This study describes the character of submarine mass movement and associated deformation as revealed by an exceptionally well-exposed portion of a seismic-scale mass-transport deposit (MTD) within the upper Miocene Mohakatino Formation (Taranaki Basin, New Zealand). The North Awakino MTD is at least 55 m thick and crops out along the northern Taranaki coastline for ∼11 km in wave-cut platforms and in cliffs as much as 100 m high. Spectacular soft-sediment deformation features are developed in remobilized sediment gravity flow deposits that initially accumulated within a low-gradient intraslope basin. Sedimentary facies within the North Awakino MTD comprise laterally extensive, thick- to thin-bedded volcaniclastic sandstone and mudstone. Distinct postdepositional deformation styles are associated with bedding type: folds developed in thick-bedded sandstone are larger (fold heights to tens of meters) and more laterally continuous (to 1 km) than those developed in thinner bedded facies.
Regional geologic relationships suggest that nearly the full width of the North Awakino MTD is exposed in outcrop, providing a rare opportunity to observe lateral relationships between the marginal and central portions of the MTD. We conducted a rigorous paleoslope analysis of slump fold, fault, and bedding orientations using both existing and newly proposed methodologies. Separate analysis of seven subregions within the North Awakino MTD reveals that the predicted MTD transport direction varies widely along the outcrop extent. Most notably, slump folds and faults within the inferred margins have mean orientations that are suborthogonal to those within the central portions of the MTD. This relationship is hypothesized to be a consequence of edge effects that may be related to lateral compression along the margins of the MTD. Our analysis demonstrates the importance of accounting for spatial heterogeneity in slump structure orientations when determining the paleoslope orientation through kinematic analysis. Our inference of west-directed translation suggests that the North Awakino MTD formed in response to a local change in the bathymetric slope orientation that was likely the result of tectonically induced basin deformation.
Submarine landslides, or mass-transport deposits (MTDs), have become well documented in bathymetric and seismic-reflection data sets and are recognized to be important agents of sediment delivery and reorganization along the world’s continental slopes (Nardin et al., 1979; Haflidason et al., 2004; Normark et al., 2004; Posamentier and Martinsen, 2011, and references within). In particular, improved documentation of MTDs within slope and basin-floor settings has emphasized their importance as a fundamental building block of deep-marine stratigraphic sequences (Moscardelli et al., 2006; Hubbard et al., 2010; Shipp et al., 2011). For this reason, study of both modern and ancient MTDs has specific relevance to basin and petroleum systems analysis, geotechnical hazards to submarine infrastructure, and tsunami-generating potential (Moore et al., 1989, 1994; Locat and Lee, 2002; Masson et al., 2006).
Exposures of MTDs of a scale similar to that observed on the seafloor or within the subsurface are relatively uncommon in the ancient record (Woodcock, 1979a; Macdonald et al., 1993; Lucente and Pini, 2008; Odonne et al., 2011). Most ancient outcrops only preserve a fraction of the original MTD extent, and thus published outcrop studies lack detailed information of the large-scale geometries that are commonly observed in remote sensing data sets (e.g., Posamentier and Martinsen, 2011). Although many outcrop studies of MTDs have documented changes in MTD deformation from headwall to toe (e.g., Farrell, 1984; Gawthorpe and Clemmey, 1985; Farrell and Eaton, 1987), few have documented the lateral relationships between the marginal and central portions of MTDs (Debacker et al., 2009). In general, knowledge of how MTD deformation structures (e.g., slump folds and faults) change in character or orientation laterally is often hampered by a lack of regional context and/or subsequent tectonic deformation.
This paper aims to document the internal architecture, kinematic history, and spatial patterns of deformation within a seismic-scale MTD that is spectacularly exposed for ∼11 km along the northern Taranaki coastline of New Zealand, referred to here as the North Awakino mass-transport deposit (NAMTD) (King et al., 2011; Fig. 1). The NAMTD consists of remobilized sediment gravity flow deposits that were deposited within a deep-marine (bathyal) depocenter within the larger Taranaki Basin during late Miocene time (Utley, 1987; King et al., 1993). Subsequent submarine mass movement has resulted in a wide range of ductile and brittle deformation structures that include spectacular slump folds and faults (sensu Martinsen, 1994). Several factors contribute to the importance and uniqueness of the NAMTD: (1) excellent semicontinuous cross-sectional exposure (in cliffs as high as ∼100 m) and plan-view exposure (in wave-cut platforms) over the extent of the study area; (2) preservation of a nearly complete lateral cross section through the compressional domain that preserves both marginal and central portions within an MTD; (3) limited tectonic deformation; and (4) regional context provided by two-dimensional (2D) and 3D seismic-reflection surveys that are located ≥3 km offshore of the study area (e.g., King et al., 2011).
This study provides nearly unique insight, based on detailed outcrop observations and measurements, into how deformation styles and geometry change laterally within an MTD. For example, we address the question of how the orientations of slump deformation structures (herein slump structures) vary with regard to their location within a deforming MTD and consider implications for estimation of the paleoslope orientation via analysis of slump folds and faults (i.e., paleoslope analysis; Woodcock, 1979b). Correspondingly, a major goal of this study is to review and critically examine methods of paleoslope analysis that have been previously applied to the lateral and central portions of MTDs (e.g., Debacker et al., 2009; Alsop and Marco, 2012). In doing so, we highlight both the utility and drawbacks of particular paleoslope methods and demonstrate the importance of incorporating spatial context into kinematic analysis of slump structures.
An important secondary goal of this study is to consider the implications of mass movement on the paleogeographic setting and tectonic history of the northeastern Taranaki Basin. We aim to resolve an ongoing debate about the transport direction of the NAMTD that has been variously interpreted as east-, south-, or west-directed (Utley 1987; Nagel, 2010; King et al., 2011) by conducting a comprehensive paleoslope analysis using both existing and new methodologies (e.g., Woodcock, 1979b; Strachan and Alsop, 2006). We also incorporate sedimentological observations from both within the NAMTD and from surrounding units to constrain the depositional setting prior to subsequent remobilization. In doing so, we document the internal architecture of postdepositional, soft-sediment deformation of sand-rich deep-water lobes (sensu Prélat et al., 2009), and provide a type example of tectonically induced mass movement within a basin-floor sedimentary succession.
The NAMTD is located within the northeastern Taranaki Basin (Fig. 1), a Cretaceous to Holocene sedimentary succession that has a complex and multiphase basin history. The Taranaki Basin initiated during Late Cretaceous time, when the New Zealand subcontinent rifted from Gondwana (eastern Australia) and thereafter developed into a passive margin setting (King and Thrasher, 1992, 1996; King, 2000). The fledgling modern Australian-Pacific plate boundary subsequently propagated through the New Zealand region from middle Eocene time (King, 2000). Convergence in the north started around late Eocene time (Bache et al., 2012) and was manifested in the Taranaki Basin as emplacement of the west-verging Taranaki thrust fault, which displaced basement more than 7 km vertically (Fig. 1; King and Thrasher, 1996; Stagpoole and Nicol, 2008). The eastern Taranaki Basin may have evolved into an underfilled foreland during that time as a response to crustal loading to the east (King and Thrasher, 1992; Holt and Stern, 1994). Major activity on the Taranaki fault in the vicinity of the study area is thought to have ended by early Miocene time (Stagpoole and Nicol, 2008), and a sedimentary succession subsequently developed within a bathyal depocenter defined by uplifted basement to the east and southeast (Herangi-Patea-Tongaporutu submarine highs) and by a submarine andesitic volcanic arc to the west (Figs. 1 and 2; Nodder et al., 1990b; King and Thrasher, 1992; King et al., 1993, 2007a, 2007b).
The NAMTD was emplaced within the basal portion (Mohakatino Formation) of an ∼2-km-thick, upward-shoaling sedimentary succession of late Miocene (Tortonian) age that is well exposed along the Taranaki coast for more than 65 km south of Tirua Point (Figs. 2 and 3). The Mohakatino Formation comprises volcaniclastic gravel, sandstone, and mudstone thought to have been derived from the submerged chain of stratovolcanoes (ca. 15–6.5 Ma) to the north and west (King and Thrasher, 1996; Giba et al., 2013; Fig. 1). Foraminiferal assemblages suggest that the Mohakatino Formation was deposited in lower bathyal water depths (>1500 m), and some suggested that this unit is composed of both sediment gravity flow deposits and suspension sedimentation of volcanic debris through the water column within a basin-floor fan or fan apron setting (Utley, 1987; Nodder et al., 1990b; King et al., 1993, 2011). The Mohakatino Formation is overlain by the Mount Messenger, Urenui, and Kiore Formations that record a pronounced progradation of the continental slope to the north and west from ca. 10.5 to 9 Ma (King et al., 1993, 1994; Browne and Slatt, 2002; Arnot et al., 2007a, 2007b; Browne et al., 2000, 2007a, 2007b; King et al., 2007a, 2007b, 2007c; Maier, 2012; Masalimova, 2013; Rotzien et al., 2014; Fig. 3). Subsequent tectonic deformation along the northern Taranaki coastline has been limited and is restricted to (1) monoclinal regional tilting (generally <10° to the west or southwest; King and Thrasher, 1996), (2) gentle long-wavelength folding (Nodder, 1987; Utley, 1987), and (3) normal faulting with generally low (< 20 m) amounts of displacement (Childs et al., 2007, 2009; Giba et al., 2010).
Our study of the NAMTD includes a variety of approaches that include both outcrop- and subsurface-based investigations. (1) Sedimentological observations and measurements were collected in the field, including measurement of ∼245 m of stratigraphy from 18 sections at a centimeter to meter scale from within and adjacent to the NAMTD (measured sections are provided in Appendix DR1 in the Supplemental File1). Sections were strategically collected to document the stratigraphic context of the MTD, correlate key marker beds along the outcrop extent, and assist with interpreting the sedimentary facies and depositional setting of MTD stratigraphy. We also collected paleocurrent measurements from flute casts, ripples, and dunes (Fig. 2). (2) A structural data set of 134 slump fold and 242 slump fault orientations was generated for kinematic analysis of the MTD (Tables DR1 and DR2 in the Supplemental File [see footnote 1]). Although some fold orientations were measured directly in the field, most were calculated stereographically (e.g., Davis et al., 2012). When possible, one or more bedding measurements were collected from each limb and hinge of the fold, and the general orientation and character of the fold was recorded. The fold axis orientation is taken to be either the β or π axis (Davis et al., 2012), and the axial plane orientation is taken to be the plane that bisects the fold limbs (additional description of methods in Appendix DR2 in the Supplemental File [see footnote 1]). In addition to considering the NAMTD as a whole, we also subdivide the coastal transect into 7 subregions (R1–R7; Fig. 2) to test for systematic spatial variability in the estimated transport direction of the MTD. The number and locations of subregion boundaries were chosen to maximize north to south resolution while allowing for a suitable number of slump structure measurements within each subregion (Tables DR1 and DR2 in the Supplemental File [see footnote 1]). Subregions R1–R3 and R7 were further defined according to noted changes in the orientations of slump folds and faults within these portions of the NAMTD (see following discussion; Fig. 2). (3) A high-resolution photographic database was collected that includes complete coverage of the coastal cliffs (taken from a boat ∼300–500 m offshore) and wave-cut platforms (taken vertically downward from an airplane). Individual photos of the coastal cliffs were stitched together using Adobe Photoshop to form a nearly seamless photomosaic of the entire study area (Appendix DR3 in the Supplemental File [see footnote 1]). Georeferenced orthophotos were produced from the aerial photos of the wave-cut platforms using the photogrammetry software Agisoft PhotoScan Professional (Appendix DR4 in the Supplemental File [see footnote 1]). (4) Regional data were provided by open-source, industry-acquired 2D and 3D seismic-reflection surveys and petroleum wells located offshore of the study area (Fig. 1; see also King et al., 2011).
NORTH AWAKINO MTD
The NAMTD collectively refers to an exposed interval of extensive soft-sediment deformation related to submarine mass movement that extends for ∼11 km along the coastline between Pitone and Paparahia Streams (Henderson and Ongley, 1923; Utley, 1987; King et al., 2011; Fig. 2). The majority of the NAMTD can be classified as a submarine slump on the basis of bedding being largely preserved, even within zones that have been intensely folded (Martinsen, 1994). However, portions of the NAMTD are only slightly deformed, and thus this MTD also displays characteristics of a translational slide (sensu Posamentier and Martinsen, 2011). Exposures of the NAMTD are nearly continuous along the entire study area, but the coastal cliffs are considerably higher in the northern 6 km of the study area (to ∼100 m; Fig. 4) compared to those in the southern 5 km (typically <10 m). A well-developed wave-cut platform extends along nearly the entire coastline and is typically 150–200 m wide (Figs. 2 and 5).
The upper surface of the MTD is well exposed over most of its outcrop extent and can be traced over long distances (Fig. 4). Deformed strata within the MTD are often abruptly truncated at this contact (Fig. 4). Although this surface is quite flat, some vertical relief (typically to several meters) is present over broad distances. Such relief has locally been healed by deposition of overlying turbidite sand (e.g., north of Waihi Stream; Fig. 4E). Otherwise, the upper surface of the MTD is overlain by well-bedded, variably volcaniclastic strata (lowermost Mount Messenger Formation; Utley, 1987) that can be traced for large distances laterally.
The flatness of the NAMTD upper surface is somewhat enigmatic because MTD upper surfaces are often marked by significant rugosity following emplacement (Lucente and Pini, 2008; Bull et al., 2009; Armitage et al., 2009). However, similarly flat and apparently erosional upper surfaces have been described in other MTDs in both outcrop and in the subsurface (Woodcock, 1976; Farrell, 1984; Alsop and Marco, 2011; King et al., 2011). The upper surface of the NAMTD may have been modified by erosion following emplacement by either turbidity currents or possibly strong contour currents (King et al., 2011). Alternatively, the upper surface may have been levelled in a process described as relaxation by Alsop and Marco (2011). In this scenario, seafloor topography generated by submarine mass movement was subject to secondary translation of material from highs into adjacent troughs (Alsop and Marco, 2011).
MTD Thickness and Extent
The thickness of the MTD can only be constrained to a minimum estimate (locally ≥55 m) along much of its outcrop extent because the basal detachment surface is generally not exposed above beach level (Fig. 4). However, we have identified and mapped a local exposure of the basal detachment surface for ∼400 m along the coastline at Paritutu Stream where both the upper and lower bounding surfaces of the MTD are clearly visible in the wave-cut platform (Figs. 2 and 5D). This contact separates variably deformed strata from undeformed strata below but otherwise juxtaposes similarly thin-bedded facies of the Mohakatino Formation (Appendix DR1 in the Supplemental File [see footnote 1]). The calculated thickness of the MTD in this region is ∼40 m based on a measured bedding dip of ∼11° to the southwest (Fig. 5D).
The areal extent of the NAMTD is similarly unknown. Isolated local exposures of the MTD are present inland from the coast (Utley, 1987), but dense vegetation precludes detailed mapping, such that the eastern margin of the MTD is poorly constrained. The offshore region west of the MTD outcrop extent is covered by a 3D seismic-reflection survey (Kahu) that extends to within ∼3 km of the study area (Figs. 1 and 6). However, poor data quality within the eastern portion of the 3D volume precludes confident identification of mass transport–related seismic facies (Figs. 6A, 6B). Although data quality improves ∼12 km west of the coastline, the Kahu 3D volume lacks strong evidence for the presence of an MTD within the Mohakatino–Mount Messenger interval in this region (Fig. 6C). Thus, we infer that the NAMTD does not extend past ∼12 km west of the coastline or has a thickness that is below seismic resolution (Figs. 2 and 6). The theoretical vertical resolution of the Mohakatino Formation within the Kahu 3D survey is ∼15 m (25% of the dominant wavelength) based on an estimated seismic velocity of 2650 m/s (King et al., 2011).
Semicontinuous exposures of the Mohakatino Formation along the coastline allow the north-south extent of the MTD to be better constrained than its east-west dimensions (Fig. 2). Field relationships suggest that the southern margin of the MTD is likely located just north of Pitone Stream based on the observation that the MTD is absent in the vicinity of Pitone Stream where the unconformable basal contact of the Mohakatino Formation is exposed (Fig. 3). This interpretation implies that the MTD thickness decreases from ∼40 m to 0 m in <1 km. This interpretation is at odds with the hypothesis that the NAMTD is cogenetic with the Otukehu Composite MTD that is exposed ∼15 km to the south between the Mohakatino and Tongaporutu Rivers (King et al., 2011). Although these MTDs are similar in scale and deformational characteristics (King et al., 2011; Masalimova, 2013), we infer that these MTDs are not correlative on the basis of different emplacement timing. Specifically, distinctive foraminifera in strata immediately overlying the NAMTD yield an estimated depositional age of 10.9–10.5 Ma (Appendix DR5 in the Supplemental File [see footnote 1]). This age range is distinctly older than a volcanic ash bed sampled from within the Otukehu Composite MTD (9.66 ± 0.28 Ma; Maier, 2012). In addition, exposures of the Mohakatino Formation south of the Awakino River lack evidence for being remobilized, suggesting that the MTD does not extend into this region (Fig. 2).
The northern margin of the NAMTD is not exposed, but several lines of evidence suggest that the MTD does not extend much farther north than Paparahia Stream (Fig. 2). Coastal and inland exposures of the Mohakatino Formation are widespread north of Paparahia Stream and are only rarely deformed by submarine mass movement (Nodder, 1987). In particular, a thick (>80 m), fault-bounded sequence of undeformed Mohakatino Formation is present at Opito Point, just ∼2.5 km north of exposures of the NAMTD (Fig. 2). Because the estimated thickness of the Mohakatino Formation is only ∼120 m in this area (Nodder et al., 1990a), it is likely that the NAMTD does not extend as far north as Opito Point. Similarly, the NAMTD is not present at Waikawau Beach, where the Mohakatino–Mount Messenger contact is well exposed and is undeformed. These observations suggest that the NAMTD likely only extends to 2.5 km, and probably not past 6 km, from its northernmost outcrop location (Fig. 2). Taken together, these observations suggest that the outcrop exposures of the NAMTD probably include a minimum of 60%–85% of its original north-south extent (Fig. 2).
Stratigraphy and Sedimentology
We designate two general end-member lithofacies that make up the majority of the NAMTD and of surrounding exposures of undeformed Mohakatino Formation: (1) thick- to medium-bedded volcaniclastic sandstone (LF1) and (2) thin-bedded volcaniclastic sandstone and mudstone (LF2; Fig. 7). These lithofacies are defined based on associations of lithologies characterized by distinctive bedding thickness, grain size, sedimentary structures, and inferred processes of deposition (Ghosh and Lowe, 1993). There is significant spatial variation in the distribution of these two lithofacies within the NAMTD: LF1 predominates in the northern ∼5 km and gradually transitions to LF2, which becomes dominant south of Poporotaupo Stream (Fig. 4). This same general facies transition is observed within surrounding undeformed exposures of the Mohakatino Formation. LF1 is abundant in northern outcrops at Waikawau Beach and at Opito Point, whereas LF2 is more abundant to the south where exposed along the banks of the Awakino and Mokau Rivers (Fig. 2). The commonality of lithofacies types and proportions between the MTD and surrounding undeformed intervals strongly suggests that the NAMTD originated locally and underwent limited translation.
LF1: Thick- to medium-bedded volcaniclastic sandstone
Description. Thick- (>50 cm) to medium-bedded (∼20–50 cm) volcaniclastic sandstone (LF1; Figs. 7A, 7C) is the most abundant lithofacies in the northern 5 km of the NAMTD, between Ounutae and Paparahia Streams. Although LF1 is dominated by medium- to very coarse grained sand and gravel, interbedded thin beds of fine sandstone and mudstone are also present. Coarse-grained volcaniclastic sandstone beds occur as amalgamated packages as much as 10 m thick (Fig. 7C). Normal grading (including coarse-tail grading) and massive bedding are typical, and common sedimentary structures include water escape structures (sheets, pillars, and dish structures); sole marks (flutes); and planar, ripple, and convoluted laminations. Intraformational clasts of sand and mud (to ∼1 m in diameter) are abundant in the thick-bedded sandstone and often form intraclast and lithic fragment breccio-conglomerates.
LF1 is characterized by a high degree of lateral bedding continuity. Individual beds can be correlated between outcrops within the northern 3 km of the MTD and demonstrate little change in bedding thickness over this distance despite significant structural shortening. This same lithofacies can be observed undeformed several kilometers north of the MTD at Opito Point, where individual beds can be traced laterally for >2 km.
LF1 also contains rare beds of nonvolcaniclastic or mixed-volcaniclastic composition that appear to have a Mount Messenger Formation affinity (Utley, 1987). These siltstone and sandstone beds are conspicuous because they weather to gray instead of the brown typical of the surrounding volcaniclastic units. Because these beds are relatively uncommon, they constitute useful marker beds that aid correlation along the coastline (e.g., green highlighted marker bed; Fig. 4). It is interesting that these isolated beds are only found within LF1 north of Poporotaupo Stream, but are also present within undeformed Mohakatino strata at Waikawau beach, ∼6 km north of the NAMTD.
An ∼22-m-thick package of intensely deformed, broken, and chaotic medium- to thick-bedded sandstone is present in the upper portions of the NAMTD north of Kopuai Stream (Fig. 4B). A detachment surface separates this unit from underlying beds that are deformed by longer wavelength folding (Fig. 4B). Although the precise timing between emplacement of the intensely deformed package and the NAMTD is ambiguous, it is incorporated into and crosscut by larger folds within the NAMTD (Figs. 4B and 5A). We interpret this package to have been a precursor MTD within the original stratigraphy that was subsequently remobilized within the NAMTD.
Interpretation. Thick-bedded strata of LF1 are interpreted to have been deposited primarily by high-density turbidity currents, forming S2 and S3 units of Lowe (1982) and Tbc units of Bouma (1962). Medium-bedded strata are interpreted to have been deposited both by high- and low-density turbidity currents, characterized by Ta, Tb, and Tc divisions (Bouma, 1962).
LF2: Thinly bedded volcaniclastic sandstone and mudstone
Description. Thin-bedded (≤20 cm) volcaniclastic sandstone and mudstone, LF2, become abundant south of Waioroko Stream; LF2 is the dominant lithofacies south of Poporotaupo Stream (Fig. 7B). This lithofacies is very well bedded and has a notably bimodal grain-size distribution; medium sand to granule bases are capped by fine sand to mud tops (Fig. 7D). Sedimentary structures include normal grading, massive texture, planar and convoluted laminations, ripple laminations, and sole marks (flame and loading structures). LF2 is also characterized by a high degree of bedding continuity; packages of thin-bedded sandstone can be traced for >2 km at Opito Point. Bioturbation has destroyed primary sedimentary structures in many cases, often resulting in mixing of medium to very coarse volcaniclastic sand within the fine-grained matrix. This lithofacies is only rarely observed to have undergone mass wasting prior to the NAMTD event, and lacks the interbedded nonvolcaniclastic beds found within LF1.
Interpretation. The thin-bedded sandstone and mudstone of LF2 are interpreted to have been primarily deposited by one of two mechanisms. Beds that have current structures and display Bouma (1962) Tb and Tc divisions indicate deposition from waning, low-density turbidity currents. Other beds, characterized by normal grading and otherwise massive texture, indicate suspension sedimentation of volcanic ash through the water column.
Interpreted Depositional Environment
Several lines of evidence suggest that both LF1 and LF2 were deposited as sheets or lobes within a low-gradient, laterally unconfined deep-marine environment: (1) the abundance of massive sandstone beds with dish structures (S3 divisions; Lowe, 1982) suggesting collapsing flows and rapid sedimentation rates due to a loss of confinement and/or decrease in slope; (2) the high degree of lateral bedding continuity (Johnson et al., 2001); and (3) the absence of incisional or large-scale erosional features associated with channelized confinement.
Although the channels that fed the Mohakatino lobes were not observed in outcrop, we suspect that these conduits were located proximally to LF1. Abundant intraformational rip-up clasts within LF1 suggest that considerable erosion occurred upstream prior to deposition. Also, the coarse-grained, thick-bedded, and amalgamated nature of LF1 is consistent with deposition within the proximal (inner) or axial portions of a submarine lobe (Figs. 7A, 7C; Prélat et al., 2009). The LF2 lithofacies was likely deposited in a lobe off-axis to lobe fringe setting (Prélat et al., 2009), based on its varying character and fabric relative to LF1, i.e., (1) a decrease in grain size and bedding thickness, (2) an increase in bioturbation, and (3) more abundant preservation of volcanic ash horizons that may have been deposited by suspension sedimentation through the water column. Thus LF2 likely was deposited in a more quiescent environment relative to LF1, although LF2 sedimentation was periodically interrupted by deposition of medium-bedded, current-structured sandstone (e.g., Fig. 7F).
Measurements of paleocurrent indicators from undeformed strata within the Mohakatino Formation demonstrate a wide variety of flow directions (Fig. 2). Paleocurrent measurements north of the NAMTD suggest flow predominantly toward the southwest, south, or southeast (Fig. 2). Populations of north-flowing ripples were also measured from Opito Point (Fig. 2). West- to southwest-directed flow is indicated by measurement of ripples and dunes within the basal Mohakatino Formation along the north bank of Pitone Stream and the south bank of the Awakino River (Fig. 2). Paleocurrent measurements from current ripples immediately above the MTD include northeast-directed flow in a broad, shallow (∼3 m) channel form at Waiakapua Stream and south-directed flow at Paparahia Stream (Fig. 2).
Other than two isolated measurements of flute casts south of Ounutae Stream that suggest northwest-directed flow, all measured flute casts within the northern NAMTD (subregions R1–R3) indicate flow toward the eastern quadrant. These measurements were collected from the bases of two distinctive thick sand beds where they are repeatedly exposed at beach level several times over a distance of 2 km within subregions R1–R3 (Fig. 2). The mean paleocurrent vector changes from 54° ± 7° to 121° ± 10° from north to south within these measurements, suggesting an apparent flow direction that varies as much as ∼67° (Fig. 2). This pattern could result from spatial variability in the current flow direction while the sand bed was being deposited. Alternatively, it can be assumed that the flute casts were originally parallel, and that those to the north were subsequently rotated in a counterclockwise direction relative to those collected near Waioroko Stream (Fig. 2).
Slump Deformation Structures
Slump folds form some of the most conspicuous and exquisite deformational features present within the NAMTD (e.g., Figs. 4 and 8). These features exhibit a considerable variety of shapes, sizes, and orientations that vary systematically along the outcrop transect (Fig. 9). The largest folds have heights of as much as several tens of meters and are most commonly found in the north (subregions R1–R4) where LF1 predominates (e.g., Figs. 8A, 8B, and 9). These folds tend to occur within semicontinuous fold trains (Figs. 5A, 5B, and 10), and individual folds locally extend as much as ∼1 km laterally (e.g., Figs. 5A, 5B). Folds in the south (subregions R5–R7) are typically disharmonic, tend to occur as rootless individuals (e.g., Fig. 8C), and have an average fold height that is less than that found in subregions R1–R4. Continuous fold trains are uncommon except where present locally in subregion R7 (Fig. 11).
Fold axes within the NAMTD are typically shallowly plunging (<25°) and display a wide range of axis orientations (Figs. 12 and 13). The best-fit girdle to fold axes correspondingly dips shallowly to the west (192°/12°W; Fig. 13). Fold axes have been assumed to be dispersed within the plane of the MTD (Woodcock, 1979b) and thus can be used to infer the overall orientation of the MTD. However, we suspect that the west-dipping girdle is likely an artifact of the slight (typically ≤10°) westward regional tilting of the Taranaki coastal exposures, such that the original slump sheet orientation was nearly horizontal following emplacement. A low amount (generally <20°) of fold axis curvature has been observed in some large folds (e.g., Fig. 8A). Sheath folds (fold axis curvature >90°; Ramsay and Huber, 1987) have not been identified within the NAMTD.
Fold attitudes range from upright to nearly recumbent and interlimb angles are typically <60° or >100° (Fig. 12). Although many fold axial surfaces are approximately planar at the scale of outcrop, some display considerable amounts of curvature (e.g., Fig. 14). For example, the upward-convex geometry observed in some folds (e.g., Figs. 14A, 14B) may reflect modification during layer-parallel simple shear (e.g., Farrell and Eaton, 1987, fig. 3A therein). Similar structures to the large folds shown in Figure 14B have been experimentally produced by deformation of layered sand and clay during simple shear (Dasgupta, 2008).
Folds demonstrate variable thickening or thinning of limbs relative to the hinge, and fold layer shapes vary from class 1A to class 2 as measured in profile view (Ramsay, 1967). Calculation of dip isogons for several folds of varying scales demonstrates that most folds exhibit a shape (class 1C) that is intermediate between parallel and similar ideals (Ramsay, 1967).
Both flexural-slip and flexural-flow folding (Donath and Parker, 1964) occur within many folds; interbedded fine-grained lithologies tend to preferentially flow relative to the coarser grained lithologies (e.g., Figs. 14A, 14B). However, even thick sandstone beds exhibit thickening in the hinge region (class 1C folds), and thus have been subjected to a certain amount of flexural flow during folding. Fold detachment is repeatedly observed where parallel folding produces a space problem within the hinge zones (e.g., Figs. 14A, 14B, and 15). Flow of finer grained lithologies into the fold hinge creates a décollement zone that often contains sandstone boudins within a pervasively sheared fine-grained matrix (Fig. 15; Davis et al., 2012). Within some isoclinal folds found in subregions R1–R3, the décollement zone can extend for more than 50 m in profile view, thereby separating hinges of formerly adjacent sandstone beds (Fig. 14A). Similar structures (X folds) have been described from the Pleistocene Dead Sea Basin and are inferred to have formed in response to extreme ductility contrast within layered sequences (Alsop and Marco, 2013). One well-exposed series of recumbent, nearly isoclinal folds in a cliff face (subregion R3) demonstrates that fold detachment can result in the same sandstone beds being structurally repeated as many as three times (Figs. 14C and 15).
Deformation intensity can be qualitatively ranked based on the degree of fold coherency (fold trains versus isolated, rootless folds), fold hinge tightening, axial plane rotation (recumbent versus upright), and amount of fold hinge curvature (Strachan, 2002). Using these metrics, the intensity of deformation varies significantly within the NAMTD, and in general is greatest within the central portion of the outcrop extent (subregion R6). However, intensely deformed, or high strain, zones are often observed to transition laterally or vertically into low-strain domains that are characterized by horizontally to shallowly dipping bedding, occasional low-angle discontinuities, and relatively few open to tight folds. These regions can be clearly identified on the wave-cut platforms as having continuous bedding that is tilted to the west and only disrupted by upright, long-wavelength folding (Fig. 5). Low-strain zones are particularly well developed in three locations: north of Kopuai Stream, between Poporotaupo and Waiongaro Streams, and between Paritutu and Ngatupaku Streams (Figs. 2 and 5). Upward-increasing strain is also observed in several locations where folds become tighter, more disharmonic, and increasingly dismembered toward the upper surface of the MTD. For example, folds exhibit an increasing level of deformation vertically from the basal detachment in the vicinity of Paparahai Stream (Fig. 5D). Other regions have highly deformed intervals that are intercalated with low-strain domains (e.g., Figs. 5C, 5D).
Two general groups of faults can be identified within the NAMTD: those that accommodated brittle deformation during submarine mass movement (both normal and reverse slump faults) and those that postdate the MTD and are of a tectonic origin (exclusively normal faults; Childs et al., 2007; Giba et al., 2010, 2013). The younger group of tectonic normal faults generally strike northeast-southwest or east-west (Utley, 1987; Nodder et al., 1990a; Giba et al., 2010), and can be identified in outcrop because they extend through the MTD and crosscut the overlying succession (e.g., Fig. 4). Slump normal faults can be distinguished from the tectonic faults based on several defining characteristics: (1) they are confined to bedding within the MTD; (2) they often have welded surfaces that lack mineralization or fractures (e.g., Fig. 16A); (3) they have surfaces that are often irregular and curved (Figs. 16B, 16C); and (4) they usually preferentially displace coarser grained units and tip out within fine-grained intervals (Debacker et al., 2009). The following discussion will focus exclusively on slump faults.
Slump faults are abundant in the NAMTD and demonstrate a wide range of orientations and styles of deformation (Figs. 16 and 17). Planar discontinuities commonly separate rootless folds in regions that have undergone significant amounts of deformation (e.g., Dasgupta, 2008). Similar discontinuities form boundaries to coherent blocks of stratigraphy and commonly have low dip angles. These features may represent low-angle detachment or thrust faults that have likely been modified by progressive deformation. Because the displacement magnitude is typically greater than the scale of the outcrop, the relative sense of displacement cannot usually be determined. Other faults are preferentially located in fold hinge zones, suggesting that these faults formed during folding (e.g., Fig. 16E).
We measured the orientation and relative displacement sense of 242 slump faults that were observed to displace bedding and can be identified as having a relative normal or reverse sense of motion (Figs. 17 and 18). We avoided measuring faults that appeared to have formed locally as a result of folding (e.g., Fig. 16E). Normal faults typically dip between 30° and 70°, and reverse faults dip between 5° and 35° (Fig. 17). Although fault displacements can be as great as several meters or more, the large majority of measured slump faults within the NAMTD have apparent displacement magnitudes that are <∼20 cm (Fig. 17). Thus, the scale of brittle deformation is typically as much as several orders of magnitude less than the scale of ductile strain accommodated in folds. Indicators of fault slip direction (e.g., slickenlines) were not observed, and thus fault displacement magnitude and sense can only be approximately constrained.
The orientations of slump structures have long been considered to be related to the orientation of the underlying bathymetric slope (e.g., Hahn, 1913; Jones, 1939). For example, early workers noted that some slump fold axes tend to align in an along-slope direction in a pattern analogous to a “corrugated carpet” (Woodcock, 1979b, p. 83). However, slump structures can also be variably orientated with respect to the underlying paleoslope (Hansen, 1971; Woodcock, 1979b; Farrell and Eaton, 1987), posing challenges to paleoslope analysis. For example, slump folds axes that initially formed parallel to the strike of slope (herein along-slope folds) can develop oblique or downslope orientations during progressive simple shear (e.g., Farrell and Eaton, 1987). Oblique or downslope folds can also be generated in situ as a result of layer-normal shear (Alsop and Marco, 2013) or as a result of shortening across the width of the MTD.
Our approach is to conduct a rigorous paleoslope analysis by using as many fold- and fault-based methods as applicable and then comparing the results of each method (see discussions in Strachan and Alsop, 2006; Debacker et al., 2009; Alsop and Marco, 2012). We also test for spatial variability in the NAMTD by estimating the MTD transport direction separately within seven divided subregions and for the MTD as a whole (Fig. 2). In the following we provide an explanation of each fold- and fault-based paleoslope method used in this study (Table 1).
Methods Based on Slump Fold Orientations
Mean Axis, Separation Arc, and Downslope Average Axis Methods
The mean axis method (MAM) postulates that the mean fold axis orientation approximates the paleoslope strike (Jones, 1939; Fig. 18; Table 1). The MAM assumes that fold hinges have undergone little downslope rotation and that the deforming MTD underwent little strike-parallel length changes (Woodcock, 1979b). This method yields two possible MTD transport directions that are 180° apart (Fig. 19). The paleoslope dip direction is usually chosen based on regional paleogeographic constraints, fold asymmetry (folds are assumed to verge in the downslope direction), and/or fold facing directions (beds become younger in a downslope direction within fold hinges; Woodcock, 1976; Bell, 1981).
A major disadvantage to the MAM is that it does not allow for fold axes that are oriented parallel or oblique to the downslope direction, even though such distributions are thought to occur in slump sheets (Hansen, 1971; Lajoie, 1972). In these cases, the separation arc method (SAM; Hansen, 1965) and/or downslope average axis method (DAM; Woodcock, 1979b) are more suitable, as the MAM can yield an estimate of the paleoslope orientation that is as much as 90° different from the true value (Woodcock, 1979b).
Fold axes within the NAMTD are markedly dispersed and show little preferred orientation (Fig. 13), suggesting that folds are variably oriented with respect to the underlying paleoslope. Furthermore, it is ambiguous whether the mean fold axis (279°/38°) is oriented parallel, perpendicular, or possibly obliquely to the bathymetric slope. Although the MAM predicts a north- or south-directed transport direction (9° or 189°), the wide distribution of fold axes casts doubt on whether a reliable fold axis mean can be calculated (Woodcock, 1979b). Similarly, the complete overlap between Z and S folds precludes the use of the SAM or DAM in determining the paleoslope orientation (Figs. 13 and 19; Table 1).
Even though interpreting fold axes within the entire NAMTD is ambiguous, 6 of the 7 subregions (R1–R5 and R7) have interpretable fold axis distributions that yield markedly different mean values from each other (Fig. 13; Table 2). Subregions R1–R3 have only 5 or 6 fold measurements each, and thus caution should be used when interpreting the statistical significance of the fold-based results in these areas. However, any statistical limitations of these small sample sets are partially offset by the large size of these folds (e.g., Figs. 4A–4C) and the consistency in their orientations that allows the calculation of a meaningful mean fold axis. Application of the MAM suggests either north- or south-directed transport for subregions R1 and R7 (14°–194° and 179°–359°, respectively; Fig. 13). These results differ from the east-west or southeast-northwest transport directions inferred for subregions R2–R5 (from 72°–252° to 148°–328°; Fig. 13). Subregions R1–R2 display a dominant vergence direction that suggests south- and southwest-directed transport, respectively (Fig. 13). However, the other subregions (R3–R7) are characterized by bimodal fold vergence distributions that only allow the MTD transport direction to be constrained to one of two possibilities (Fig. 13). The SAM and DAM cannot be applied because either only S folds were measured (subregions R1–R2) or Z and S folds do not occupy distinct fields (subregions R3–R6; Fig. 13). Subregion R7 displays only partial overlap of opposing vergence senses; most S folds occur north of the 89°–269° mean vector (7 of 10) and most Z folds occur south of the mean vector (5 of 8). Although a separation angle cannot be defined, the DAM indicates an east-west transport direction for subregion R7 (Fig. 13).
Axial Planar and Axial Planar Intersection Methods
Fold axial surfaces (approximated as axial planes) yield important kinematic information in a similar way as fold axes (Woodcock, 1976, 1979b; Farrell and Eaton, 1987). Axial plane orientations have the added benefit of including the sense of fold vergence. Poles to axial planes tend to fan in a great circle about the mean fold axis, and thus the strike of the best-fit girdle to these poles gives a bidirectional estimate of the MTD transport direction (axial planar method, APM; Woodcock, 1979b; Fig. 19; Table 1).
A similar method assumes that the MTD transport direction is related to the intersection of the mean S and Z fold axial planes (axial planar intersection method, APIM; Alsop and Holdsworth, 2002; Strachan and Alsop, 2006; Fig. 19; Table 1). This method was initially developed on folds from mid-crustal shear zones (Alsop and Holdsworth, 2002), but has also been successfully applied to paleoslope analysis of ancient MTDs (Strachan and Alsop, 2006) and recent slump deposits (Alsop and Marco, 2012). The MTD transport direction is inferred to be the trend of the S and Z fold mean plane intersection (designated APIM-N) for regions characterized by layer-normal shear (e.g., wrench strains at the margin of a submarine landslide; Strachan and Alsop, 2006). However, the transport direction is orthogonal to the intersection trend for regions characterized by layer-parallel shear (designated APIM-P; Strachan and Alsop, 2006). Thus, the APIM yields three possible transport directions. We have considered the fourth direction for axial planar intersections with low plunges (<10°) because the point of intersection is within error of shifting 180° on an equal-area stereonet (e.g., Fig. 13).
As with fold axes, axial planes within the NAMTD are widely dispersed and show little preferred orientation (Fig. 13). The best-fit girdle of the poles is oriented 349°/64°E, and thus the APM predicts an MTD transport direction parallel to 169°–349° (Fig. 13; Table 2). Application of the APIM to all axial planes in the NAMTD yields an estimated transport direction of 10°–190° (APIM-P) or toward 280° (APIM-N; Table 2). However, these results should be viewed with caution, because the best-fit girdle is poorly constrained in such scattered distributions as a result of being sensitive to outlying points (Fig. 13).
The APM can be applied to each subregion within the MTD and yields results similar to the MAM (Fig. 13). Subregions R1 and R6–R7 yield north or south transport directions (Fig. 13). Subregions R2–R5 display a variety of predicted transport directions that range from northeast-southwest (53°–233°) to southeast-northwest (158°–338°; Fig. 13; Table 2). The mean axial plane variably dips to the west or east in these regions. The APIM can be applied to 4 of 7 subregions (R3–R4 and R6–R7) and yields results (within 25°) similar to the APM, with the exception of subregion R6, which is 75° different from the APM (Fig. 13; Table 2).
Methods Based on Modification of Fold Geometry during Progressive Simple Shear
Farrell and Eaton (1987) proposed a model in which layer-parallel simple shear modifies fold geometries during progressive deformation during submarine mass movement; folds are predicted to progressively tighten, become recumbent, and rotate into a downslope direction. As a result, the relationship between fold hinge azimuth and fold interlimb angle (HIM; Strachan and Alsop, 2006) or axial plane dip (HAM; Farrell and Eaton, 1987) can be used to identify downslope hinge rotation and determine the direction of MTD transport (Fig. 19; Table 1). In both cases, the downslope direction can be identified on a plot of interlimb angle or axial plane dip versus fold axis azimuth as a V shape with the apex coinciding with the direction of MTD transport (Fig. 19; Farrell and Eaton, 1987; Strachan and Alsop, 2006). Correspondingly, folds that have undergone little hinge rotation will not display a systematic relationship between fold axis azimuth and interlimb angle or axial plane dip (Fig. 19). Debacker et al. (2009) proposed a similar method that uses the same relationship between axial plane orientation and fold interlimb angle to infer downslope fold rotation with progressive fold tightening (axial surface strike and interlimb angle method; ASIM; Table 1).
The HIM, HAM, and ASIM cannot be confidently applied to the entire NAMTD or any subregion within because there is no clear relationship between fold interlimb angle and the attitude of fold axes or axial planes (Fig. 20). These observations suggest that folds underwent limited systematic downslope hinge rotation (Strachan and Alsop, 2006). A single possible exception may be subregion R5, the fold axes and axial planes of which display a poorly developed V pattern when plotted against interlimb angles with the apex oriented toward ∼148° (Fig. 20). Subregions R1–R4 and R7 have fold orientations that plot in one of two groups that each have a range of interlimb angles (Fig. 20). Application of the HIM, HAM, and ASIM in these regions yields results that are functionally equivalent to the MAM and APM. The wide scatter of fold attitudes within subregion R6 precludes the application of HIM, HAM, or ASIM in this area (Figs. 13 and 20).
Axial Surface Dip and Dip Direction Method
Debacker et al. (2009) proposed that the relationship between the dip magnitude and dip direction of fold axial surfaces is related to the transport direction in the lateral to oblique portions of MTDs (axial surface dip and dip direction method; DDM; Fig. 19; Table 1). This method assumes that several conditions are present in the lateral and/or oblique portions of MTDs: (1) layer-normal shear predominates, (2) fold axial surfaces are steeper than in the central and/or frontal portions of the MTD, and (3) fold axes and axial planes fan about the paleoslope dip direction (Debacker et al., 2009; Alsop and Marco, 2013). Axial surfaces are presumed to steepen with progressive layer-normal shear, and this process is manifested as a negative relationship between interlimb angle and axial surface dip (Debacker et al., 2009, fig. 15C therein). However, Alsop and Holdsworth (2007) predicted the opposite relationship to occur within differential shear zones.
A major challenge to the application of the DDM is the requirement of prerequisite knowledge about relative position within the MTD (Table 1). This method can only be confidently applied when independent evidence indicates a position within the lateral and/or oblique portion of the MTD. The margins of the NAMTD are not exposed, but geologic relationships suggest that subregions R1 and R7 may be within the lateral portions the MTD (Fig. 2). Application of the DDM to subregions R1 and R7 yields east- or west-directed transport (Fig. 13). However, several considerations suggest that subregions R1 and R7 do not strictly meet the assumptions of the DDM (Table 1). (1) Plots of interlimb angle versus axial plane dip show a positive correlation, suggesting that slump folds become more reclined during progressive fold tightening (Fig. 20). This pattern is typical of slump folds that have been deformed by layer-parallel simple shear (e.g., Farrell and Eaton, 1987), but differs from the negative relationship predicted to occur at MTD margins (Debacker et al., 2009). (2) Although fold axial surfaces are predicted to be steeper along the MTD margins than in the center (Debacker et al., 2009; Table 1), this relationship is not well developed along the inferred margins of the NAMTD (Fig. 20). The mean axial plane dip is slightly greater for subregion R7 than other subregions (mean 71°), but this is not so for subregion R1 (mean 48°; Fig. 20). (3) There is no clear systematic relationship between the orientation and dip of fold axial planes (Fig. 20). Subregions R1 and R7 display little (≤10°) difference between fold axial surface attitudes above and below the mean dip (Fig. 20). Taken together, these observations suggest that the DDM is not likely an applicable method for determining the transport direction of the NAMTD.
Methods Based upon Slump Fault Orientations
Mean Fault Orientation Method
The mean fault orientation method (MFOM) simply presumes that the MTD transport direction is subparallel to the mean fault dip direction (updip for reverse faults and downdip for normal faults; Farrell, 1984; Martinsen and Bakken, 1990; Debacker et al., 2009). This method assumes that there is not a significant component of oblique- or strike-slip fault motion (Debacker et al., 2009). Unfortunately, the degree of oblique- or strike-slip fault movement is difficult to assess because slip direction indicators (e.g., slicken-lines) were not observed on slump fault surfaces.
Reverse faults within the NAMTD display little preferred orientation when combined together (Fig. 18). The mean reverse fault is nearly flat-lying (320°/1°NE), and thus the MFOM cannot be reliably applied to the NAMTD as a whole (Fig. 18). Normal faults similarly display a wide range of orientations but tend to be clustered into two main groups that are dipping to the east and north. As a result, the mean fault surface dips toward the northeast (332°/19°NE). Smaller populations of south- and west-dipping faults are also present (29% of total). Application of the MFOM to all normal faults within the NAMTD suggests a transport direction toward the northeast (43°; Fig. 18; Table 2).
As with slump folds, reverse fault orientations vary within different subregions of the NAMTD (Fig. 18). In general, poles to reverse fault planes form a steeply dipping girdle with poles clustering in opposing quadrants (Fig. 18). This pattern reflects the presence of conjugate sets of reverse faults that have subparallel strike orientations but opposing dip directions. As a result, the MFOM can only constrain the MTD transport direction to one of two possibilities in regions where there is not a dominant vergence orientation. This is the case for all subregions except R4, where most (9 of 11) measurements are dipping to the east. Estimated transport directions range from northeast-southwest to southeast-northwest for R1–R5 (from 36°–216° to 116°–296°) and north-south for R6–R7 (169°–349° and 117°–357°, respectively; Fig. 18; Table 2).
Normal faults display comparatively less systematic variation in orientation from north to south than reverse faults (Fig. 18). Most subregions have major populations of east- and north-dipping faults with subordinate populations of south- and/or west-dipping faults. As a result, the best-fit girdles dip to the west, southwest, or south (Fig. 18). Application of the MFOM to normal faults is somewhat less reliable due to the complex nature of fault orientation distributions, and low numbers of measurements in the north preclude robust analysis of normal faults in these areas (i.e., subregions R1–R2). Even so, the MFOM yields consistent northeast- and east-directed MTD transport directions (36°–102°) for subregions R3–R7, assuming that the most abundant populations of normal faults are synthetic with respect to the paleoslope (Fig. 18; Table 2). Alternatively, southwest- and west-directed MTD transport (126°–192°) is suggested if the dominant fault population is assumed to be antithetic with respect to the underlying paleoslope (e.g., Alsop and Marco, 2014). These results are generally inconsistent with the results from the fold- and reverse fault-based methods, with the possible exception of subregion R4 (Figs. 13 and 18).
Best-Fit Girdle to Fault Poles Method
A similar approach to the MFOM is to approximate the MTD transport direction as the strike of the best-fit girdle to fault plane poles (GFPM; Fig. 21; Table 1). This method has previously been applied to slump faults that were interpreted to have been oriented oblique to the paleoslope strike (Debacker et al., 2009). However, this method is equally applicable to faults that strike at low angles to the paleoslope strike (along-slope faults). We present a modified version of the GFPM that is applicable to both normal and reverse faults, allows for both along-slope faults and faults with orientations that fan obliquely about the paleoslope strike, and can be applied when both synthetic and antithetic faults are present (Fig. 21).
The estimated MTD transport direction is taken to be in the strike direction of the best-fit girdle for faults with inferred along-slope orientations (GFPM-P; Figs. 21A, 21B). Such faults can be recognized by poles that are ideally dispersed along a steeply dipping girdle (Figs. 21A, 21B). Synthetic and antithetic faults will be positioned on opposite sides of the stereonet center point. The MTD transport direction is taken to be in the downdip or updip direction for the synthetic population of normal and reverse faults, respectively (Fig. 21A), and can only be constrained to one of two possibilities if there is ambiguity about which fault population is synthetic versus antithetic. Note that the GFPM should be applied with caution for tightly clustered (unimodal) data sets (e.g., Fig. 21A) because the best-fit girdle is imprecise for such distributions; the MFOM yields a more reliable result in this case.
For faults with orientations that fan about and are oblique to the paleoslope strike, the MTD transport direction is taken to be parallel to the dip direction of the best-fit girdle (GFPM-N; Figs. 21C, 21D). Faults with obliquely fanning axes can be recognized by poles that are ideally aligned along a moderately dipping girdle (Figs. 21C, 21D). The sense of displacement is taken to be the dip direction for normal faults and the opposite of the dip direction for reverse faults (Debacker et al., 2009). Application of the GFPM-N becomes more complicated if both synthetic and antithetic faults are present (Fig. 21D). Ideally, separate best-fit girdles can be drawn if groups of synthetic and antithetic fault poles can be identified, and these girdles will have similar strike directions but opposite dips (Fig. 21D). The MTD transport direction is constrained only to one of two possibilities if uncertainty exists regarding which set is synthetic versus antithetic (Fig. 21D).
Use of the GFPM requires evaluation of normal and reverse fault pole distributions to determine (1) whether faults have along-slope or obliquely fanning orientations and (2) if both synthetic and antithetic populations are present (Fig. 21). Normal fault distributions within the NAMTD and most subregions (excluding R4) have moderately dipping (38°–78°) girdles and a dominant set of east- and north-dipping faults (Fig. 18). Most subregions also contain a second set of south- and west-dipping faults (Fig. 18). These two populations can be interpreted to represent two conjugate sets of fanning normal faults (e.g., Fig. 21D). If correct, the GFPM-N is applicable, and this method yields consistent southwest or northeast (40°–54° or 220°–234°) transport directions for the entire NAMTD and for subregions R3 and R5–R7 (Fig. 18). If the most abundant conjugate set (east and north dipping) is taken to be synthetic to the paleoslope, the inferred transport direction is toward the northeast. Normal fault poles in subregion R4 are aligned along a steeply dipping girdle, and we thus infer that these normal faults have primarily along-slope orientations (e.g., Fig. 21B). For this reason, we use the GFPM-P that yields an inferred transport direction toward the east (110°; Fig. 18), assuming that the dominant east-dipping fault population is synthetic to the paleoslope.
Reverse faults within most subregions of the NAMTD have poles aligned along steeply dipping girdles (82°–89°; Fig. 18), and thus appear to have along-slope orientations (Figs. 21A, 21B). Application of the GFPM-P yields results very similar to the MFOM, and uncertainty about which population of faults is synthetic versus antithetic in subregions R1–R3 and R6–R7 limits the interpreted transport direction to one of two possibilities (Fig. 18). Fault poles in subregion R5 display a comparatively wide spread and are not clearly aligned along a girdle (Fig. 18); thus the GFPM-P yields an ambiguous result in this case.
Mean Bedding Strike Method
In a way similar to that of folds, bedding attitudes within an MTD are expected to be preferentially oriented with respect to the paleoslope (e.g., Jones, 1939). Although individual bedding orientations will vary widely within any given MTD, we propose that the MTD transport direction can be related to the mean strike of bedding (mean bedding strike method; MBSM). This method is analogous to the MAM because the transport direction is taken to be parallel to the mean strike for MTDs with along-slope fold axes. Ideally, the mean bedding strike will also be subparallel with the mean fold axis and axial plane strike.
The MBSM can be applied by measuring a sufficient number of randomly selected bedding attitudes within an MTD and calculating the mean bedding orientation. We use a modification of this approach that utilizes the well-exposed wave-cut platforms along the study area (Fig. 5). Because the wave-cut platform surface is nearly horizontal, bedding traces are close approximates for bedding strike. Bedding traces are outlined on georeferenced aerial photos of the wave-cut platforms (Fig. 5), and strike is plotted on a bidirectional, length-weighted rose diagram from which the mean vector and associated uncertainty are calculated (Fig. 13). In addition, we measured the orientation of the axial surface traces for large folds visible on the wave-cut platforms (Fig. 5). In general, the orientations of fold axial surfaces show a close correspondence with the mean bedding strike, with the exception of subregion R6 (Table 2). This approach has the advantage of including a large number of measurements without having to collect numerous individual bedding attitudes in the field.
A disadvantage to our approach is that any subsequent tilting of the MTD will change the orientation of bedding planes and cause flat-lying beds to strike in one direction on a wave-cut platform. For example, the NAMTD has been gently tilted to the west (≤10°), and originally flat-lying beds now strike in a north-south direction (Fig. 5). Bedding orientations cannot be easily corrected for postemplacement tilting or folding because bedding dip magnitudes are generally not known with our approach. Because the NAMTD has only been tilted slightly (≤10°), we avoid this potential bias by excluding regions of the wave-cut platform that are characterized by shallowly dipping strata (<∼20°). In general, the MBSM should be used with caution in areas that have undergone significant tectonic deformation (folding or faulting), and may also be less precise for gently plunging, curvilinear folds.
Bedding strike orientations from the entire NAMTD display a wide range of attitudes but have a dominant north-south trend when plotted on a length-weighted rose diagram (Fig. 13). Correspondingly, the MBSM yields an interpreted east-west transport direction for the entire NAMTD (81°–261°; Fig. 13; Table 2). However, as with fold and reverse fault orientations, bedding strike distributions change systematically between subregions (Fig. 13). Bedding in the far north (R1) and south (R7) is east-west striking (transport toward 4°–14°or 184°–194°), and bedding in the central portion of the study area (R2–R6) ranges from northeast-southwest to southeast-northwest striking (transport toward 62°–112° or 242°–292°; Fig. 13). The MBSM shows close agreement with the results from fold-based methods in most subregions, and has a comparatively smaller margin of error (Table 2). In addition, the mean bedding strike closely agrees with the mean fold axial surface orientation for subregions excepting R6 (Table 2).
Comparison between Paleoslope Methods
Rather than relying on one or two methods to infer the direction of MTD translation, we conduct a rigorous comparison among the results of all applicable paleoslope methods for the entire NAMTD and for each of the seven subregions (Fig. 22). This approach has the advantage of allowing a more realistic assessment of uncertainty about the inferred paleoslope orientation based on the overlap between multiple, independent paleoslope estimations (e.g., Debacker et al., 2009). For example, the overlap of fold- and reverse fault-based results in subregions R1–R4 and R7 provides increased confidence to the paleoslope interpretation in these regions (Fig. 22). Regions with nonoverlapping results (e.g., subregion R6) should be interpreted with caution (Fig. 22). Several important conclusions can be drawn from comparing the results of each paleoslope method applied to the NAMTD.
Our paleoslope analysis is inconclusive when applied to the NAMTD as a whole due to nonoverlapping results of the few methods that are applicable (Fig. 22).
Although analysis of the MTD as a whole yields an ambiguous result, subregions R1–R4 and R7 have overlapping results that allow a defensible paleoslope interpretation to be made in each of these regions (Fig. 22). In particular, methods based on folds, reverse faults, and bedding all show good agreement. However, the interpreted transport direction varies by as much as 90° between these subregions (e.g., R1 versus R3; Fig. 22). This change can be readily observed in the wave-cut platforms in the northern portion of the NAMTD where east-west–trending folds transition rapidly into northeast-southwest– and finally northwest-southeast–trending folds (Figs. 5A, 5B).
The transport direction can only be constrained to one of two possible directions that are 180° apart in subregions R3–R5 and R7 (Fig. 22). This ambiguity is primarily a consequence of both folds and reverse faults having opposing vergence senses in nearly equal abundances except for subregions R1 and R2 that only have south- and west- or southwest-verging folds, respectively (Figs. 5B, 10, and 11).
Subregions R5 and R6 show little overlap between paleoslope methods, suggesting that the transport direction is poorly constrained in these regions. Although the cause of ambiguity in subregions R5 and R6 is unclear, we suspect that the wide spread in fold and fault orientations relates to the characteristically intense deformation that LF2 has undergone within the central portions of the NAMTD. For example, some noted that fold hinge orientations become increasingly scattered with increasing deformation as a result of several factors that include (1) the development of multiple generations of crosscutting folds, (2) folding and faulting in response to locally generated stresses rather than to the paleoslope gradient, and (3) variable fold rotation toward the transport direction (e.g., Alsop and Marco, 2013). The difficulty in interpreting folds and faults in subregions R5–R6 relative to other areas within the NAMTD suggests that some portions of the MTD are better suited for paleoslope analysis than others. These differences likely relate to deformation intensity that in turn is controlled by the stratigraphic character of the deformed strata or possibly the relative position within the MTD.
Our paleoslope analysis suggests that some methods are more applicable and yielded more consistent results than others. For example, the APIM-N does not overlap with other fold-based methods and probably does not yield an accurate estimate of the downslope direction (Fig. 22). In addition, methods that assume the mean fold axis is parallel to the downslope direction (e.g., SAM, DAM) are generally not applicable or yield ambiguous results. Fold axes are correspondingly interpreted to have an along-slope mean in at least subregions R1–R4 and R7. This conclusion is supported by the observation that thrust faults verge suborthogonally to the trend of the mean fold axes (Fig. 22). In addition, there is no compelling evidence for significant downslope rotation of fold axes (e.g., development of sheath folds or a systematic relationship between fold hinge azimuth and interlimb angle; Fig. 20). For these reasons, methods that rely on progressive layer-parallel simple shear (i.e., HIM, HAM, and ASIM) were found to be ineffective. The MBSM yielded the most constrained results for all subregions and was in good agreement with most other methods (Figs. 13 and 22). Whereas fold- and fault-based methods were ambiguous in subregions R5–R6, the MBSM yields a generally east- or west-directed transport that is consistent with adjacent subregions R2–R4 (Figs. 13 and 22).
Normal fault-based methods suggest northeast-directed transport, or possibly southwest-directed transport if the dominant fault population is antithetic with respect to the paleoslope (e.g., Alsop and Marco, 2014). However, these results rarely overlap with those based on folds, reverse faults, and bedding (Fig. 22). Although normal faults are typically assumed to strike on average in a direction parallel to the paleoslope (Farrell, 1984; Martinsen and Bakken, 1990), normal faults within the NAMTD have orientations that are markedly oblique to the inferred transport directions based on the other methods (Fig. 18). It is possible that some normal faults formed early and were subsequently overprinted by progressive deformation that altered their original orientations. However, this hypothesis is difficult to evaluate due to the scarcity of definitive crosscutting relationships between normal faults and folds. Alternatively, some normal faults could have formed in response to local stresses within the deforming MTD. For example, Alsop and Marco (2011) documented conjugate sets of normal faults that were oriented oblique to the presumed transport direction, and suggested that these faults formed in response to lateral spreading along the width of the MTD. In such cases, interpretation of normal fault orientations alone is likely to yield an ambiguous or erroneous estimate of the paleoslope orientation. We instead rely on the consistency between folds, reverse faults, and bedding orientations to aid in interpreting the overall MTD transport direction (Fig. 22).
Transport Direction of the NAMTD
Kinematic analysis of slump folds and faults within the NAMTD reveals that the predicted transport direction varies significantly within the MTD outcrop extent (Fig. 22). Moreover, the variability in the predicted MTD transport direction appears to be systematic: north or south transport directions are predicted along the inferred lateral margins of the MTD (subregions R1 and R7) and generally east or west transport directions are predicted for the central portions of the MTD (subregions R2–R4 and possibly R5; Fig. 22). In general, these observations can be explained by two end-member scenarios: (1) the NAMTD is composed of several separate MTDs that have different kinematic histories, or (2) the NAMTD is one large MTD that is characterized by spatially heterogeneous slump structure orientations.
The first scenario may be viable in the south due to the complex nature of deformation and occasional breaks in outcrop continuity that make it difficult to conclude whether the NAMTD represents a single, large mass-wasting event or is actually a composite of two or more juxtaposed events. However, field observations suggest that the first scenario is not tenable in the northern portions of the MTD (Figs. 4, 5A, and 5B). South-verging folds north of Kopuai Stream (subregion R1) can be directly traced into west-verging folds to the south (subregion R2) with no break in outcrop continuity (Figs. 5A, 5B). These same beds can then be correlated across a short break (∼200 m) in outcrop to northwest- and southeast-verging folds in subregion R3 and finally to west- and east-verging folds in subregion R4 (Figs. 5A, 5B, and 14). The continuity and repeated deformation of the same stratigraphic interval throughout subregions R1–R4 strongly suggests that the northern 4 km of the NAMTD was emplaced during a single mass-wasting event. Thus the dramatic north to south differences in fold and fault orientations must be related to spatial heterogeneity within the deforming MTD rather than separate, juxtaposed slumps. Moreover, only subregions within the inferred lateral portions of the MTD (subregions R1 and R7) display fold, fault, and bedding orientations that are systematically oriented in an east-west direction (Figs. 13, 18, and 22). Thus, we contend that the orientations of slump structures in subregions R1 and R7 most likely diverge from the rest of the NAMTD as a result of edge effects near the lateral portions of a single large MTD (Fig. 23).
The recognition of distinct central and lateral portions of the MTD greatly clarifies interpretation of the overall transport direction. Because portions of the MTD inferred to be in a central position generally yield east- or west-directed transport (67°–123° or 247°–303°), we infer that the MTD was likely emplaced along a north-south–striking paleoslope (Fig. 23). Determining whether the MTD traveled toward the west or east is somewhat ambiguous because most paleoslope methods are only able to constrain MTD transport to one of two possible directions (Fig. 22). However, several lines of evidence suggest that the NAMTD was most likely emplaced in an overall westward direction (toward ∼245°–300°).
Both bedding strike and fold orientations can be observed to fan ∼35° about a westward direction in the wave-cut platforms on either side of Waioroko Stream (subregions R3–R4; Figs. 5A, 5B, and 22). This fanning pattern is highly reminiscent of the typically arcuate pattern of pressure ridges found nearly ubiquitously on MTD surfaces on the modern seafloor and in the subsurface (e.g., Prior et al., 1984; Prather et al., 1998; Frey-Martínez et al., 2006; Bull et al., 2009). Because this fanning pattern is observed over a distance of only 1.5–2 km, this region likely represents a subordinate slump lobe, or possibly a second-order flow cell (sensu Alsop and Marco, 2014), within the larger NAMTD. We interpret the westward convexity of this slump lobe to be a strong kinematic indicator of west-directed transport in subregions R3–R4 (Figs. 5A, 5B).
Thrust faults in subregion R1 verge slightly obliquely (∼25° clockwise) to the large slump folds in this area (Fig. 22), which is consistent with an overall west-directed transport direction.
The nearby offshore presence of the NAMTD is equivocal due to problematic data quality in the Kahu 3D seismic survey, but it does not appear to project westward from ∼12 km west of the coastline where the Mohakatino interval is well imaged (Figs. 1 and 6). Had the MTD originated in the west, we might expect to see depositional evidence for an east-facing slope, including perhaps the headwall scarp of the MTD in the area covered by the Kahu 3D survey.
An angular unconformity between the Mohakatino Formation and the underlying Manganui-Mokau Group (undifferentiated) suggests that westward tilting occurred between middle and late Miocene time (Utley, 1987, fig. 6.3 therein). West-directed transport is consistent with the continuation of west-directed tilting into the early part of late Miocene time (see following discussion).
Other MTDs present in the northeastern Taranaki Basin within overlying stratigraphic intervals (e.g., the Otukehu MTD, lower Mount Messenger Formation) have inferred west- to north-directed transport directions based on both slump structure orientations and kinematic indicators observed in offshore seismic surveys (Strachan, 2002; King et al., 2011; Fig. 3).
Why Are Folds and Faults within the MTD Margins Oriented Parallel to the Inferred Downslope Direction?
Slump structures within the lateral margins of the MTD (subregions R1 and R7) have an average orientation that is subparallel to the inferred downslope dip direction, based on an assumed overall west-directed MTD transport direction (Figs. 22 and 23). Thus, the NAMTD provides important verification that downslope-parallel slump structures occur in large MTDs inasmuch as published models of margin-to-center structural relationships within MTDs have been largely based on terrestrial analogues, including a small tundra landslide (Hansen, 1971) and a slumped snow bed (Lajoie, 1972). We propose that the downslope-parallel slump structures within the margins of the NAMTD could have formed as a result of two end-member mechanisms that each can be accommodated via interrelated processes: (1) rotation of originally along-slope structures via downslope rotation during progressive layer-parallel shear or bulk vertical-axis rotation, and/or (2) in situ formation via layer-normal shear, shortening parallel to the paleoslope strike, or variable slope bathymetry.
Rotation of Slump Structures into a Downslope-Parallel Orientation
Downslope rotation of originally along-slope slump structures can occur via progressive layer-parallel simple shear (e.g., Farrell and Eaton, 1987) and/or bulk vertical-axis rotation. The former process seems unlikely to explain patterns of deformation within the NAMTD because folds within the marginal portions (subregions R1 and R7) show no evidence for progressive rotation with increasing deformation (e.g., lack of sheath folds and no systematic relationship between hinge orientation and interlimb angle; Fig. 20). However, bulk vertical-axis rotation at the NAMTD margins could conceivably account for oblique- or downslope-oriented slump structures within subregions R1 and R7. This process likely contributes to the typically arcuate pattern of downslope-fanning pressure ridges that are commonly observed within MTDs that undergo greater amounts of material translation in their central portions than in their margins (e.g., Prather et al., 1998, fig. 13 therein). Approximately 90° of clockwise and counterclockwise rotation would have to have occurred along the margins of the NAMTD within subregions R1 and R7, respectively, to account for observed slump folds and fault orientations (Figs. 13 and 18). Although the extent of vertical-axis rotation is poorly constrained along the southern extent of the NAMTD, several lines of evidence suggest that such large amounts of vertical-axis rotation probably did not occur along the northern margin of the NAMTD.
Flute casts on the base of a correlated sand bed in subregions R1 and R3 all have flow toward the eastern quadrant, suggesting limited vertical-axis rotation between these subregions (Fig. 2). The observed ∼60° divergence could be interpreted as counterclockwise rotation of subregion R1 relative to R3. However, this interpretation is inconsistent with a preferred west-directed overall MTD transport direction that would suggest that R1 was rotated clockwise relative to R3. Even if the MTD was emplaced toward the east, the 60° flute divergence is not great enough to account for the 110° of counterclockwise rotation required by fold orientations (Fig. 13).
Folds within subregion R1 verge toward the south and not toward the north, as would be expected if originally west-verging folds had been rotated clockwise to become downslope parallel at the northern margin of the NAMTD.
The transition from downslope- to along-slope–parallel fold axes occurs abruptly south of Kopuai Stream (Fig. 5A). The absence of structures with transitional orientations seems to be at odds with gradual rotation to a downslope orientation at the MTD margins.
Large-scale rotation of R1 relative to R2–R3 would likely create a space problem that is not observed in outcrop. As described here, the same stratigraphic interval can be traced from south- to west-verging folds without any apparent suggestion of rotation.
In Situ Formation of Downslope-Parallel Slump Structures
Some suggested that slump folds can form with orientations at low angles to the downslope direction (e.g., Debacker et al., 2009). In particular, folds formed within the margins of MTDs can undergo layer-normal shear that is predicted to result in steeply dipping axial surfaces and fold hinges that fan about the downslope direction (Alsop and Holdsworth, 2002; Debacker et al., 2009; Alsop and Marco, 2011). However, it is questionable whether the slump structures within the margins of the NAMTD were formed by layer-normal shear (Table 3). For example, fold distributions within subregions R1 and R7 display a positive relationship between interlimb angle and axial surface dip (Fig. 20), the opposite of the relationship predicted to occur as a result of layer-normal shear (Debacker et al., 2009; Table 3). Large slump folds within subregion R1 show evidence for layer-parallel simple shear (e.g., convex-upward axial surfaces) and display geometries similar to those of folds within subregions R3–R4 (Fig. 14). Thus it seems unlikely that downslope-parallel folds within subregion R1 and possibly R7 resulted from layer-normal shear.
Alternatively, downslope-parallel slump structures could have formed as a response to shortening across the width of the MTD (i.e., lateral compression) and/or variably sloping seafloor bathymetry. These processes could have acted in concert as changes in the seafloor gradient to the north and south of the NAMTD may have resulted in local confinement and inward-directed compression within the margins of the MTD (Fig. 23). We favor this interpretation for several reasons. (1) This scenario is consistent with the abrupt transition in fold orientations and bedding strike between subregions R1–R3 (e.g., Figs. 5A) and (2) may also help explain why the large folds present in subregion R1 are all verging south toward the interior of the MTD and not toward the north, as one would expect based on clockwise vertical-axis rotation of originally west-verging structures. (3) Confinement of the northern margin of the NAMTD by a generally south-dipping paleoslope is supported by paleocurrent measurements collected at Opito Point and Ngarupupu Point (Fig. 2). (4) Other examples of lateral compression at the margins of MTDs have been identified in seismic-reflection data sets (e.g., Posamentier and Walker, 2006, figs. 152 and 160 therein).
Spatial Variability in MTD Structural Architecture
As with slump structure orientations, the internal structural architecture of the NAMTD varies widely along its outcrop extent (e.g., Fig. 9). However, there is a less clearly defined relationship between spatial position within the margin-to-center transect and the style of MTD deformation. Rather, the lithologic character of the host stratigraphy likely played a major role in influencing the type and scale of slump structures. In general, regions dominated by LF1 (subregions R1–R4) tend to form the largest and most continuous folds found in the NAMTD (Fig. 9). Slump structures formed in LF2 are marked by an increase in the intensity and complexity of deformation that includes generation of typically smaller scale rootless, disharmonic folds that are intercalated with low-strain zones. Thus, we hypothesize that the dramatic change in MTD structural architecture that occurs south of subregions R3 is primarily a result of the decrease in LF1 relative to LF2 within subregions R4–R5 (Fig. 9).
The relative abundance of normal and reverse slump faults also changes markedly from north to south. Normal faults are comparatively uncommon in subregions R1–R3 relative to reverse faults (15 versus 60 normal and reverse fault measurements, respectively; Fig. 13). The opposite is true for subregions R4–R7 (129 normal versus 48 reverse measurements; Fig. 13). The scarcity of normal faults in subregions R1–R3, except where developed around fold hinges, suggests that this portion of the MTD did not undergo significant extension during initiation, translation, or while evolving to a compressional regime (Farrell, 1984).
Implications for Paleoslope Analysis
The NAMTD demonstrates that significant spatial heterogeneity in slump structure orientations and internal architecture can greatly complicate determination of the paleoslope orientation via kinematic analysis. This point is illustrated by the ambiguity that results from conducting a paleoslope analysis of all folds and faults without regard for their relative location within the NAMTD (Fig. 22). We found that a robust paleoslope interpretation was only made possible by accounting for spatial changes in fold and fault orientations along the MTD exposure (Fig. 22). This finding may account for the conflicting results yielded by previous estimates of the MTD transport direction that did not account for spatial variability along the outcrop extent (east- and south-directed transport; Utley, 1987; Nagel, 2010). Thus the NAMTD highlights the importance of interpreting slump structures within their spatial context and demonstrates that failure to do so can yield an erroneous paleoslope interpretation.
Our analysis also highlights the challenges of interpreting slump structures characterized by widely dispersed fold and fault orientations and senses of asymmetry. In some cases, scattered fold and fault orientations may be a consequence of complex and intense patterns of deformation (e.g., subregion R6; Alsop and Marco, 2013). Alternatively, such distributions may result from failure to account for systematic spatial variations in slump structure orientations. In such cases, we strongly recommend that the data set be examined for systematic spatial variability, and fold and fault measurements be interpreted within their spatial context.
In the case of the NAMTD, discrimination between the lateral and central portions of the MTD is critical for interpreting the overall transport direction and is largely made possible by regional geologic constraints. Unfortunately, recognition of lateral margins in outcrops of ancient MTDs is typically hampered by a lack of exposure and by subsequent tectonic deformation. Although criteria for recognizing lateral margins of MTDs based on slump structures have been proposed (e.g., Debacker et al., 2009), many of these criteria are not fulfilled in the regions that we infer to be within the margins of the NAMTD (Table 3). We suspect that this incongruity is a result of lateral compression along the margins of the MTD, rather than deformation via layer-normal shear. Thus the NAMTD demonstrates an important methodological limitation of some paleoslope methods that generally assume little shortening across the width of the MTD (Woodcock, 1979b) and suggests that it can be difficult to confidently discriminate between lateral and central portions of MTDs based upon the characteristics of slump structures alone, particularly when MTDs undergo lateral compression due to confinement against basin margins or local seafloor topography.
Paleogeography of the Northeastern Taranaki Basin
Our study of the NAMTD documents evidence for a spectacular episode of mass wasting in the northeastern Taranaki Basin during late Miocene time and corroborates, clarifies, and provides new insights into the previously inferred paleogeographic setting and tectonic evolution of the region (e.g., Nodder et al., 1990b; King et al., 1993). The predeformation sheet-like geometry and lateral continuity of deep-water Mohakatino Formation beds exposed along the coast between Opito Point and the Awakino River suggest that deposition by sediment gravity flows was primarily within an unconfined basin-floor setting. However at a slightly broader scale, this depocenter was likely surrounded on all sides by bathymetric highs (Fig. 24). The submarine volcanic arc that was the source of Mohakatino detritus formed seafloor topography to the west and north, effectively isolating the northeastern portion of the Taranaki Basin (Fig. 1). The Herangi submarine high likely provided an eastern flank to this subbasin and a backstop for Mohakatino fan deposition (King et al., 1993). Containment to the south is suggested by the inferred location of the late Miocene slope that prograded northward during and following Mohakatino Formation deposition (Masalimova, 2013; Rotzien et al., 2014). A depositional position near the subbasin center is supported by variable paleocurrent directions within the Mohakatino Formation (primarily west-, south-, and east-flowing; Fig. 2) and local interfingering of the Mount Messenger and Mohakatino Formations that suggest flows entered the basin from a variety of directions and with varying provenance (Figs. 2 and 24; Utley, 1987; Nodder et al., 1990a, 1990b). Although depositional facies of the Mohakatino Formation suggest a basin-floor setting, the ultimate base of slope was far to the west within the New Caledonian Basin (Fig. 1; Uruski and Wood, 1991). Thus the NAMTD was likely positioned within a local basin that was perched, or possibly ponded, within the western continental slope of New Zealand (sensu Prather, 2003; Prather et al., 2012; Fig. 24).
A surprising finding of our study is that the NAMTD was likely translated back toward the submarine volcanic arc that was the original source of sediment (Fig. 1). Although the transport distance of the NAMTD is uncertain, the distribution of lithofacies within the NAMTD closely matches that of the surrounding undeformed intervals. This strongly suggests that the NAMTD was generated locally and probably did not travel far (Fig. 24). Our interpreted west-directed MTD transport direction suggests that a change occurred in the orientation and gradient of the seafloor slope following deposition of the south- to east-flowing LF1 lithofacies of the Mohakatino Formation (Fig. 24). It is likely that minor westward tilting of the basin occurred even prior to deposition of the Mohakatino Formation based on local observations of a subtle angular relationship between middle Miocene bathyal mudstone and the overlying basal Mohakatino Formation (Utley, 1987). Within the past few million years, strata within the entire northern Taranaki coastal region have been progressively uplifted and tilted to the west or southwest within a regional monocline (King and Thrasher, 1996). Taken together, these observations suggest that the Taranaki Miocene sedimentary succession underwent a long history of regional westward tilting, both during and after deposition of the Mohakatino Formation.
Although the precise cause of the local change in the bathymetric slope orientation within the northeastern Taranaki Basin is unclear, it is likely that tectonically induced basin deformation played a major role. We speculate that the eastern basin margin (the Herangi submarine high) was uplifted during late Miocene time prior to ca. 10.5–10.9 Ma (Appendix DR5 in the Supplemental File [see footnote 1]) and that this tectonism resulted in some west-directed tilting and an associated slight reconfiguration of the subbasin depocenter (Fig. 24). A possible driver for uplift of the eastern basin margin is activity on the buried Taranaki fault that nominally forms the eastern boundary of the Taranaki Basin (Fig. 1). This interpretation is supported by the location of the NAMTD overlying the leading edge of the Taranaki fault hanging wall (Figs. 2 and 24). However, major displacement on the Taranaki fault in this vicinity is thought to predate ca. 19–18 Ma (Stagpoole and Nicol, 2008), and thus any continued activity during late Miocene time was probably relatively minor. Alternatively, disruption of the paleoslope may have been triggered by the coeval active volcanism and associated normal faulting that was occurring to the west (Fig. 1; Giba et al., 2013).
Mass Movement within Basin-Floor Successions
The NAMTD provides an important example of mass movement within a stratal succession originally deposited within a low-gradient basin floor, possibly in a perched or ponded basin. Mass failure in this setting is surprising because MTDs most commonly originate in outer shelf, upper slope, and mid-slope settings where seafloor gradients are typically higher (Posamentier and Martinsen, 2011). Correspondingly, documented examples of mass movement within basin-floor facies are somewhat uncommon and have generally been ascribed to one of two processes: (1) basin deformation that alters local bathymetric gradients and produces slope instabilities in basin-floor facies (Trincardi and Argnani, 1990; Haughton, 2000; Lucente and Pini, 2008), and (2) entrainment and/or deformation of basin-floor facies within an MTD that originated higher in the slope profile (van der Merwe et al., 2009; Posamentier and Martinsen, 2011). We do not favor the latter option because the NAMTD is composed exclusively of basin-floor facies and lacks evidence for strata derived from slope or shelf settings (e.g., van der Merwe et al., 2009).
We contend that the NAMTD provides a type example of mass movement of a deep-water basin-floor succession induced by tectonism. Basin tilting created a slightly steeper and destabilized slope gradient that caused sand-rich, basin-floor lobes of the Mohakatino Formation to be remobilized downslope to form the NAMTD (Fig. 24). A similar mechanism may account for mass movement of the lower Mount Messenger Formation (e.g., Otukehu MTD), which also incorporates lobe-like facies (Fig. 2; Masalimova, 2013). The Oligocene–Miocene Apennine foredeep provides a comparable example where basin margin uplift created a temporary slope from which basin-floor facies were remobilized orthogonally to the regional paleoflow direction (Lucente and Pini, 2008). Because of the inherently low slope gradients in basin-floor settings, any tectonic movements can readily create changes in seafloor slope orientation. Correspondingly, it is common for MTDs developed within basin-floor successions to have transport directions that are quite different from the regional paleoflow direction (e.g., Miyata, 1990; Haughton, 2000; Lucente and Pini, 2008). This scenario helps account for the coarse-grained sand-rich nature of the NAMTD that reflects the lithologic character of the staging area, which included proximal portions of coarse-grained lobe deposits. This example suggests that MTDs that comprise basin-floor lobes can be sand rich relative to those that originate on the continental slope where facies are more mud dominated (Posamentier and Martinsen, 2011).
The late Miocene NAMTD is spectacularly exposed in coastal cliffs and wave-cut platforms, provides a world-class example of mass movement in a deep-water basin-floor setting, and exemplifies the scale, style, and heterogeneity that is possible within this type of deposit. Moreover, the fortuitous coincidence of west-directed transport with north-south–oriented coastal exposures has resulted in a lateral cross section that extends across nearly the full width of the compressional domain of the NAMTD. This provides an exceptional, and possibly unique, opportunity to study the spatial variability in slump structure orientations and internal architecture from both the marginal and central portions of an MTD >10 km in width.
The NAMTD developed in strata deposited as sheets or lobes within a perched or ponded intraslope basin along the western paleocontinental margin of New Zealand. Sedimentary lithofacies within the NAMTD can be categorized into two end members that are interpreted to have been deposited within both inner lobe and outer lobe and/or lobe fringe environments, i.e., thick- to medium-bedded volcaniclastic sandstone (LF1) and thin-bedded volcaniclastic sandstone and mudstone (LF2), respectively. LF1 predominates in the northern 5 km of the study area but transitions gradually to LF2 toward the south. These lithofacies correspond with distinctly different styles of soft-sediment deformation. Slump folds developed in LF1 are typically larger (amplitudes to tens of meters), often form semicontinuous fold trains, and contain décollement zones within the fold hinges. Slump folds developed in LF2 are typically smaller, disharmonic, and occur as rootless individuals. Intensely deformed zones are often intercalated vertically or horizontally with low-strain domains.
Kinematic analysis of slump folds and faults compiled for the entire NAMTD yields an ambiguous interpretation of the transport direction. However, a clearer picture has emerged through separate analysis of seven subregions within the NAMTD that allows a robust interpretation of the paleoslope orientation to be made for five of the subregions that display overlap of multiple paleoslope estimations. Populations of both upslope- and downslope-verging folds and reverse faults are present throughout much of the NAMTD, and the paleoslope dip direction can only be constrained to two possibilities in these regions. By discriminating and accounting for these spatial changes in slump structure orientations and considering regional geologic constraints, we demonstrate that the NAMTD was likely emplaced within an overall west-facing paleoslope. Moreover, slump structures within the inferred margins of the MTD have orientations that are subperpendicular to those in the central portions of the MTD. This relationship is best expressed in the northern extent of the MTD, where the same stratigraphic interval is repeatedly deformed in a series of spectacular folds with orientations that systematically change by ∼110° over 1.5 km. Thus the NAMTD verifies that downslope-parallel fold axes can occur within the lateral portions of MTDs, possibly as a result of a variably sloping seafloor and/or lateral shortening within the MTD margins. Our analysis demonstrates the importance of studying slump structures in their spatial context and using multiple, independent methods of estimating the paleoslope orientation.
The submarine Mohakatino volcanic arc is the acknowledged source of volcaniclastic sediment to the subbasin; the inferred presence of an east-, southeast-, or south-facing paleoslope formed on the volcanic flanks is generally supported by paleocurrent measurements within and north of the MTD. Our interpretation of west-directed transport of the NAMTD, back toward the volcanic arc, suggests that a modification of bathymetric slope orientation must have occurred prior to emplacement of the MTD. We speculate that this occurred via local basin tilting that may have resulted from uplift of the Herangi submarine high to the east. Possible tectonic drivers of basin deformation include continued activity on the buried Taranaki fault and/or differential subsidence associated with active volcanism and normal faulting to the west. Such tectonism helps account for the development of a large MTD in sand-rich strata that were originally deposited in a likely ponded subbasin with very low seafloor gradients. The NAMTD may provide an analog for other unconfined, deep-water deposits in intraslope basins that have undergone mass wasting associated with basin deformation via tectonic or gravitational processes (e.g., salt or mobile shale movement).
Financial support for this study was provided by a Geological Society of America graduate research grant and by the Stanford Project on Deep-water Depositional Systems. We also thank the individuals from GNS Science who contributed to this study, including Greg Browne, Rob Funnell, Malcolm Arnot, Martin Crundwell, and Andy Nicol. This study benefited from discussion with Tim Debacker, George Hilley, Don Lowe, and Lorna Strachan. We thank Blair Chan, Tess Menotti, and Nora Nieminski for assistance in the field. We also thank the many New Zealanders who made this project possible, including Craig Rain at Paparahia Station; Alistair Bryant and Maree Bryant at Onetai Station; Karl Reipen at Awakino Estate; Dawn Colman and Neil Colman; John Potroz and Angela Potroz; Shane Marsden, Jenny Marsden, and Graham Marsden; and Alan Jones and his family. We would also like to thank Andrea Fildani, Ian Alsop, and one anonymous reviewer for constructive feedback and suggestions that greatly improved this manuscript.