Lithospheric Structure of the Hikurangi Subduction Margin

In this thesis, controlled-source seismic data acquired during two regional-scale experiments are analysed to determine the offshore lithospheric structure at the Hikurangi subduction margin in New Zealand. Subduction of the ∼120 Myr old Hikurangi Plateau occurs beneath the east coast of North Island, New Zealand. Because the plateau is an oceanic large igneous province, where the crustal thickness is about 50% greater than normal oceanic crust, there are different dynamics and seismicity patterns compared to the subduction of a regular oceanic crust. On interseismic time-scales, the plate interface in the south is locked down to depths of 30 km and experiences deep (30-45 km) slow-slip events (SSEs). In contrast, the plate interface is creeping in the north and experiences shallow (5-10 km) SSEs. It is important to understand the seismic velocity structure from the upper crust down to the base of the lithosphere in order to gain insights into the observed variations in subduction thrust slip behaviour, SSEs and the structure of an oceanic plateau in general. The thesis consists of three projects, each focusing on different aspects of the lithosphere at the Hikurangi subduction margin.

Project one investigates the crustal and upper mantle structure of the subducting Hikurangi Plateau at the southern Hikurangi margin. In this study, onshore-offshore seismic data acquired during the Seismic Array HiKurangi Experiment (SAHKE) are used. By forward-model raytracing the travel-times of observed refractions and wide-angle reflections in common receiver gathers, the Hikurangi Plateau crustal thickness is estimated to be 12±1 km. A ∼10% reduction in P-wave-speeds in the Hikurangi Plateau crust beneath the trough is observed. Refractions provide evidence for two upper mantle layers: an upper layer with regular upper mantle P-wave-speeds (8.0±0.2 km/s); and a deeper layer with a high P-wave-speed (8.7±0.2 km/s) at a depth of 50±2 km beneath the Hikurangi trough. Similarly fast upper mantle P-wave speeds are reported along margin-parallel azimuths under the North Island, about 100 km down-dip of the subduction zone at depths of ∼8-10 km from the Moho suggesting that the upper mantle of the Hikurangi Plateau is characterised by anomalously high P-wave-speeds along all azimuths. A velocity reduction of ∼10%, similar to that in the crust, is deduced to extend down to 25±2 km in the upper mantle beneath the trough, as a result of the formation of a low-velocity zone in the faster upper mantle layer. It is proposed that this is due to the serpentinisation of mantle peridotite by hydration through bending-induced normal faults and/or due to crack porosity introduced by thermal cracking, further enhanced by bending-related faulting. This implies that the “regular mantle” (VP ∼8 km/s) is not regular, but rather the faster upper mantle has mechanically bent, fractured and altered. The onset of seismicity in the lower band of the double seismic zone and high upper mantle VP under the North Island is observed at depths of ∼50 km. This is consistent with the hypothesis that the lower band of earthquakes in a double seismic zone is due to antigorite dehydration processes, a hydrous mineral in the low-velocity zone in the upper mantle beneath the trough. Despite the differences in crustal thickness and high upper mantle P-wave speeds, subduction-related upper mantle hydration and dehydration are analogous with other margins where regular oceanic crust is subducting.

The second project is focused on a series of long-offset, late-arriving wide-angle seismic reflections observed in the onshore-offshore common receiver gathers from the first project. Results from modelling these wide-angle reflections using forward-model raytracing, amplitude versus offset modelling and synthetic waveform modelling, are consistent with a series of reflective horizons approaching sub-lithospheric depths of the subducting Pacific Plate. A ∼3-5 km thick, azimuthally anisotropic layer with a P-wave anisotropy of 13-15% is proposed to exist at a depth of 70 km. A 5 km thick layer with low P-wave velocity (7.6 km/s) and a high VP/VS ratio (>>2.5) is then required below the anisotropic layer. It is followed by another ∼3-5 km thick layer with slightly lower (7.4 km/s) or higher (7.8 km/s) P-wave velocity and a regular VP /VS ratio (∼1.85). The higher VP/VS ratio in the upper layer indicates that it contains either melt or volatiles, whereas the relatively low VP/VS ratio in the lower layer may indicate a relatively lower fluid content. These two layers comprise a composite low VP layer with a thickness of ∼8-10 km beneath the anisotropic layer, and is interpreted to be the lithosphere-asthenosphere boundary (LAB) channel. It is consistent with the down-dip projected depths of the LAB channel found from an earlier study. The most prominent discovery here is the azimuthally anisotropic layer whose fast azimuth is subparallel to the direction of absolute plate motion (perpendicular to the margin). Strong shearing occurring at the LAB channel due to the differential movement of the lithosphere on top of the asthenosphere is suggested to give the preferential alignment of olivine crystals along the direction of maximum finite shear strain and produce the observed azimuthal anisotropy. Results from this study show, therefore, that it is not a single low-velocity channel that makes up the LAB, but a series of layers that make up an LAB zone. In addition, the study highlights the key role that wide-angle reflections from controlled-sources can play in investigating the fine-scale structure of the LAB boundary zone due to the short wavelengths and the generation of enhanced amplitudes when the reflections approach the critical angles (∼55◦).

The primary objective of the third research project is to estimate the VP/VS ratio of the Hikurangi margin forearc using mode-converted seismic phases from airgun shots recorded by an array of multi-component ocean bottom seismographs (OBS) deployed as a part of the Seismogenesis at Hikurangi Integrated Research Experiment (SHIRE). PPS mode-conversions at the sediment-basement interface and the top of the subducting crust are identified. Estimated average VP/VS ratios for the topmost sediments range from 2.5±1.0 to 6.0±2.5 and are consistent with water-saturated, unconsolidated sediments. Average VP/VS ratios for the entire column of sediments and sedimentary rocks above the subducting crust range from ∼1.55±0.08 to 2.20±0.08. Low-average VP/VS ratios between ∼1.55±0.08 and ∼1.78±0.12 are estimated for a region of higher sediment thickness in the southern Hikurangi margin. The thick sediments may result in a higher degree of compaction. The low VP/VS ratios are also coincident with the offshore extension of the Pahau Torlesse Terrane which consists mainly of low-porosity, highly compacted, Cretaceous greywackes. In contrast, high VP/VS ratios between ∼1.85±0.10 and 2.22±0.08 are observed in regions with lower sediment thickness, which may reflect effects of lower degree of compaction and lithology. Furthermore, the average VP/VS of the rocks and sediments above the subducting crust show a weak correlation with the slip-rate deficits on the subduction thrust. Shear-wave splitting results indicate an anisotropy of ∼3.5% localised in the top layer (∼1-2 km) of sediments beneath the seafloor. Fast polarisation directions are oriented perpendicular to the plate interface contours, suggesting stress-aligned, fluid-filled cracks.

The work presented in this thesis provides constraints on the lithospheric structure of the Hikurangi subduction margin, from the upper plate down to the base of the lithosphere, using controlled-source seismology. The results provide insights on the physical properties of the materials and their association with geodynamic processes. The outcomes of this thesis advance our knowledge of the Hikurangi subduction margin and contribute to our knowledge of plate tectonics in general.