-parallel orientation of our fast azimuth measurements are likely due to the NE-SW lattice-preferred orientation of olivine in the mantle wedge due to shear deformation associated with oblique convergence and trench-parallel mantle flow. Previous studies have observed apparent isotropy, dominantly west of the central volcanic region, and have suggested that isotropy may extend as far south as the TRL. Our results show that this region is anisotropic and we suggest that the boundary between apparent isotropy to the north and anisotropy in the south is located north of the RATTIL network (-38.75 ◦). Overall, this work represents an extensive analysis of seismic anisotropy across North Island. The results provide a detailed view of the lateral variations of seismic anisotropy across North Island and they inform us about the underlying anisotropic sources. Moreover, these results shed new light on the temporal and spatial dynamics of a slow slip event and reveal its mechanics. The gained insights will be useful for further development of SWS and future anisotropy analyses.