The upper 8-12 km of the Alpine Fault, South Island, New Zealand, accommodates relative Australia-Pacific plate boundary motion through coseismic slip accompanying large-magnitude earthquakes. Earthquakes occur due to frictional instabilities on faults, and their nucleation, propagation, and arrest is governed by tectonic forces and fault zone properties. A multi-disciplinary dataset is presented on the lithological, microstructural, mineralogical, geochemical, hydrological, and frictional properties of Alpine Fault rocks collected from natural fault exposures and from Deep Fault Drilling Project (DFDP-1) drillcore. Results quantify and describe the physical and chemical processes that affect seismicity and slip accommodation. Oblique dextral motion on the central Alpine Fault in the last 5-8 Myr has exhumed garnet-oligoclase facies mylonitic fault rocks from depths of up to 35 km. During the last phase of exhumation, brittle deformation of these mylonites, accompanied by fluid infiltration, has resulted in complex mineralogical and lithological variations in the fault rocks. Petrophysical, geochemical, and lithological data reveal that the fault comprises a central alteration zone of protocataclasites, foliated and nonfoliated cataclasites, and fault gouges bounded by a damage zone containing fractured ultramylonites and mylonites. Mineralogical results suggest that at least two stages of chemical alteration have occurred. At, or near, the brittle-to-ductile transition (c. >320 °C), metasomatic alteration reactions resulted in plagioclase and feldspar replacement by muscovite and sausserite, and biotite (phlogopite), hornblende (actinolite) and/or epidote replacement by chlorite (clinochlore). At lower temperatures (c. >120°C), primary minerals were altered to kaolinite, smectite and pyrite, or kaolinite, smectite, Fe-hydroxide (goethite) and carbonate, depending on redox conditions. Ultramylonites, nonfoliated and foliated cataclasites, and gouges in the hanging wall and footwall contain the high-temperature phyllosilicates chlorite and white mica (muscovite/illite). Brown principal slip zone (PSZ) gouges contain the low-temperature phyllosilicates kaolinite and smecite, and goethite and carbonate cements. The frictional and hydrological properties of saturated intact samples of central Alpine Fault surface-outcrop gouges and cataclasites were investigated in room temperature experiments conducted at 30-33 MPa effective normal stress (σn') using a double-direct shear configuration and controlled pore fluid pressure in a triaxial pressure vessel. Surface-outcrop samples from Gaunt Creek, location of DFDP-1, displayed, with increasing distance (up to 50 cm) from the contact with footwall fluvioglacial gravels: (1) an increase in fault normal permeability (k = 7.45 x 10⁻²⁰ m² to k = 1.15 x 10⁻¹⁶ m²), (2) a transition from frictionally weak (μ=0.44) fault gouge to frictionally strong (μ=0.50’0.55) cataclasite, (3) a change in friction rate dependence (a–b) from solely velocity strengthening to velocity strengthening and weakening, and (4) an increase in the rate of frictional healing. The frictional and hydrological properties of saturated intact samples of southern Alpine Fault surface-outcrop gouges were also investigated in room temperature double-direct shear experiments conducted at σn'= 6-31 MPa. Three complete cross-sections logged from outcrops of the southern Alpine Fault at Martyr River, McKenzie Creek, and Hokuri Creek show that dextral-normal slip is localized to a single 1-12 m-thick fault core comprising impermeable (k=10⁻²⁰ to 10⁻²² m²), frictionally weak (μ=0.12 – 0.37), velocity-strengthening, illite-chlorite and trioctahedral smectite (saponite)-chlorite-lizardite fault gouges. In low velocity room temperature experiments, Alpine Fault gouges tested have behaviours associated with aseismic creep. In a triaxial compression apparatus, the frictional properties of PSZ gouge samples recovered from DFDP-1 drillcore at 90 and 128 m depths were tested at temperatures up to T=350°C and effective normal stresses up to σn'=156 MPa to constrain the fault's strength and stability under conditions representative of the seismogenic crust. The chlorite/white mica-bearing DFDP-1A blue gouge is frictionally strong (μ=0.61–0.76) across a range of experimental conditions (T=70–350°C, σn'=31.2–156 MPa) and undergoes a stability transition from velocity strengthening to velocity weakening as T increases past 210°C, σn'=31.2–156 MPa. The coefficient of friction of smecite-bearing DFDP-1B brown gouge increases from μ=0.49 to μ=0.74 with increasing temperature and pressure (T=70–210°C, σn'=31.2–93.6 MPa) and it undergoes a transition from velocity strengthening to velocity weakening as T increases past 140°C, σn'=62.4 MPa. In low velocity hydrothermal experiments, Alpine Fault gouges have behaviours associated with potentially unstable, seismic slip at temperatures ≥140°C, depending on mineralogy. High-velocity (v=1 m/s), low normal stress (σn=1 MPa) friction experiments conducted on a rotary shear apparatus showed that the peak coefficient of friction (μp) of Alpine Fault cataclasites and fault gouges was consistently high (mean μp=0.69±0.06) in room-dry experiments. Variations in fault rock mineralogy and permeability were more apparent in experiments conducted with pore fluid, wherein the peak coefficient of friction of the cataclasites (mean μp=0.64±0.04) was higher than the fault gouges (mean μp=0.24±0.16). All fault rocks exhibited very low steady state coefficients of friction (μss) (room-dry mean μss=0.18±0.04; saturated mean μss=0.10±0.04). Three high-velocity experiments conducted on saturated smectite-bearing principal slip zone (PSZ) fault gouges had the lowest peak friction coefficients (μp=0.13-0.18), lowest steady state friction coefficients (μss=0.02-0.10), and lowest breakdown work values (WB=0.07-0.11 MJ/m²) of all the experiments performed. Lower strength (μ < c. 0.62) velocity-strengthening fault rocks comprising a realistically heterogeneous fault plane represent barrier(s) to rupture propagation. A wide range of gouges and cataclasites exhibited very low steady state friction coefficients in high-velocity friction experiments. However, earthquake rupture nucleation in frictionally strong (μ ≥ c. 0.62), velocity-weakening material provides the acceleration necessary to overcome the low-velocity rupture propagation barrier(s) posed by velocity-strengthening gouges and cataclasites. Mohr-Coulomb theory stipulates that sufficient shear stress must be resolved on the Alpine Fault, or pore fluid pressure must be sufficiently high, for earthquakes to nucleate in strong, unstable fault materials. A three-dimensional stress analysis was conducted using the average orientation of the central and southern Alpine Fault, the experimentally determined coefficient of friction of velocity-weakening DFDP-1A blue gouge, and the seismologically determined stress tensor and stress shape ratio(s). Results reveal that for a coefficient of friction of μ ≥ c. 0.62, the Alpine Fault is unfavourably oriented to severely misoriented for frictional slip.

This thesis presents new insights into Alpine Fault structures at the drill site of the Deep Fault Drilling Project (DFDP)-2B at Whataroa in New Zealand. Despite the challenging conditions for seismic imaging within a glacial valley filled with sediments and steeply dipping valley flanks, several structures related to the valley itself as well as the tectonic fault system are imaged. The Alpine Fault at the West Coast in New Zealand is a major plate boundary forming a significant geohazard as large earthquakes (magnitude 7-8) occur regularly and the next earthquake is expected relatively soon. A major effort has been made to study the fault characteristics through scientific drilling in the Deep Fault Drilling Project (DFDP) Alpine Fault with the deepest DFDP-2B borehole located in the Whataroa Valley. A great variety of seismic data are newly acquired. First, the WhataDUSIE (Whataroa Detailed University Seismic Imaging Experiment) data set is a ~5 km long 2D profile acquired in 2011 prior to the drilling. As the 2D profile could not fully explain the 3D structures in the Whataroa Valley, an extended surface and borehole data set was acquired in 2016 after the drilling. This data set consists of shorter 2D lines (< 3 km), a dense 3D-array, and vertical seismic profiling (VSP) using the DFDP-2B borehole including the fibre-optic cable. 3D seismic data proved to be essential to understand the complex 3D structures of the glacial valley and the major fault. First-arrival travel time tomography and prestack depth migration (PSDM) are applied to obtain a P-wave velocity model and seismic images of the subsurface (<5 km). In this complex setting, the Fresnel volume migration (a focusing PSDM method) proved to best obtain structural information about the subsurface. Analysing the results of the seismic data processing, two major outcomes are achieved: improved knowledge about the glacial structures of the Whataroa Valley and structural images of the Alpine Fault zone. The Whataroa Valley is an overdeepened glacial valley with details of the basement topography visible in the seismic images. A deep trough is identified south of the DFDP-2B borehole with horizontal layering of the sediments. Valley flanks are identified in both the seismic images and the P-wave velocity model, particularly the western valley flank. Thus, Quaternary and glacial processes can be analysed with the help of the newly derived seismic images. The Alpine Fault is directly imaged with the seismic data, which is the first time in this region at shallow depths (<5 km). Several shorter fault segments between depths of 0.2 km and 2.2 km dipping 40-56° to the southeast are directly imaged. Further identified reflectors and faults are interpreted to represent Alpine Fault structures in the form of a damage zone and induced faults adding further complexity to the fault zone. In conclusion, the 3D seismic results presented in this thesis provide new insights into the Whataroa subsurface. Hence, the new results form a good basis for a deeper understanding of the Alpine Fault structures and underlying processes which is important for potential future drilling but also for the estimation of the geohazard in the region.

The plate-boundary Alpine Fault runs immediately offshore of the popular tourist destination of Milford Sound, which is visited by more than half a million tourists each year. Glaciers retreated from the fiord between ~24-16 ka, leaving behind a legacy of extreme topography, including some of the world's highest sea cliffs, which tower nearly 2 km above the fiord. Visitors come to view the spectacularly steep and rugged landscape, with many cruising the fiord by boat. This project utilizes surface exposure dating (TCND) of glacially modified surfaces, to gain further insight into the glacier retreat history of Milford Sound. Exposure dates from strategic locations near the entrance to the fiord indicate that the main trunk glacier had retreated about 9 km from its peak LGM position by ~18 ka. Additional TCND and calibrated Schmidt Hammer data from a range of positions within the Milford catchment provide strong evidence that the main trunk glacier receded rapidly after about 18 ka, retreating a further 16 km to a position near the present-day confluence of the Tutoko and Cleddau rivers, by ~16 ka. Available seismic reflection data suggest that post-glacial sediment infill has been strongly influenced by massive deposits of rock avalanche debris. New high-resolution bathymetric and seismic reflection data reveals the presence of at least 18 very large post-glacial rock avalanche deposits which blanket ~40% of the fiord bottom. Geomorphic mapping and field investigation reveal the presence of at least ten additional very large to giant terrestrial landslide deposits in the lower Milford catchment; radiocarbon and surface exposure dating indicate that these events occurred during the Holocene, between ~9-1 ka. Ages of six of these deposits are in agreement with published rupture dates on the southern on-shore portion of the Alpine Fault.