Towards a tectonic framework for normal faults in Waitematā Group rocks, North Island, Aotearoa-New Zealand

ABSTRACT

Analyses of active and past normal fault behaviour in extensional settings provide key insights into regional-scale tectonic processes driven by plate boundary forces. To better understand past tectonic environments in the Tāmaki Makaurau-Auckland region of northern Aotearoa-New Zealand, we examine normal faults with Miocene to Quaternary movements dissecting the Early Miocene Waitematā Group. Structural data were collected from coastal outcrops, including fault geometry and kinematics, and reveal a dominant NE-trending strike direction, indicating NW-SE extension. Faults dissecting coeval and younger Waitākere Group rocks west of the study site exhibit the same NE-striking fabric, and thus may represent the westward continuation of Waitematā Group extensional structures. Furthermore, a NE-trending normal fault fabric is present within the adjacent Taranaki Basin, suggesting that the observed extension in the Waitematā Group was widespread over northern Aotearoa, with some extension still occurring today in southern parts of the Taranaki Basin. We propose that many of the observed normal faults in the Waitematā Group represent the northern portion of the Taranaki Basin fault system, with the Taranaki faults likely representing the younger continuation of the extension recorded in the Waitematā Basin, resulting from the migration of back-arc extension due to roll-back of the Hikurangi subduction margin.

KEYWORDS:

• Waitematā Basin
• back-arc rifting
• normal fault
• extensional tectonics
• Taranaki Basin

Introduction

Detailed geometric and kinematic investigations of ancient and active normal fault systems across the North Island of Aotearoa-New Zealand can provide insights into the geodynamic setting that has evolved in response to plate boundary processes (e.g. Knox 1982; Rowland and Sibson 2001; Villamor and Berryman 2001; Lamarche et al. 2006; Giba et al. 2010; Downs et al. 2014). In particular, the volcano-tectonic setting of the North Island of New Zealand fundamentally changed during the Eocene to Miocene when the Hikurangi Subduction Margin was established (Cole and Lewis 1981; Nicol et al. 2007; Seebeck et al. 2014a, 2014b). Though several rift basin and volcanic arc systems are known to have formed during this period (Hayward 1979; Herzer 1995; Booden et al. 2012), various aspects of the Miocene volcano-tectonic setting of northern Aotearoa-New Zealand are still debated. For example, there is still uncertainty regarding the tectonic and geodynamic processes driving the development of the Early Miocene Waitematā Basin, as well as the location and genesis of the surrounding volcanic systems that sourced the basin’s sediments through time (e.g. Shane et al. 2010).

To better understand the tectonic environment of northern Aotearoa from Miocene to present, we examine a section of normal faults with Miocene to Quaternary movements in the Waitematā Group. The aim of this work is to determine the likely formation scenario for many of the observed Waitematā Group faults, by determining the dominant fault type, orientation and kinematics using structural data from fault outcrops along the east coast of Tāmaki Makaurau-Auckland.

Geological background

The Waitematā Basin is located in northern Aotearoa-New Zealand, occupying the southern part of the Northland Region (Figure 1). The basin is thought to be an elongate depocentre, exposed onland ∼130 km southeast from near Whangarei to South Auckland, with a width of ∼60 km (Hayward 1993). It formed during the Upper Oligocene to Early Miocene (Ballance 1976) on a predecessor of the present Tonga-Kermadec subduction system (Stratford et al. 2022).

Figure 1. A, Location and approximate extent of the Waitematā Basin (Hayward 1993) in the North Island of Aotearoa-New Zealand, and the general extent of the study area in this study. The approximate extent of the Waitākere Group is based on maps of Hayward (1976). B, Inset showing the location of A with respect to Aotearoa’s plate boundary and the distribition of key basement terranes (Dun Mountain Ophiolite Belt and Median Batholith Terranes), illustrating the ‘Z-shape’ structure either side of the Alpine Fault associated with oroclinal bending (Lamb and Mortimer 2020).

The Waitematā Basin has a complex geology due to renewed Cenozoic convergent movement between the Australian and Pacific plates after the breakaway of Zealandia from Gondwana (Hayward 1976; Gaina et al. 1998; Laird and Bradshaw 2004; Mortimer 2004; Bache et al. 2014). In addition, its detailed stratigraphy is difficult to decipher, because of a loss of fossils due to diagenetic processes (Davidson and Black 1994; Raza et al. 1999) and an apparent lack of continuous marker lithologies at multiple stratigraphic levels. The basement that underlies the basin consists of three northwest-trending terranes of Palaeozoic to Mesozoic rocks (Mortimer 2004; Edbrooke 2017): the Murihiku terrane in the west, the central Dun Mountain-Maitai terrane, and the Waipapa terrane in the east. These originated during plate convergence between Panthalassa and Gondwana and were finally consolidated and metamorphosed by late Cretaceous time (Woldemichael and Black 2002; Adams et al. 2013; Jiao et al. 2014). The narrow, ophiolitic Dun Mountain-Maitai Terrane is a major shear feature of New Zealand and has experienced repeated deformations (Spörli and Aita 1995). Overlying these basement terranes unconformably, in part together with Eocene-Oligocene sedimentary rocks, the Waitematā Group consists of the Kawau Subgroup, East Coast Bays, Pakiri, Blockhouse Bay, Paremoremo, Timber Bay, and Cornwallis formations, and is dominated by quartz-poor, clay-rich turbidites sourced from volcanics on the margins of the basin (Ballance 1976; Hayward 1993; Edbrooke 2001; Spörli and Rowland 2007), with a few thick horizons of volcaniclastic conglomerate also present in many outcrops. This volcaniclastic conglomerate has been previously termed the ‘Parnell Grit’ (Gregory 1969; Ballance 1976), though we opt for the use of the term Parnell Volcaniclastic Conglomerate of Shane et al. (2010) to describe these horizons, as the term grit is a redundant term no longer applied in grain size schemes. Towards the west, the Waitematā Group transitions laterally and upward into the coarser-grained Waitākere Group, which hosts erupted material from the large and numerous volcanoes of the initial volcanic arc associated with the renewed subduction in the Oligocene (Ballance 1976; Hayward 1976, 1977, 1979, 1993). There is general agreement that this arc has subsequently migrated eastward relative to New Zealand to a present location at the southern end of the Tonga-Kermadec chain (Hayward 1993; Mortimer et al. 2010), with a southward migration component also suggested (Giba et al. 2010). In addition, some of the input of volcaniclastics to the Waitematā Group is not related to subduction-related andesitic volcanics, and instead has an ocean island basalt (OIB) chemistry, indicating a possible influence of lateral slab detachment, a localised slab window, or lateral termination of the subducting slab (Shane et al. 2010). In all, the degree to which this tectonically complex setting may have influenced the structural and lithological development of the Waitematā Basin is currently unknown.

The extensional tectonics during the separation of Zealandia from Gondwana (Bache et al. 2014), which preceded the establishment of the Oligocene to Present subduction system, created a widespread rhombic marginal network of crustal block faults. These block faults have continued to influence the sedimentary, tectonic and topographic development of Zealandia, often through their re-activation (Spörli 1980). For example, the interpreted re-activation of these inherited fault blocks is well-documented in the ‘basin and range’ landscape of the southern South Island some distance away from the active Alpine fault zone (McSaveney and Stirling 1992; Jackson et al. 1996, 2002). An analogue sub-rectangular NNW- and ENE-trending structural pattern is observed in the Auckland area of the North Island, also some distance away from the active plate boundary, and aligns sub-parallel to the basement tectonic grain and has likely been reactivated by younger faults (Spörli 1980; Kenny et al. 2012) (Figure 2). Examples of the NNW-striking set of faults include the active Kerepehi Fault in the Hauraki Rift (Hochstein and Nixon 1979; Hochstein and Ballance 1993; Persaud et al. 2016), and the active Drury and Wairoa Faults of the Hunua Ranges (Wise et al. 2003) (Figure 2). Importantly, this NNW-striking set of faults parallel the deep-seated NNW-trending fabric of the underlying basement terranes (e.g. Eccles et al. 2005; Ensing et al. 2022). The ENE-striking fault set is represented by the deep-seated, Port Waikato fault, as well as the Pōkeno and Mangatangi faults (Hochstein and Nunns 1976; Ensing et al. 2022) (Figure 2). Whilst there are a few active faults within the Auckland area and Waitematā Basin, the overall seismic hazard is low (Kenny et al. 2012; Sherburn et al. 2007), and the region is considered one of New Zealand’s most tectonically stable regions with only 80 earthquakes recorded above M_L 2.5 from 1983 and 2007 (Sherburn et al. 2007).

Figure 2. Locations of key field sites examined in this study, representing a cumulative length of ∼26 km along the eastern coast of Tāmaki Makaurau-Auckland. Also annotated are the traces of identified active faults (i.e. Seebeck et al. 2023) to the south and southeast of Tāmaki Makaurau-Auckland, which exhibit the ENE- and NNW-trending fault block pattern characteristic of the region. Dashed lines represent the approximated boundary of the faulted depression of the NNW-trending Hauraki Rift. Labelled field sites include: (1) Long Bay, (2) Torbay, (3) Browns Bay, (4) Campbells Bay, (5) Castor Bay, (6) Milford, (7) Takapuna, (8) Narrowneck, (9) St Helliers, (10) Glendowie, and (11) Musick Point.

Deformation within the Waitematā Basin is postulated to have been influenced by the establishment of the new subduction system under northern New Zealand that included movements on the Northland Allochthon, which was emplaced on top of the Northland autochthonous rocks during inception and propagation of the new plate boundary (Ballance and Spörli 1979; Hayward 1993; Isaac et al. 1994). The Northland Allochthon originated from east of the continental margin of Northland and consists of a lower level of Late Cretaceous to Palaeogene sedimentary sequences tectonically overlain by ophiolitic ocean floor volcanics (Ballance and Spörli 1979). It outcrops from the northern tip of the North Island to the vicinity of Tāmaki Makaurau-Auckland and has an equivalent in the East Cape region of the North Island. Although the timing of different fault development stages is challenging to constrain both locally and regionally, an initial compressional phase, associated with the emplacement of the Northland Allochthon, and westward thrusting in the Taranaki Basin (Giba et al. 2010), likely preceded a dominantly extensional phase occurring during later stages of Waitematā Group emplacement (Spörli and Rowland 2007).

The coastal outcrops of the Waitematā Group rocks that record these various phases of tectonic deformation vary in their structural complexity (Spörli 1989). Structurally simple areas contain horizontal or gently dipping strata, that are occasionally dissected by high-angle normal faults (Spörli 1989; Spörli and Rowland 2007; Irwin 2009). These comparatively undeformed regions alternate with zones of complex deformation that include soft sediment and weak rock deformation folds and faults (Gregory 1969; Spörli 1989; Spörli and Rowland 2007; Strachan 2008; Irwin 2009). The variations between simple and more complexly deformed zones are interpreted in part to be a product of the dynamic depositional environment (Spörli 1989; Hayward 1993; Irwin 2009), where sedimentary layers were remobilised en-masse, either from gravitational instabilities along the continental slope (Gregory 1969; Strachan 2008) and/or potentially triggered by earthquakes (Spörli and Rowland 2007; Butler et al. 2015).

Although studies have been performed previously on faults (dominantly normal) in the Waitematā Basin (Spörli 1978; Spörli and Browne 1981; Spörli 1989; Spörli and Rowland 2007; Strachan 2008; Kenny et al. 2012), no major compilation or synthesis of fault formation over a larger area has been made to date. Ballance (1964) drew attention to the presence of numerous small thrust faults in the Waitemata Group. Larger-scale thrust faults have been mapped by Codling (1970), Morris (1983), and Spörli and Rowland (2007), and are seen to predate a significant group of major normal faults. From geomorphological criteria, Kenny (2008) postulated thrust faults in the Waitematā Group, that are proposed to be related to the emplacement of the Northland Allochthon. These thrusts fan out from an inferred toe point of the allochthon. Mostly normal NW- to NNW-striking and NE- to ENE-striking faults have also been documented in many unpublished theses (e.g. Irwin 2009). Slip vectors of strike-slip faults are difficult to recognise in the Waitematā Group cliffs, especially in the sections with horizontal bedding, and have been seldom documented (e.g. Figure 3 in Spörli 1989; Spörli and Rowland 2007), although some faults could possibly have been reactivated with a dextral oblique-slip component (Wise et al. 2003; Irwin 2009). Seismic reflection surveys of the Waitematā Harbour have also revealed NE-striking, and possibly active, faults with unknown slip vectors that vertically displace the basement near the centre of Auckland City (Davy 2008).

In this study, we make an initial contribution to a regional synthesis of Waitematā Group faults to help understand the tectonic and sedimentary processes responsible for these structures, and their association with the formation of the Waitematā Basin and surrounding fault systems in the North Island of New Zealand.

Methods

Examination of the Waitematā Group faults was carried out along the coastline of the Waitematā Harbour of Tāmaki Makaurau-Auckland. The majority of normal faults observed in this study were found in less-disturbed sections with horizontal to shallow-dipping bedding. Extensional faults associated with the structurally complex zones in the Waitematā Group did not often meet our identification thresholds (outlined below), though many of these are likely to be only of local significance, as they bottom out into low-angle movement surfaces rather than continuing down to basement depths (Spörli and Browne 1981; Spörli 1989; Spörli and Rowland 2007; Fildes et al. 2011; Fildes 2013). Field sites on the North Shore of Tāmaki Makaurau-Auckland were examined between the northern end of Long Bay Regional Park (Figure 2, Field Site 1) and Narrow Neck, Devonport (Figure 2, Field Site 8). Field sites in eastern Tāmaki Makaurau-Auckland were examined between St. Heliers (Figure 2, Field Site 9) and Musick Point (Figure 2, Field Site 11). The defined extent of the surveyed area covers a coastline of ∼26 km (Figure 2) in a rough N-S direction. We acknowledge that the general ∼N-S direction of the coastline could limit our ability to identify and recognise sub-parallel structures and produce an E-W-directed bias in our structural dataset. Critically, the scale of the study region (i.e. ∼26 km-long coastline) combined with persistent local variations in outcrop orientation limit the use of a Terzaghi correction to account for this bias (e.g. Tang et al. 2018). Regardless of these limitations, we find that E-W-striking faults (i.e. those striking between 080° and 100°) make up a smaller proportion of our dataset than northerly-striking structures (i.e. those striking between 350° and 010°), equating to 6% and 13% of the data, respectively (see Results Section). Furthermore, previous local investigations of fault structures observed on shore platforms (e.g. Morris 1983) also reveal primary NE-striking structural trends similar to patterns observed in our study (see Results Section). Therefore, we find little evidence that the orientation of outcrop exposures has produced a significant directional bias in our dataset.

Structural data were acquired through observation and mapping only of cliff outcrops, as opposed to the shore platforms where irregular erosion and debris cover in many cases hindered acquisition of quantitative data and description of fault zones (Figure 3(b)), particularly when the strikes of faults were parallel to the strike of bedding. The main focus of the study was locating and analysing predominantly normal and a few steep reverse faults (many considered to be tilted normal faults, see ‘Summary of results’), as well as describing the associated stratigraphy and measuring bedding.

Figure 3. Examples of faulted outcrops in Waitematā Group rocks on the eastern coastline of Tāmaki Makaurau-Auckland. A, Example of a normal fault with decimetre displacements at Caster Bay, North Shore. White arrows indicate the position of the fault and a displaced marker bedd is annotated with an ‘X’. B, NE-striking normal faults dissecting Waitematā Group rocks on a shore platform, near Kennedy Park, North Shore. C, Fault breccia zone observed within a normal fault (down to the left) observed at Musick Point. D, Close-up image of the fault breccia in C. E, NW-SE-striking normal fault (down to the north) cemented by carbonate concretions, near Kennedy Park, North Shore. F, Example of a gouge-filled fault plane at Waiake Bay, North Shore. G, Calcite-filled faults in the Waitematā Group at Grannys Bay, North Shore. White arrows indicate the position of the calcite veins. H, Calcite-filled faults and fractures on a shore platform near Kennedy Park, North Shore. The veins and fractures seen here strike approximately NW-SE, the displaced bed has an offset of ∼40 cm.

With the exception of the conjugate faults described in this study, it is important to note that generally we cannot determine exact fault slip vectors of many of the observed faults, as only one fault striation was observed in the study region, likely due to the weak nature of the Waitematā Group rocks. Strictly speaking, the displacements visible in the cliff faces, and documented in our study, represent the observed vertical separation between the footwall and hanging wall cut-off, which we refer to as ‘fault throw’ in this study. However, because of the low bedding dips, the apparent normal fault separation observed in the cliff outcrops supports the assumption that the faults involved had a significant component of normal dip-slip. This assumption is supported by the observed conjugate fault couples, and our associated kinematic analyses of these, which support an extension direction roughly normal to the strike of the fault plane in most cases. Slickensides observed on the only observed striated fault surface in the study region also support a dominant dip-slip component (72° rake) to fault movements. We therefore continue to use the standard terms for describing fault slip in this study despite the precise slip vector being unconstrained for many of the measured faults, except for the conjugate fault examples.

To facilitate a regional-scale investigation, and ensure that the examined faults represent the major proportion of accumulated strain, faults were only analysed if they had a minimum throw of 0.2 m. A total of 154 faults met this criterion, and the following data were collected: (1) strike and dip of the fault, (2) sense of slip (i.e. normal or reverse), (3) throw of the fault, and (4) nature of the fault zone. The accuracy of throw estimates on the larger faults is less than faults with smaller offsets, as larger faults often could not be measured with a tape measure and were instead sighted from a distance. Structural fault information was recorded using the software, Field Move (2020, Petroleum Experts Limited, Edinburgh, UK), on an iPad 2016, 7th generation, and photographs were taken of every fault, with all data available in Supplementary Tables S1-S3.

Fault kinematic data were derived from the analysis of conjugate normal faults. This approach has been previously applied to faults dissecting Waitematā Group rocks (e.g. Spörli 1989; Spörli and Browne 1981).

Results

The host Waitematā Group rocks

Faults within the study area occur within the East Coast Bays Formation of the Waitematā Group (Hayward 1993) consisting of interbedded soft sandstones and mudstones ranging from finely laminated mudstones to thick-bedded, massive, indurated sandstone (Figure 3). Locally there are thicker layers of Parnell Volcaniclastic Conglomerate (e.g. Campbells Bay) with clast sizes ranging from pebbles to boulders. Typical for the East Coast Bays Formation, stretches of relatively flat lying beds (Figure 4) are interrupted by areas of complex structure (e.g. Spörli 1989; Spörli and Rowland 2007) exhibiting folds, faults and locally steep dips. This deformation is either due to purely gravitational or in part tectonically-induced mass movements (Spörli and Rowland 2007; Strachan 2008), probably associated with submarine channels (Fildes 2013).

Figure 4. Lower hemisphere stereonet plot of poles to bedding planes in the Waitematā Group in regions analysed in this study. The majority of beds are gently dipping, with occasional tilted, steeply dipping beds. All bedding data can be found in Supplementary Table S3.

Fault deformation zones

Within the Waitematā Group, a diverse range of fault cores were observed (Figure 3). Some faults were observed to have clean fault planes with little damage seen in the rocks adjacent to the plane (Figure 3(a)), while others have an extensive damage zone with arrays of slip surfaces and surrounding fractures, and multiple segments visible in some exposures (Figures 3(c,d,h)).

Four types of fault rocks were observed: (1) fault breccia, (2) fault gouge, (3) calcite vein-filled faults, and (4) fault planes cemented by concretions (Figure 3). Fault breccias are mostly observed on normal faults with throws >2 m. The breccias exhibit a clast-supported framework and contain angular clasts (Figure 3(d)). Zones of fine-grained fault gouge are seen on some faults with throws less than 1 m (Figure 3(f)). Veins are often seen within the damage zone of a fault, often linking segments together (Figure 3(h)). These veins can be complexly interconnected, but some are sub-parallel to each other. In some cases they follow the non-dominant orientation of a fault, joining the main fracture segments. The vein arrays appear to have opened due to a dilatational component to the normal faulting (i.e. extensional-shear fracturing). From our regional study, we observed only a single example of mineral striations, though further work on the geometry of the vein arrays may provide additional information on fault slip vectors. In some locations, erosion-resistant concretion walls mark the traces of normal faults (Figure 3(e)). These are most commonly observed around Musick Point (Figure 2, Field Site 11; Morris 1983) and are often located below the Parnell Volcaniclastic Conglomerate, whereas fault zones within the Parnell Volcaniclastic Conglomerate and above are comparatively uncemented.

Fault orientations and dimensions

Fault dips range from 31° to 89°, and the mean dip for normal faults is 64° and reverse faults is 69°. The mean strike for all faults analysed in our study area is 043 ± 22° (95% confidence) (Figure 5). Although a dominant NE trend in strike is observed in the dataset (Figure 5), lesser maxima appear to be present with NNW-SSE, N-S, and ENE-WSW strikes. Normal faults have a mean strike of 044°, whereas the mean strike of steep reverse faults is 033°.

Figure 5. Orientation data of faults within the Waitematā Group analysed in this study. A, Rose diagram of fault strikes. B, Lower hemisphere stereonet plot of poles to fault planes. C, Lower hemisphere stereonet contour plot of the data density of poles to fault planes. All fault directional data can be found in Supplementary Table S1.

The majority of faults analysed have throws between 0.20 and 0.40 m, with 36% of the observed faults exhibiting throws >1 m. Larger faults with more than 5 metres of throw were identified, but are less frequently observed (i.e. 11% of the data). The largest fault throw identified was 20 m.

Fault kinematics

Normal faults are the most common style of faulting, making up 82.5% of the dataset, with the remaining faults exhibiting either reverse (9.7%) or an unknown (7.8%) sense of slip. Stress directions can be derived from conjugate faults measured in the field, where the greatest and least principal stress axes bisect the acute and obtuse angles between the conjugate fault planes, respectively, and the intermediate stress axis corresponds to the intersection line between the conjugate fault planes (Anderson 1951; Spörli and Anderson 1980). These methods have also been used in the present study. Inferred extension directions from analyses of 17 conjugate fault couples (Supplementary Table S2) show two clusters (Figure 6): one in the NNW-SSE direction (k-mean 339^o) and the other in the SW-NE direction (k-mean 242^o). Two such orthogonal extension directions were also recognised by Spörli and Anderson (1980) in a regional summary of faults in the Waitematā Basin. The NNW-SSE extension direction would be compatible with the faulting on the dominant NE-striking faults (Figure 5) in our study and/or perhaps one of the inherited and reactivated fault sets (e.g. Port Waikato fault set, Figure 2). While the SW-NE extension direction does not have as many corresponding faults in our dataset, it may be an indication of the other inherited NNW-striking fault set (e.g. Drury, Wairoa and Kerepehi faults, Figure 2).

Figure 6. Lower Hemisphere stereonet plot of the orientation (dip and dip direction) of σ₃ based on the examination of conjugate fault pairs in the Waitematā Group. Mean directions of σ₃ for the two subpopulations are also plotted (k-means of 339° and 242°), which are consistent with the NE- and NNW-striking fault populations recognised in the study region. Conjugate fault data can be found in Supplementary Table S2.

Discussion

Summary of results

Measurements and descriptions of meso-scale normal faults in little disturbed sections of the Waitemata Group show a range of orientations (Figures 5 and 6) observed on a mostly NNW-trending coastline in Tāmaki Makaurau-Auckland. These orientations reveal a dominant NE-SW strike, and a suggestion of lesser maxima with NNW-SSE, N-S, and ENE-WSW strikes (Figure 5). Considering a possible outcrop bias associated with the N- to NNW-trending coastline, the NE-SW maximum may be somewhat over-emphasised, though is still considered here to be the most significant fault trend. By comparison, the NNW-SSE and N-S strikes may be somewhat under-represented (Figure 5). The ENE-WSW and the NNW-SSE minor maxima parallel the regional block faulting pattern (Spörli 1980; Edbrooke 2001; Kenny et al. 2012) inherited from post-Cretaceous separation of Zealandia from Gondwana (Adams et al. 2013; Bache et al. 2014). Of these, the ENE-WSW direction may be relatively over-emphasised in our data, whereas the NNW-SSE direction could be under-represented, again due to a possible outcrop bias associated with the N- to NNW-trending coastline. Small-scale conjugate fault couples support a dominantly normal type of fault movement with sub-horizontal extension.

Because of our sampling choice, the minimum throw documented for the faults is 0.2 m, and range up to greater than 15 m. The variety of fault rocks and damage zones (Figure 3) strongly suggests that these faults were formed over a long period of time relative to Waitematā Basin sedimentation. Specifically, these differences in the type of fault deformation require shearing in rocks that have significantly different lithification and related burial histories. A near-absence of fault surfaces (i.e. fault striations) observed along fault planes is consistent with fault slip in poorly consolidated sediments, likely not long after deposition of the Waitematā Group sediments in the Early to Middle Miocene. Similarly, carbonate concretion cemented faults (Figure 3(e)) represent faults forming during the early depositional stages of the Waitematā Group (Morris 1983). While we do not claim to have sampled the full range of these faults, it is interesting to note that quite a few of them strike in the NE sector (Morris 1983). Breccia-bearing faults (Figure 3(d)), on the other hand, necessitate fault slip in comparatively well-lithified Waitematā Group rocks, and hence represent a later stage of deformation, possibly post-dating Waitemata Group deposition. If the steep reverse faults are tilted normal faults and not the steep tips of listric thrusts (e.g. Twiss and Moores 1992), they must have formed prior to or simultaneously with this tilting. However, it is difficult to fix the relative position of this tilting in the deformation history of the Waitematā basin, and thus this tilting cannot be used to constrain the timing of faulting.

Tectonic significance of normal faults of the Waitematā Group

As previously discussed, the timing of fault movement in the Waitematā Group is challenging to constrain. Good cross-cutting relationships are very rarely observed (e.g. Spörli 1989). Since these faults cut through Early Miocene-aged lithologies, they either formed during or after deposition of the Waitematā Group; however, the documented fault rock types do support some normal fault movements early in the depositional history of the basin when the sediments were relatively unconsolidated. The orientations of some faults with post-Miocene movement documented by Kenny et al. (2012) also roughly correlate with Waitematā Group faults presented in this study, and thus some of the analysed faults could also have been active after the Miocene. For example, where the active Wairoa Fault (Wise et al. 2003) intersects the shores of the Waitematā Harbour at Beachlands a steep normal fault in the Waitematā Group rocks propagates upward into Pleistocene rhyolitic tephra (Hayward 2017).

The deformational history of the Waitematā Basin is also complex, as shown by the contrast between highly contorted and fractured sections, and sections exhibiting straight low dip beds (Spörli 1989; Spörli and Rowland 2007; Fildes et al. 2011; Fildes 2013). However, as most of our study region exposes the latter sections, we have increased the chance of capturing faults that have propagated through both the Waitematā Group and the underlying basement, and are therefore of greater regional significance than faults within complexly deformed sections, which may be dominantly intra-formational, and thus represent local responses to various tectonic and gravitational influences (e.g. Spörli and Rowland 2007; Strachan 2008; Fildes 2013).

Although the timing of fault formation is poorly constrained, we can still examine how the Waitematā Group normal fault system may fit into the tectonic and geodynamic framework of the North Island by comparing fault patterns with those observed in surrounding faults systems with known Miocene to Quaternary movements (Figure 7). In addition to the NNW- and ENE-trending basement fabric inherited from the separation of Zealandia from Gondwana (Adams et al. 2013; Bache et al. 2014), which is in part responsible for fault block patterns in South Auckland (Figure 2), we identify three other regional-scale fault and fracture systems of relevance to the Waitematā Group normal faults: (1) the Hauraki Rift, located to the southeast of the Waitematā Basin and in part derived from the inherited NNW fault pattern; (2) faults in the Waitākere Group to the west of the Waitematā Group; and (3) the Taranaki Basin, along the western side of the North Island.

Figure 7. Regional synthesis map of selected North Island faults. Pink: Locations of Waitematā Group faults measured in this study, with a mean strike of 043°. Blue: Taranaki Faults from Giba et al. (2010) in central to northern regions of the basin that were active from 12–4 Ma, with a mean strike of 033°. Red: Haruaki Rift Faults from Persaud et al. (2016), which have a mean strike of 156°. Active faults in the South Auckland region (Seebeck et al. 2023) are presented as dark grey lines.

The Hauraki Rift is a half-graben rift basin that extends from the northwestern side of the Taupō Volcanic Zone towards the Hauraki Gulf (Figures 1, 2 and 7) in a NNW direction (Persaud et al. 2016). It has been active since the Late Neogene (Hochstein and Nixon 1979; Hochstein et al. 1986), and the Kerepehi Fault, occurring within the basin’s depocentre, exhibits neotectonic traces (Lange (de) and Lowe 1990; Persaud et al. 2016). Although the geodynamic processes controlling its formation are poorly constrained, the Hauraki Rift is located in the back-arc region of the Hikurangi Subduction Margin, and it runs parallel to the extinct Coromandel volcanic arc as interpreted by Hochstein and Ballance (1993). The average fault strike of active faults within the Hauraki Rift is 156° (Figure 7) and the overall structural pattern of the rift may in part reflect the reactivation of the rhombic fault network created during the separation of Zealandia from Gondwana (Bahiru et al. 2019). Although these faults are younger than the Waitematā Group, they do parallel the minor NNW-SSE fault trend observed in our Waitematā fault dataset (Figure 4), and thus it is possible that a minor portion of the faults observed in the Waitematā Group may have moved/reactivated during the Quaternary and under the same extensional tectonic regime as the Hauraki Rift.

Miocene igneous and volcaniclastic rocks of the Waitākere Group are situated immediately west of the Waitematā Basin and are the on-land tip of a large offshore zone of volcaniclastic and igneous deposits associated with the Waitākere volcanic arc (Hayward 1979). Within these rock units, Taylor (2021) observed dominant NE-trending structures active during or after the Miocene, including normal faults, clastic and igneous dikes, and pervasive joint and vein systems (also see Hayward 1976). This dominant NE trend is observed across all western Tāmaki Makaurau-Auckland locations, where measured joint trends have a primary direction of 032° and fault strikes have a primary direction of 043°, suggesting a pervasive NE-SW structural trend. This structural grain is also reflected in the directional trends of geomorphic erosional lineaments in the Waitākere Ranges (Al-Jawadi 2020). Given the dominant NE trends observed in both Waitākere and Waitematā Group faults, and the Early Miocene age of both these groups, we suggest that some Waitākere Group faults likely represent a possible westward continuation of the Waitematā Group normal faults observed in this study.

The Taranaki basin has a more complete record of faulting compared to the Waitematā basin, derived from abundant offshore reflection seismic data collected off the western coast of the North Island for hydrocarbon exploration (King and Thrasher 1996; Giba et al. 2010). In the Taranaki Basin, sedimentation rates exceeded rates of faulting, and thus back-stripping methods can be adopted to detail the timing of fault evolution (Giba et al. 2010). The northern end of the Taranaki Basin lies close to, or is continuous with, the southern known limit of the Waitematā Basin, and arc volcanism, extensional tectonic activity, and basin development has propagated southward from the Waitematā Basin region since ∼12 Ma (Giba et al. 2010). Although the Taranaki Basin experienced a significant period of westward thrusting focused along the Taranaki Fault from 40 Ma to 12 Ma, there has been dominant tectonic extension from the Middle Miocene to Present, with a potentially still active volcanic back-arc rift basin at its most southern extent (Giba et al. 2010). The mean fault strike for recently active (∼12-0 Ma) normal faults within the Taranaki Basin is 033° (Figure 7), which is in close agreement with the dominant normal fault orientations in the Waitematā Basin (mean strike is 043°). Furthermore, mapped Taranaki Basin faults within 100 km of the Waitematā Group study site exhibit a mean orientation of 050° (Giba et al. 2010). Kinematic data collected on <12 Ma Taranaki Basin normal faults occurring onshore support dominantly NW-SE-directed extension. Fault striations on the northernmost Taranaki Basin faults, within 20 km of the Waikato River mouth, indicate a 154–334° mean extension direction (Giba et al. 2010), within 5° of the dominant extension direction (159–339°) inferred from our study of Waitematā Group faults. Extensional fault patterns are asymmetric in the Taranaki Basin, and it is interpreted that faults on the basin’s western part nucleated on pre-existing faults inherited from the separation of Zealandia from Gondwana, while the eastern set appears to be dominantly newly formed (Giba et al. 2010)

In all, given the similar trends, kinematics and possible ages of these structures, it is plausible that a significant proportion of the normal faults dissecting the Waitematā Group represent the northern extent of those also in the Taranaki Basin, which formed in a back-arc extensional setting (Giba et al. 2010). A still unresolved question is whether 40–12 Ma convergence in the Taranaki basin extended into the Waitematā Basin, causing at least some of the many thrust faults observed in outcrop in the Tāmaki Makaurau-Auckland region (Ballance 1964; Spörli 1989; Giba et al. 2010).

Given the Miocene-Quaternary movement of the Waitematā Basin normal faults, it is plausible that some of the extension recorded here in the Waitematā Basin immediately predated normal faulting in the Taranaki Basin in the Early Miocene, which is also shown to have propagated southwards since the Middle Miocene (Stern and Davey 1989; King and Thrasher 1996; Giba et al. 2010). Thus, this phase of Miocene extension recorded along the west coast of northern Aotearoa-New Zealand possibly initiated in the Waitematā Basin, before propagating southward through the Taranaki Basin.

We emphasise that, with the detailed evidence for the longevity of faults in the Taranaki basin and the presence of active faults in the Tāmaki Makaurau-Auckland region, it is likely that a portion of the normal fault slip recorded in this study also postdates deposition of the Waitematā Group. Examples may include the subordinate NNW- and ENE-trending structures, which could relate to more recent Quaternary block faulting prominently observed in the South Auckland region. However, unlike the faults with Quaternary movements observed in the South Auckland region (e.g. Wairoa and Waikopua North Faults; Wise et al. 2003; Gasston et al. 2021) (Figure 2), Waitematā Group faults mapped from coastal exposures in this study have not yet been shown to exhibit a geomorphic expression in the modern landscape, and only one has so far been observed to offset Quaternary age deposits (e.g. Beachland fault, Hayward 2017).

Finally, the Tamaki Makaurau-Auckland region lies a short distance north of the northern hinge of a regional-scale orocline defined by the Junction Magnetic Anomaly (Figure 1). This structure forms part of a larger ‘Z-shaped’ orocline observed in the North and South Islands of Aotearoa-New Zealand, and is offset by ∼750 km across the Alpine Fault (Figure 1). Regional bending likely initiated prior to 45 Ma and possibly as early as the Mesozoic (Lamb and Mortimer 2021) and is recorded in both palaeomagnetic data and the regional curvilinear pattern observed in basement terranes across both the North and South Islands. Although this bending could have driven Neogene faulting in the Tamaki Makaurau-Auckland area, oroclinal bending is frequently associated with the development of mostly strike-slip and compressional structures (Miller et al. 2002; Hollingsworth et al. 2010; Li et al. 2018), rather than extensional features (i.e. the NE-striking Waitematā Group normal faults). Furthermore, an unresolved question is whether much of this bending in the North Island is of Neogene age (e.g. Sutherland 1999; King 2000; Nicol et al. 2007; Rowan and Roberts 2008; Lamb 2011; Seebeck et al. 2014a), as some authors favour a reconstruction in which the northern oroclinal bend in the Tāmaki Makaurau-Auckland region predates deposition of the Waitematā Group (Lamb and Mortimer 2020). Although we cannot completely rule out the role of oroclinal bending as a driving mechanism for some Waitematā Group faulting, we see little evidence to support this bending as the causal mechanism for the extensional structures recorded in our dataset. On the other hand, the oroclinal bending possibly affected the rhombic fault pattern inherited from plate separation and contributed to its present pattern.

Conclusions

The fault structures presented in this study represent the most exhaustive compilation of extensional Waitematā Group faults to date. Results reveal a dominant NE-SW structural trend observed along a 26 km-long section of exposed coastal outcrops. Normal faults are the most common style of faulting, making up 82.5% of the dataset, with kinematic analyses of conjugate normal faults and the dominant NE-striking fault system supporting a primary NW-SE extension direction, in addition to a minor NNW-SSE fault trend and associated NE-SW extension.

Potentially synchronous normal fault structures are observed in the Waitākere Group rocks immediately west of the study site, which exhibit a similar NE-SW strike cluster. Due to the similarity in depositional age of the Waitematā and Waitākere Groups and the parallel structural trends, the Waitākere tectonic structures could represent the westward continuation of the normal fault system of the Waitematā Group. A similar NE-trending normal fault fabric is also present within the adjacent Taranaki Basin to the south, suggesting that the observed extension in the Waitematā Group was widespread in northern Aotearoa. Based on the current interpretation for the tectonic setting of the Taranaki Basin, we suggest that back-arc extension resulting from roll-back of the Hikurangi subduction margin likely drove much the observed extensional tectonics recorded in the Waitematā Basin. Observed normal faults in the Waitematā and Waitākere Groups are interpreted as the oldest normal faults observed in an extensional tectonic regime that subsequently propagated southward into the Taranaki Basin.

We emphasise that this study represents an initial contribution to a regional synthesis of fault types found in the Waitematā Group, focusing on a 26 km-long section of outcrops observed on the eastern coast of Tāmaki Makaurau-Auckland. Future studies can test these results and interpretations through investigations of the E-W coastal exposures on the northern side of the Manukau Harbour, nearest to the northern end of the Taranaki basin, which should also reveal the contributions of other potential major fault zones to the tectonic development of the region.

Acknowledgements

We are grateful to reviewers Frank Chanier and Mark Rattenbury who provided thorough and insightful reviews that greatly improved this manuscript. JM acknowledges a University of Auckland Early-Career FRDF award for field support during this study. Jennifer Eccles is thanked for discussions on the faults and general structure of the Auckland region.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

The datasets presented in this study can be found in an online repository at https://zenodo.org/record/7958221.