The development and activity of submarine canyons on continental margins is strongly influenced by temporal and spatial changes in sediment distribution associated with orbitally-forced sea-level cyclicity. On active margins, canyons are also strongly influenced by tectonic processes such as faulting, uplift and earthquakes. Within this framework the role of mass-wasting processes, including sediment failures, bedrock landslides and sediment gravity flows, are to: 1) transport material across the slope; 2) act as intra-slope sediment sources; and 3) shape seafloor morphology. In this project the seafloor-landscape signatures of tectonic and geomorphic processes are analysed to interpret the development of submarine canyon morphology on active margins. Datasets include high-resolution bathymetry data (Simrad EM300), multichannel seismic reflection data (MCS), high-resolution 3.5 kHz seismic reflection data, sediment cores, and dated seafloor samples. High-resolution bathymetric grids are analysed using techniques developed for terrain-roughness analysis in terrestrial landscapes to objectively map and interpret features related to seafloor mass-wasting processes. The Hikurangi subduction margin of New Zealand provides world-class examples of the control of tectonic and sedimentary processes on margin development, hosting multiple examples of deeply-incised canyon systems across a range of scales. Two main study sites, in Poverty Bay and Cook Strait, provide examples of canyon formation. From these examples conceptual and representative models are developed for the spatial and temporal relationships between active tectonic structures, geology, sediment supply, slope- and shelf-incised canyons, slope gully systems, and bedrock mass failures. The Poverty Bay site occurs on the subduction-dominated northern Hikurangi Margin, where the ~3000 km² Poverty re-entrant hosts the large Poverty Canyon system, the only shelf-break-to-subduction-trough canyon on the northern margin. The geomorphic development of the re-entrant is affected by gully development on the upper slope, and multi-cubic-kilometre-scale submarine landslides. From this site the study focuses on the initiation and development of upper-slope gullies and the role of deep-seated slope failure in upper-slope evolution. The Cook Strait site occurs on the southern Hikurangi Margin in the subduction-to-strike-slip transition zone. The 1800 km² Cook Strait Canyon incises almost 50 km into the continental shelf, with a multi-branching canyon head converging to a deeply slope-incised meandering main channel fed by multiple contributing slope canyons. Other medium-sized canyons are incised into the adjacent continental slope. Fluvial sediment supply to the coast is relatively low on the southern margin, but Cook Strait is subject to large diurnal tidal currents that mobilise sediment through the main strait area. Prior to the morphostructural analysis of the Cook Strait and Poverty study sites a revision of the tectonic structure was undertaken. In Cook Strait a revision of the available fault maps was undertaken as part of a wider, related tectonic study of the central New Zealand region. In Poverty Bay very limited prior information was available, and as part of this study the structure and stratigraphy of the entire shelf and upper slope has been interpreted. On active tectonic margins submarine canyons respond to tectonics at: 1) margin-setting scales relating to their ability to become shelf incised; 2) regional scales relating to canyon-incision response to base-level perturbations; and 3) local scales relating to propagating structures affecting canyon location and geometry. Interpretation of the spatial distribution of fluid vent sites, gully development and landslide scars leads to the conclusion that seepage-driven failure is not a primary control on the widespread instances of gully formation and landslide erosion affecting structurally-generated relief across the margin. Rather, the erosion of tectonic ridges is dominated by tectonics by: slope oversteepening; weakening of the rockmass in fault-damage zones; and triggering of slope failure by earthquake-generated cyclic loading. Deep-seated mass failures affect numerous aspects of submarine landscapes and play a major role in the enlargement of canyon systems. They enable the development of slope gully systems and represent a major intra-slope sediment source. Quantitative morphometric analysis together with MCS data indicate that landslides may evolve to be active complexes where landslide debris is remobilized repeatedly, analogous to terrestrial-earthflow processes. This process has not previously been documented on submarine slopes. A model is presented for the evolution of active margin canyons that contrasts highstand and lowstand canyon activity in terms of channel incision, sedimentary processes and slope-erosion processes. During sea-level highstand intervals, canyons become decoupled from their terrestrial/coastal sediment-supply source areas, while during sea-level lowstand intervals, canyons are coupled to fluvial and littoral sediment-supply sources, and top-down (i.e. shelf-to-lower-slope) sediment transport and channel incision is active. Canyon-head areas are incision dominated during the lowstand while mid to lower canyon reaches experience both a transient increase in sediment in storage and canyon-fill degradation and incision into bedrock. Tectonics influences the canyon landscape through both uplift-controlled perturbations to canyon base-levels and earthquake-triggering of mass movement. Following sea-level rise the sediment supply to canyon heads will be switched off at a certain threshold sea level. From this point canyon heads become aggradational. Mid to lower canyon reaches continue to incise due to continuing tectonic uplift and earthquake-triggered slope instability. Knickpoints are propagated up channel and excavate canyon and sub-canyon channels from the bottom up. Thus, while top-down infilling of non-coupled canyons occurs during sea-level highstands, the lower reaches of active margin canyons continue to incise due the influence of tectonic processes.

This thesis investigates normal faulting and its influence on fluid flow over a wide range of spatial and temporal scales using tunnel engineering geological logs, outcrop, surface fault traces, earthquakes, gravity, and volcanic ages. These data have been used to investigate the impact of faults on fluid flow (chapter 2), the geometry and kinematics of the Taupo Rift (chapter 3), the hydration and dehydration of the subducting Pacific plate and its influence on the Taupo Volcanic Zone (chapter 4), the migration of arc volcanism across the North Island over the 16 Myr and the associated changes in slab geometry (chapter 5) and the Pacific-Australia relative plate motion vectors since 38 Ma and their implications for arc volcanism and deformation along the Hikurangi margin (chapter 6). The results for each of these five chapters are presented in the five paragraphs below. Tunnels excavated along the margins of the southern Taupo Rift at depths < 500 m provide data on the spatial relationships between faulting and ground water flow. The geometry and hydraulic properties of fault-zones for Mesozoic basement and Miocene strata vary by several orders of magnitude approximating power-law distributions with the dimensions of these zones dependent on many factors including displacement, hostrock type and fault geometries. Despite fault-zones accounting for a small proportion of the total sample length (≤ 15%), localised flow of ground water into the tunnels occurs almost exclusively (≥ 91%) within, and immediately adjacent to, these zones. The spatial distribution and rate of flow from fault-zones are highly variable with typically ≤ 50% of fault-zones in any given orientation flowing. The entire basement dataset shows that 81% of the flow-rate occurs from fault-zones ≥ 10 m wide, with a third of the total flow-rate originating from a single fault-zone (i.e. the golden fracture). The higher flow rates for the largest faults are interpreted to arise because these structures are the most connected to other faults and to the ground surface. The structural geometry and kinematics of rifting is constrained by earthquake focal mechanisms and by geological slip and fault mapping. Comparison of present day geometry and kinematics of normal faulting in the Taupo Rift (α=76-84°) with intra-arc rifting in the Taranaki Basin and southern Havre Trough show, that for at least the last 4 Myr, the slab and the associated changes in its geometry have exerted a first-order control on the location, geometry, and extension direction of intra-arc rifting in the North Island. Second-order features of rifting in the central North Island include a clockwise ~20° northwards change in the strike of normal faults and trend of the extension direction. In the southern rift normal faults are parallel to, and potentially reactivate, Mesozoic basement fabric (e.g., faults and bedding). By contrast, in the northern rift faults diverge from basement fabric by up to 55° where focal mechanisms indicate that extension is achieved by oblique to right-lateral strike-slip along basement fabric and dip-slip on rift faults. Hydration and dehydration of the subducting Pacific plate is elucidated by earthquake densities and focal mechanisms within the slab. The hydration of the subducting plate varies spatially and is an important determinant for the location of arc volcanism in the overriding plate. The location and high volcanic productivity of the TVZ can be linked to the subduction water cycle, where hydration and subsequent dehydration of the subducting oceanic lithosphere is primarily accomplished by normal-faulting earthquakes. The anomalously high heat flow and volcanic productivity of the TVZ is spatially associated with high rates of seismicity in the underlying slab mantle at depths of 130-210 km which can be tracked back to high rates of deeply penetrating shallow intraplate seismicity at the trench in proximity to oceanic fluids. Dehydration of the slab mantle correlates with the location and productivity of active North Island volcanic centres, indicating this volcanism is controlled by fluids fluxing from the subducting plate. The ages and locations of arc volcanoes provide constraints on the migration of volcanism across the North Island over the last 20 Myr. Arc-front volcanoes have migrated southeast by 150 km in the last 8 Ma (185 km since 16 Ma) sub-parallel to the present active arc. Migration of the arc is interpreted to mainly reflect slab steepening and rollback. The strike of the Pacific plate beneath the North Island, imaged by Benioff zone seismicity (50-200 km) and positive mantle velocity anomalies (200-600 km) is parallel to the northeast trend of arc-front volcanism. Arc parallelism since 16 Ma is consistent with the view that the subducting plate beneath the North Island has not rotated clockwise about vertical axes which is in contrast to overriding plate vertical-axis rotations of ≥ 30º. Acceleration of arc-front migration rates (~4 mm/yr to ~18 mm/yr), eruption of high Mg# andesites, increasing eruption frequency and size, and uplift of the over-riding plate indicate an increase in the hydration, temperature, and size of the mantle wedge beneath the central North Island from ~7 Ma. Seafloor spreading data in conjunction with GPlates have been used to generate relative plate motion vectors across the Hikurangi margin since 38 Ma. Tracking the southern and down-dip limits of the seismically imaged Pacific slab beneath the New Zealand indicates arc volcanism in Northland from ~23 Ma and the Taranaki Basin between ~20 and 11 Ma requires Pacific plate subduction from at (or beyond) the northern North Island continental margin from at least 38 Ma to the present. Pacific plate motion in a west dipping subduction model shows a minimum horizontal transport distance of 285 km preceding the initiation of arc volcanism along the Northland-arc normal to the motion vector, a distance more than sufficient for self-sustaining subduction to occur. Arc-normal convergence rates along the Hikurangi margin doubled from 11 to 23 mm/yr between 20 and 16 Ma, increasing again by approximately a third between 8 and 6 Ma. This latest increase in arc-normal rates coincided with changes in relative plate motions along the entire SW Pacific plate boundary and steepening/rollback of the Pacific plate.

The development and activity of submarine canyons on continental margins is strongly influenced by temporal and spatial changes in sediment distribution associated with orbitally-forced sea-level cyclicity. On active margins, canyons are also strongly influenced by tectonic processes such as faulting, uplift and earthquakes. Within this framework the role of mass-wasting processes, including sediment failures, bedrock landslides and sediment gravity flows, are to: 1) transport material across the slope; 2) act as intra-slope sediment sources; and 3) shape seafloor morphology. In this project the seafloor-landscape signatures of tectonic and geomorphic processes are analysed to interpret the development of submarine canyon morphology on active margins. Datasets include high-resolution bathymetry data (Simrad EM300), multichannel seismic reflection data (MCS), high-resolution 3.5 kHz seismic reflection data, sediment cores, and dated seafloor samples. High-resolution bathymetric grids are analysed using techniques developed for terrain-roughness analysis in terrestrial landscapes to objectively map and interpret features related to seafloor mass-wasting processes. The Hikurangi subduction margin of New Zealand provides world-class examples of the control of tectonic and sedimentary processes on margin development, hosting multiple examples of deeply-incised canyon systems across a range of scales. Two main study sites, in Poverty Bay and Cook Strait, provide examples of canyon formation. From these examples conceptual and representative models are developed for the spatial and temporal relationships between active tectonic structures, geology, sediment supply, slope- and shelf-incised canyons, slope gully systems, and bedrock mass failures. The Poverty Bay site occurs on the subduction-dominated northern Hikurangi Margin, where the ~3000 km² Poverty re-entrant hosts the large Poverty Canyon system, the only shelf-break-to-subduction-trough canyon on the northern margin. The geomorphic development of the re-entrant is affected by gully development on the upper slope, and multi-cubic-kilometre-scale submarine landslides. From this site the study focuses on the initiation and development of upper-slope gullies and the role of deep-seated slope failure in upper-slope evolution. The Cook Strait site occurs on the southern Hikurangi Margin in the subduction-to-strike-slip transition zone. The 1800 km² Cook Strait Canyon incises almost 50 km into the continental shelf, with a multi-branching canyon head converging to a deeply slope-incised meandering main channel fed by multiple contributing slope canyons. Other medium-sized canyons are incised into the adjacent continental slope. Fluvial sediment supply to the coast is relatively low on the southern margin, but Cook Strait is subject to large diurnal tidal currents that mobilise sediment through the main strait area. Prior to the morphostructural analysis of the Cook Strait and Poverty study sites a revision of the tectonic structure was undertaken. In Cook Strait a revision of the available fault maps was undertaken as part of a wider, related tectonic study of the central New Zealand region. In Poverty Bay very limited prior information was available, and as part of this study the structure and stratigraphy of the entire shelf and upper slope has been interpreted. On active tectonic margins submarine canyons respond to tectonics at: 1) margin-setting scales relating to their ability to become shelf incised; 2) regional scales relating to canyon-incision response to base-level perturbations; and 3) local scales relating to propagating structures affecting canyon location and geometry. Interpretation of the spatial distribution of fluid vent sites, gully development and landslide scars leads to the conclusion that seepage-driven failure is not a primary control on the widespread instances of gully formation and landslide erosion affecting structurally-generated relief across the margin. Rather, the erosion of tectonic ridges is dominated by tectonics by: slope oversteepening; weakening of the rockmass in fault-damage zones; and triggering of slope failure by earthquake-generated cyclic loading. Deep-seated mass failures affect numerous aspects of submarine landscapes and play a major role in the enlargement of canyon systems. They enable the development of slope gully systems and represent a major intra-slope sediment source. Quantitative morphometric analysis together with MCS data indicate that landslides may evolve to be active complexes where landslide debris is remobilized repeatedly, analogous to terrestrial-earthflow processes. This process has not previously been documented on submarine slopes. A model is presented for the evolution of active margin canyons that contrasts highstand and lowstand canyon activity in terms of channel incision, sedimentary processes and slope-erosion processes. During sea-level highstand intervals, canyons become decoupled from their terrestrial/coastal sediment-supply source areas, while during sea-level lowstand intervals, canyons are coupled to fluvial and littoral sediment-supply sources, and top-down (i.e. shelf-to-lower-slope) sediment transport and channel incision is active. Canyon-head areas are incision dominated during the lowstand while mid to lower canyon reaches experience both a transient increase in sediment in storage and canyon-fill degradation and incision into bedrock. Tectonics influences the canyon landscape through both uplift-controlled perturbations to canyon base-levels and earthquake-triggering of mass movement. Following sea-level rise the sediment supply to canyon heads will be switched off at a certain threshold sea level. From this point canyon heads become aggradational. Mid to lower canyon reaches continue to incise due to continuing tectonic uplift and earthquake-triggered slope instability. Knickpoints are propagated up channel and excavate canyon and sub-canyon channels from the bottom up. Thus, while top-down infilling of non-coupled canyons occurs during sea-level highstands, the lower reaches of active margin canyons continue to incise due the influence of tectonic processes.

Cold seep ecosystems in the deep sea are fuelled by chemosynthetic processes based on methane emission to the sediment surface from gas hydrate disassociation, methanogenesis or thermogenic processes. While cold seep ecosystems have been studied in the last three decades worldwide, little is known about New Zealand’s cold seep habitats and associated fauna. A joint German-New Zealand cruise to the Hikurangi Margin in early 2007 enabled biological and sediment sampling to investigate the biological and sedimentological relationships and variability of seeps and their faunal diversity. Multi-disciplinary approaches were employed that included Xray radiography, stratigraphic descriptions, lebensspuren traces analysis, sediment grain size analysis, determination of total organic content, carbonate content and its stable isotopic composition, and analysis of benthic invertebrate assemblages of seep habitats. The results of this study revealed three distinctive habitats and associated fauna based on the sediment characteristics and faunal type. Habitat 1 includes all sites pertaining to Omakere Ridge, a seep-related habitat comprised of layers of very poorly sorted, sandy silt, shell hash and bands of methane-derived authigenic aragonitic carbonate nodules with low total organic content (TOC). Due to the characteristics of the sediments and death assemblages of molluscs, it is inferred that Habitat 1 methane seepage is actively diffusive, waning or dormant. Habitat 2 describes sites that are either non-seep or relic and applies to those at Bear’s Paw and Kaka. Habitat 2 constituted of shell hash overlain with very poorly sandy silt, and low carbonates content and low to medium TOC. Habitat 3 describes non-seep related habitats, and includes all sites of the Wairarapa region and one reference site from Kaka also falls into this category. Sediments for Habitat 3 constituted poorly sorted silt with high TOC and low carbonate content which can be explained by their close proximity to land and converging sea currents. The mineral components of the background siliciclastic sediments for all sites studied originated in the Tertiary mudstone of the East Coast Basin. The characteristics of seep habitats of the Hikurangi Margin were comparable to that of the Northern Hemisphere modern seep counterparts, although the abundance and distributions of seep fauna were low. Results from this research have enhanced our understanding on the spatial and variability of methane fluxes and their affects on the duration of cold seep ecosystems, especially for New Zealand. However, more such studies are essential to increase our understanding of seep sediments and explain disturbance-sediment-benthic invertebrate interactions.

Cold seep ecosystems in the deep sea are fuelled by chemosynthetic processes based on methane emission to the sediment surface from gas hydrate disassociation, methanogenesis or thermogenic processes. While cold seep ecosystems have been studied in the last three decades worldwide, little is known about New Zealand’s cold seep habitats and associated fauna. A joint German-New Zealand cruise to the Hikurangi Margin in early 2007 enabled biological and sediment sampling to investigate the biological and sedimentological relationships and variability of seeps and their faunal diversity. Multi-disciplinary approaches were employed that included Xray radiography, stratigraphic descriptions, lebensspuren traces analysis, sediment grain size analysis, determination of total organic content, carbonate content and its stable isotopic composition, and analysis of benthic invertebrate assemblages of seep habitats. The results of this study revealed three distinctive habitats and associated fauna based on the sediment characteristics and faunal type. Habitat 1 includes all sites pertaining to Omakere Ridge, a seep-related habitat comprised of layers of very poorly sorted, sandy silt, shell hash and bands of methane-derived authigenic aragonitic carbonate nodules with low total organic content (TOC). Due to the characteristics of the sediments and death assemblages of molluscs, it is inferred that Habitat 1 methane seepage is actively diffusive, waning or dormant. Habitat 2 describes sites that are either non-seep or relic and applies to those at Bear’s Paw and Kaka. Habitat 2 constituted of shell hash overlain with very poorly sandy silt, and low carbonates content and low to medium TOC. Habitat 3 describes non-seep related habitats, and includes all sites of the Wairarapa region and one reference site from Kaka also falls into this category. Sediments for Habitat 3 constituted poorly sorted silt with high TOC and low carbonate content which can be explained by their close proximity to land and converging sea currents. The mineral components of the background siliciclastic sediments for all sites studied originated in the Tertiary mudstone of the East Coast Basin. The characteristics of seep habitats of the Hikurangi Margin were comparable to that of the Northern Hemisphere modern seep counterparts, although the abundance and distributions of seep fauna were low. Results from this research have enhanced our understanding on the spatial and variability of methane fluxes and their affects on the duration of cold seep ecosystems, especially for New Zealand. However, more such studies are essential to increase our understanding of seep sediments and explain disturbance-sediment-benthic invertebrate interactions.

Cold seep-carbonates are the microbially mediated by-products of the anaerobic oxidation of methane (AOM) at seafloor cold seeps, and are widespread about modern continental margins and in the geologic record. Some modern and Miocene examples of cold seep-carbonates from the East Coast Basin, North Island, New Zealand have been analysed in this study, to characterise and determine their carbonate fabrics, elemental and mineralogical composition, and stable δ18O and δ13C isotope signatures, so as to provide insights into the diagenetic changes associated with the lithification and burial of seep-carbonates. The ancient samples were collected from the onshore middle Miocene Tauwhareparae (TWP) seep deposit, while the modern samples were obtained from the National Institute of Water and Atmosphere (NIWA) Cruise TAN0616 (November 2006) from Ritchie Ridge, offshore Hikurangi Margin. A paragenetic sequence of diagenetic events involving early aragonitic phases, followed by late calcitic phases is defined for the seep-carbonates. This sequence likely has relevance for understanding the fluid-cement histories of seep-carbonates more widely. Two main carbonate mineralogies occur in each of the sample groups - modern samples are aragonitic or dolomitic, while the ancient ones consist dominantly of either aragonite or calcite. Thus, aragonite common to both sample groups, and is interpreted to represent the initial primary carbonate precipitate in hydrocarbon seep provinces under specific fluid flux and local pore-water chemistry conditions. Aragonite morphologies range from microcrystalline carbonate ('micarb'), to acicular aragonites that may form botryoids or spherulites. Dolomite occurs in those modern samples which appear to constitute exhumed remnants of a former subsurface 'seep plumbing system', and so are strictly not true seabed 'seep-carbonates', but instead are part of the larger hydrocarbon seep province. Calcite in the ancient samples is either a product of alteration and neomorphic transformation of aragonite, or derives from late stage cementation from burial fluids. As a result of their formation processes, the calcites are generally recrystallised and have equant or 'cellular' textures. Stable δ13C and δ18O isotope cross-plots reveal a large spread of values for the sample groups. Ancient samples range from δ13C -8 to -50 PDB and δ18O -5.5 to +2 PDB. Modern samples have δ13C values from -6 to -41 PDB and δ18O values ranging from +2.6 to +6.7 PDB. The δ13C values suggest the majority of the methane that formed these seep-carbonates is of thermogenic origin, although some mixing from other carbon sources may have occurred. The positive δ18O signatures are suggestive of carbonate formation during dissociation of gas hydrates, while the negative values possibly indicate that some of the formation fluids were warmer than normal in the 17 - 30 C range.