Earthquake history at the eastern boundary of the South Taupo Volcanic Zone, New Zealand

ABSTRACT

At the eastern boundary of the south Taupo Rift, the NE-striking, rift-bounding Rangipo and the ENE-striking Wahianoa active normal faults intersect. We investigate their intersection at the Upper Waikato Stream to understand the kinematics of a rift termination in an active volcanic area. The Upper Waikato Stream Fault is a previously unrecognised seismogenic source also at the eastern boundary, capable of producing a M_W6.5 and up to M_W7.1 earthquake if it ruptures in conjunction with the Rangipo or Wahianoa faults. We found a minimum of 12 surface-rupturing earthquakes in the last 45.16 ka on the Upper Waikato Stream Fault (mean slip-rate c. 0.5 mm/yr), and a minimum of nine surface-rupturing earthquakes in the last 133 ka on the Wahianoa Fault (mean slip-rate c. 0.2 mm/yr). Periods of highest slip-rate on these faults may coincide in time with Taupo, Ruapehu or Tongariro eruptions, but, despite their intersection, movement was not coincident across all faults. The Upper Waikato Stream Fault responded to a major Taupo Volcano eruption, the Wahianoa to a major eruptive sequence from Mt Tongariro and the Rangipo to major explosive events from Mt Ruapehu.

KEYWORDS:

• Paleoseismology
• Rangipo Fault
• slip-rate
• Taupo Rift
• Taupo Volcanic Zone
• tectonics
• Tongariro Volcanic Centre
• Upper Waikato Stream Fault
• Wahianoa Fault

Introduction

At the southern end of the Taupo Rift (and Taupo Volcanic Zone) in the Central Plateau of the North Island of New Zealand, local reorientation of the stress tensor axes is responsible for a complex tectonic structure, marking the abrupt closure of the rift (Villamor & Berryman 2006a; Villamor et al. 2007). Three sets of active normal faults occur in this area, which are highly oblique to each other (Figure 1).

Figure 1. A, Tectonic setting of New Zealand. Green rectangle marks location of Figure 1B. B, Location of the Taupo Rift (active faults in red from Langridge et al. 2016). Yellow rectangle marks location of Figure 1C. C, Location map of the Mt Ruapehu Graben, Upper Waikato Stream (UWS), Wahianoa and Rangipo faults in the southern part of the Tongariro Volcanic Complex, in southern Taupo Rift. Potential intersection area marked with a purple rectangle (UWS).

NE-striking faults delineate the Taupo Rift margins and axis, whereas E–W- and ENE-striking faults delineate the southern termination. At the southernmost part, mapping of active fault scarps has shown E–W- and ENE-striking faults hard linked with NE-striking faults. For example, the NE-striking Raurimu Fault intersects the E–W-striking Ohakune Fault in the southwest, and the NE-striking Rangipo Fault intersects the ENE-striking Shawcroft Road Fault in the southeast (Villamor & Berryman 2006) (Figure 1). Along the eastern boundary and north of the intersection between the Rangipo and Shawcroft Road faults, prior studies have identified the Rangipo and Kaimanawa faults as rift-bounding faults between volcanic sediments and the greywacke Kaimanawa Mountains; although only the Rangipo Fault is currently mapped as active (Langridge et al. 2016). Here, the ENE-striking Wahianoa Fault trends towards the rift boundary but there is no geomorphic expression of intersection with the Rangipo Fault, due to cover by young, undisturbed recent sediments. However, active faults do traverse the banks of the Upper Waikato Stream, suggesting that it is very likely that the Rangipo and/or the Wahianoa faults extend northeast, beyond their intersection (Figure 1).

Understanding the geometry, kinematics and paleoearthquake history of this fault intersection and rift termination may have important implications for seismic hazard assessment in the Tongariro Volcanic Centre (TgVC). In particular, confirming the prolongation of the Rangipo and Wahianoa faults into the Upper Waikato Stream area would increase their potential rupture lengths and thus the associated earthquake magnitudes.

This study characterises the geometry, kinematics and faulting history of the Upper Waikato Stream and Wahianoa faults. Subsequently, we compare their faulting histories with that of the Rangipo Fault (Villamor et al. 2007). The comparison allows us to assess whether these faults rupture simultaneously or interdependently, and whether their activity is temporally associated with major eruptions from the nearby volcanic centres (Figure 1).

Geological and structural setting

The Taupo Volcanic Zone (TVZ) is a NNE-trending Quaternary volcanic arc associated with subduction of the Pacific Plate below the Australian Plate offshore of the North Island of New Zealand along the Hikurangi margin (Wallace et al. 2004). The volcanic arc is known for its 25 caldera-forming eruptions in the last 1.6 Ma, active andesitic stratovolcanoes, parasitic vents, such as scoria cones, geothermal and hydrothermal activity within an active fault belt with horst and graben structures (Wilson et al. 2009; Leonard et al. 2010; Pardo et al. 2014) (Figure 1). Faults within the volcanic arc (Rowland & Sibson 2001) define an intra-arc rift, the Taupo Rift (Acocella et al. 2003). Extension is created by fore-arc clockwise rotation (Wallace et al. 2004) and slab rollback (Seebeck et al. 2014). Extension across the rift is associated with a combination of dike intrusion and normal faulting (Rowland et al. 2010; Villamor et al. 2011).

Volcanic history and stratigraphy

The TgVC is the southern part of the TVZ. It is a 13 km long and 5 km wide basaltic–andesite to dacitic volcano–vent corridor complex (Hackett & Houghton 1989) within the Mt Ruapehu Graben (Villamor & Berryman 2006). The TgVC includes two major stratovolcano systems, and a number of minor vents. Ruapehu is the southernmost and largest volcano and is one of the most active volcanoes in New Zealand. It is a complex andesitic stratovolcano with a crater lake over its active vent, and is one of the world’s most regular producers of lahars (Cronin et al. 1997a; Procter et al. 2010). The eastern flanks of Ruapehu (Whangaehu River, Mangatoetoenui Stream and Wahianoa River) have been a major pathway for lahars overflowing towards the Desert Road since prehistoric times (Lecointre et al. 2004). The eastern extent of the Whangaehu River drains into the Upper Waikato Stream and eventually the Tongariro River (Figure 1).

Detailed studies of the volcanic stratigraphy in this region have provided excellent constraints on the timing of active faulting. Many volcanic deposits provide diachronic chronological markers as they are displaced by different amounts due to repeated fault movement (Villamor et al. 2007). For this reason, we present here a summary of the chronostratigraphic framework used to assess the paleoearthquake history of the Upper Waikato Stream and Wahianoa faults.

In the Upper Waikato Stream area, 33 andesitic tephras comprise a comprehensive chronostratigraphic framework (Topping 1973; Donoghue et al. 1995; Cronin et al. 1996a, 1996b; Donoghue & Neall 2001; Pardo et al. 2012). These are interbedded with lahar deposits from c. 18 separate events (Cronin & Neall 1997; Cronin et al. 1997a; Donoghue & Neall 2001) and the distal fall deposits of 19 rhyolitic eruptions from Taupo and Okataina volcanic centres (Froggatt & Lowe 1990; Wilson et al. 1995; Cronin et al. 1996b; Lowe et al. 2013) (Figure 2).

Figure 2. Chronological stratigraphic summary of the laharic and lava formations and andesitic tephras from the Tongariro Volcanic Centre interbedded with distal rhyolitic tephras from the Taupo Volcanic Zone calderas. Notice the general stratigraphic record for the Upper Waikato Stream and Wahianoa Fault areas at the right. *The oldest laharic formations’ ages are based on the Rotoehu ash age (45.16 cal ka BP; Danišík et al. 2012). Ru, Ruapehu; Tg, Tongariro; Ng, Ngāuruhoe; Tp, Taupo; Ok, Okataina, R# corresponds to the lahar episodes as described by Cronin & Neall (1997).

The largest of the TgVC volcanoes, Mt Ruapehu, has a near-continuous eruptive history from > 250 ka to the present (Hackett & Houghton 1989; Tost et al. 2015). The most complete tephra records of Mt Ruapehu occur east of the volcano, along the Desert Road. The Upper Waikato Stream includes some of the oldest and deepest outcrops with a continuous record of regional volcanic activity over the past > 50 ka (Cronin et al. 1996a; Cronin & Neall 1997), making it an ideal site for our study (Figure 2). Lahars have been channelled down the Upper Waikato Stream and the Tongariro River several times (Cronin et al. 1997b), but since c. 25 ka they stopped flowing this way, due to the formation of a moraine in the upper Whangaehu River during the last glaciation (Cronin et al. 1997a) (Figure 3).

Figure 3. Geological map and general stratigraphy of the SE Mt Ruapehu area (after Townsend et al. 2008; Lee et al. 2011), showing the location of the study sites (sections 1–3) and active surface fault traces (Langridge et al. 2016). See Figure 2 for detailed stratigraphic record.

Most of the rhyolitic tephras are radiocarbon dated (Topping 1973; Froggatt & Lowe 1990; Lowe et al. 2013), combined with isotopic dating methods such as ²³⁸U/²³⁰Th isotopes and (U–Th)/He dating of zircon (Danišík et al. 2012). Identification of the tephra is normally carried out by glass and mineral chemistry using electron microprobe analysis (Cronin et al. 1996a; Moebis et al. 2011). Isotopic K/Ar and ⁴⁰Ar/³⁹Ar dating have been used for lava flows (Gamble et al. 2003; Conway et al. 2015). Cosmogenic nuclide dating is also used for glacial sediments (Eaves et al. 2015).

Active faulting: the Rangipo and Wahianoa faults

At the southern termination of the TVZ and the Taupo Rift, radial extension (strain) and local reorientation of the stress axes are responsible for a complex fault pattern around Mt Ruapehu (Rowland & Sibson 2001; Villamor & Berryman 2006; Villamor et al. 2007). Three major active fault sets form the southern termination of the rift: (1) the NNE-trending Mt Ruapehu Graben; (2) the ENE-trending Karioi fault set; and (3) the E–W Raetihi fault set (Villamor & Beryman 2006). Immediately northwards, at the latitude of Tongariro and Ngāuruhoe volcanoes, normal fault displacements together with extensional fissures and dike intrusions accommodate extension along a NNE trend (Rowland & Sibson 2001). Volcanic vents are also NNE aligned, along the axis of the Mt Ruapehu Graben (Nairn et al. 1998; Acocella et al. 2003; Villamor & Berryman 2006).

Rangipo Fault is located at the eastern margin of the Mt Ruapehu Graben (Villamor & Berryman 2006). It is an active, NNE-striking, normal fault downthrown to the west (Figure 1). The fault is at least 32 km long with a clear fault scarp ranging from a few metres to tens of metres in height. To the south, the Rangipo Fault is bounded by the Snowgrass Fault, but its northern extent remains uncertain due to erosion of the scarp by the Whangaehu River and/or by burial with recent lahar deposits. Prior to this study, the fault was mapped only as far north as the intersection with the Wahianoa Fault (Villamor et al. 2007). Donoghue et al. (1991) estimated a displacement rate of 3 mm/yr, but Villamor et al. (2007) calculated from paleoseismic trenches and geomorphic analyses a displacement rate of 1.4 ± 1 mm/yr from 17.7–27 cal ka BP to the present day, and of 0.2 mm/yr from 13.8 cal ka BP to the present day. These data imply an average long-term slip-rate of 1.5 mm/yr for the last c. 65–230 kyr and a period between 17.7–25.4 cal ka BP and 13.64 cal ka BP, where slip-rate could have been up to 9 mm/yr. Past earthquakes on the Rangipo Fault had estimated magnitudes of between 5.9 and 7.1 with recurrence intervals from 3.4 ka to > 14 ka, depending on whether rupture was segmented or not (Villamor et al. 2007).

The Wahianoa Fault is an ENE-striking normal fault downthrown to the SE, and is the main fault of the Karioi fault set. It has a prominent geomorphic expression for at least 10 km across the SE flanks of Mt Ruapehu, but the northeastern side of the fault is obscured by younger deposits of lahars and tephras from the TgVC. Preliminary slip-rates were assessed for this fault by Villamor and Berryman (2006a) at 0.3 ± 0.2 mm/yr. It is not certain whether the Wahianoa Fault extends beyond the current mapped trace and it might be buried under young lahar deposits. Based on the current National Seismic Hazard Model, this fault is capable of producing M_W6.7 earthquakes based on a 27 km fault length (Stirling et al. 2012; Litchfield et al. 2014). Earthquake magnitude could be considerably more if the fault is, in fact, longer.

Methods

To assess the possible connection between the Rangipo and Wahianoa faults in the Upper Waikato Stream area, we initially reviewed the current active fault mapping (Langridge et al. 2016) in the area of intersection, before analysing fault offsets of known stratigraphic units in outcrops, and determining offsets of geomorphic surfaces of known age. The data were used to reconstruct the timing of fault rupture and estimate fault slip-rates, earthquake magnitudes and recurrence intervals for the main structures in this area.

Fault mapping

Fault mapping in the southeastern TVZ was based on aerial photography interpretation from stereoscopic pairs from the 1940s and 1960s, at 1:50,000 and 1:16,000 scales (NZ Aerial Mapping, runs 1727, 1738, 2436, 2564 and 2572) along with Google Earth images. Grid references on maps are in NZTM2000.

A 2 m resolution photogrammetry-based digital surface model (DSM) was used as a base map for mapping and to calculate surface displacements. The DSM of the entire study area, within the boundary of the Tongariro National Park, was developed from two different aerial surveys conducted in November 2010 and November 2012, respectively, by New Zealand Aerial Mapping Ltd as part of their contract to provide regional aerial coverage to Horizons and Waikato regional councils. The 2010 and 2012 aerial surveys were flown in N–S and E–W orientations, respectively, and captured imagery using VEXCEL UltraCam Xp and UltraCamX large-format digital frame cameras (Microsoft) from a flight height of c. 7200 m a.s.l., which resulted in a ground sampling distance of 40 and 50 cm, respectively. The aerial blocks, consisting of 343 and 68 frames for 2010 and 2012 surveys, respectively, were triangulated using an iterative least-square bundle block adjustment in ERDAS Photogrammetry software.

The extraction of the DSM was conducted using the enhanced Automatic Terrain Extraction (eATE) module of ERDAS Photogrammetry. The algorithm used in eATE applied certain steps to achieve 3D generation from stereo-pairs. The steps encompassed hierarchical image matching, including reverse matching, by the mean of normalised cross correlation and followed by the refinement of matching results using least squares fitting. The eATE method generated a 3D coordinate for every pixel within the overlapped area of each stereo-pair and used maximum three images for simultaneous matching thus generated dense point-cloud data in laser format. These point cloud data from both survey blocks were further processed in LP360 software using an inverse distance weighted surface algorithm to achieve 2 m resolution DSM. The ground control points used to triangulate the 2010 aerial block contained ellipsoidal heights in NZGD2000 (NZ Geographic Datum 2000) that returned DSM vertical heights different from the 2012 aerial block, which was triangulated using normal-orthometric heights in NZVD2009 (NZ Vertical Datum 2009). The NZ Geoid2009 height model was subtracted from the 2010 DSM that attained its vertical heights compatible with the 2012 DSM. The DSM and a hillshade map were analysed with ArcMap 10.2.1 software (Esri).

Existing fault and geological mapping information were taken from: the GNS Active Fault Database (Langridge et al. 2016); the 1:250,000 Hawke’s Bay geological map (Lee et al. 2011), as well as mapping soon to be completed as part of a new geological map of Tongariro National Park (Leonard et al. 2014).

Fault geometry and displacements

We measured fault geometry (strike and dip), movement sense and fault displacements (mainly throw but occasionally also net slip) using terrestrial laser scanning (TLS) and geomorphic analysis of the DSM. Fault strike and dip were measured in the field using a geological compass. Dip–slip offsets were measured with a tape measure. The tape measure error is 0.1 m plus 8.4% of the measurement (equivalent to the highest percentage error using TLS).

TLS is a ground-based technique used to measure the position and dimension of objects with a rotating laser beam in which thousands of individual points are acquired (x, y, z). In the Upper Waikato Stream, we used a Leica Nova MS50 MultiStation TLS to acquire panorama pictures of the river exposures in conjunction with 3D geo-referenced points, with varying spacing between 2 and 4 cm at c. 30 m distance, which give a 0.03 m resolution on the rasterised data. We used eight scanning locations; each surveyed by a real-time kinematic GPS, which was used as a back-sight point during laser scanning (Figure 4). The data were processed with Cyclone 8.1 (Leica) and ArcGIS 10.2.1 (Esri) software packages. The merged point cloud data were meshed using the 3DReshaper Application 2014 (Technodigit) to obtain a digital outcrop model of the Upper Waikato Stream in order to measure fault displacements. The reported values (Table 1) are the average of multiple measurements of the offsets in the digital outcrop model.

Table 1. Net-slip, single-event displacements, progressive displacements and timing of rupture of the Upper Waikato Stream Fault, at section 1.

Faults	1		2		3		4		5		Events
Offset (m)	Net-slip	SED/PD	Net-slip	SED/PD	Net-slip	SED/PD	Net-slip	SED/PD	Net-slip	SED/PD	N event (Total SED/PD; unknown %)
Taupo pumice (1.72 ka)		?		?		?					?
Pahoka Tephra (11 ka)		?		?		?					?
Waiohau Tephra (13.64 ka)		?	?	?		?					?
Rerewhakaaitu Tephra (17.7ka)		?	–	?		?					?
Okareka Tephra (21.9 ka)		?	–	?		?					?
Hokey Pokey eruptive period		?	–	?	?	?		?	0.01 ± 0.15	0.01 ± 0.15^a	9 (0.01 ± 0.2 to 0.09 ± 0.2 m; 90%)^a
Ōruanui Tephra (25.4 ka)		?	1.44 ± 0.24	1.44 ± 0.2^a,b	0.12 ± 0.11	0.12 ± 0.11^a		?	0.16 ± 0.24	0.15 ± 0.29	8 (1.71 ± 0.4 to 2.2 ± 0.4 m; 36%)
R10 lahars (26.45 ka)		?	3 ± 0.34	1.57 ± 0.41	0.46 ± 0.12	0.35 ± 0.16		?	–	?	7 (1.93 ± 0.4 to 2.73 ± 0.4^b m; 29.6%)
Marker unit 1 (olive tephra)	?	?	3.66 ± 0.39	0.65 ± 0.51	0.46 ± 0.12	0 ± 0.17	?	?	–	?	6 (0.65 ± 0.5 to 0.92 ± 0.5 m; 29.6%)
R11 lahars (26.45 ka)	0.69 ± 0.22	0.69 ± 0.22^b	3.92 ± 0.44	0.26 ± 0.58	0.81 ± 0.23	0.35 ± 0.26	1.73 ± 0.25	1.73 ± 0.3^a,b	0.43 ± 0.17	0.27 ± 0.3^c	5 (1.03 ± 0.7 to 2.76 ± 0.8^b m; 0%)
Okaia Tephra (28.6 ka)	–	?	?	?	–	?	–	?	?	?	?
Omataroa Tephra (28.2 ka)	–	?			0.92 ± 0.16	0.12 ± 0.28^c	2.08 ± 0.36	0.35 ± 0.44^c			4 (0.47 ± 0.5 to 1.24 ± 0.5 m; 62.2%)^c
Orange lapilli	–	?			–	?	2.54 ± 0.43	0.46 ± 0.56			3 (0.46 ± 0.6 to 1.79 ± 0.6 m; 74.3%)
Elephant surge (marker unit2)	1.04 ± 0.17	0.35 ± 0.28^c			1.15 ± 0.25	0.23 ± 0.3	2.54 ± 0.35	0 ± 0.55			2 (0.58 ± 0.7 to 1.21 ± 0.7 m; 51.94%)
R12 lahars	1.04 ± 0.39	0 ± 0.43			?	?	?	?			?
Hauparu Tephra (36.1 ka)	1.62 ± 0.31	0.58 ± 0.5									1 (0.58 ± 0.5 to 5.64 ± 0.5^b m; 89.7%)^c
Rotoehu Ash (45.16 ka)	?	?									?
R13 lahars (> 45 ka)
Total	4.39 ± 0.57	4	12 ± 0.71	4	3.93 ± 0.43	6	8.89 ± 0.71	4	0.6 ± 0.33	3	9
Fault observations	Wall a 006-026/75-86NW		Wall b & c 037/68NW		Wall b & c 029/>80NW		Wall b & c 005-020/81-84NW		Wall b & c 012-043/72-89SE
Fault termination	R11 lahars		Covered by <17.57 ± 0.6 ka paleo-channel		25.4 ± 0.2 ka		Covered but inferred from fault 5 in 23.6 ± 1.8ka		Hokey Pokey		Unknown
Mean slip-rate (mm/yr)^d	0.03 ± 0.02 (36.1 ka)		0.11 ± 0.03 (26.45 ± 1.1 ka)		0.02 ± 0.02 (32 ka)		0.07 ± 0.02 (34.05 ± 2.1 ka)		0.01 ± 0.01(26.45 ± 1.1 ka)		0.07 ± 0.06 (36.1 ka)
Other comments	Scarp-derived colluvial wedge > R12 lahars, filled with Hauparu Tephra and a grey ash deposit, event shortly after 36.1 ka		0.3 m wide fissure filled with alluvial debris of reworked tephra smeared in fault		Multiple steps around R11 lahars		One event just after 36.1 ka, over-thickening layer above Elephant surge (free fault-face eroded). Multiple-steps at Elephant surge		Antithetic normal fault of fault 4

Net-slip, total displacement associated with each chronological maker; SED/PD, single event displacement/progressive displacement; N event, minimum number of rupture events; total SED/PD, minimum and maximum total progressive displacement; unknown %, unknown percentage of total PD relative to the 100% displacement measured on R11 lahars. Note that we do not differentiate between single-event displacement or progressive displacement here because of these values could represent one rupture (single-event displacement) and more than one rupture event. All values are in metres. All the ages are cal ka BP. See Methods section for data accuracy. Vertical displacements and more detailed description of each fault can be seen in Table S1.

^aLast event known, uncertain timing. –, no value; ?, not sure; ___, fault termination.

^bMore than one event.

^cCould have happened after this time.

^dSlip rate based on net-slip.

Displacement measurements for the Wahianoa Fault (section 3) were taken from 10 locations along the fault where a clear scarp displaces geomorphic surfaces of various age. For this, the 2 m resolution DSM was used, leading to an uncertainty of 1 m, plus 10% of the scarp height, due to the irregularity of the ground surface. For error propagation, we used the theory of errors (Taylor 1982), as described in Villamor and Berryman (2006a).

Stereonet 9.2 software (RW Allmendinger) was used for plotting the fault planes on the stereographic projection and azimuth frequency with rose diagrams. The net displacements measured in the field were converted into vertical offsets; vertical displacements obtained by the TLS were transformed into dip–slip offset using a fault dip range between 50° and 70°, as suggested by other studied faults within the Taupo Rift (Villamor & Berryman 2001; Nicol et al. 2006; Mouslopoulou et al. 2008). The Plane Vector Convention was used to report the dips (Dip/Dip Dir).

Earthquake history and single-event displacement

Paleoseismological investigations provide direct evidence from stratigraphic data to reconstruct the rupture history of the main faults in the Upper Waikato Stream and Wahianoa faults. Earthquake history was assessed by analysing progressive displacement on tephra layers of increasing age. Most faults have produced multiple earthquakes, as evidenced by the increase in fault displacement with increasing age of the tephras. Event horizons were used in association with differential displacements to identify new faulting events (McCalpin 1996).

Subtraction of consecutive total displacement values results in an estimate of co-seismic displacement or single-event displacement. This value may correspond to multiple events (progressive displacement) if several fault ruptures have occurred between depositions of horizons of known age.

Slip-rate, earthquake magnitude and recurrence interval

Knowing the fault displacement accrued over a specific period allows calculation of the fault slip-rate. Variations in fault slip-rates over time are used to compare the faulting history of the Upper Waikato Stream Fault with that of the Rangipo and Wahianoa faults, and with periods of enhanced volcanism in the Taupo Rift.

Earthquake moment magnitudes (M_W) are calculated using the fault scaling relationship developed for the TVZ (equation 1) (Villamor et al. 2007; Stirling et al. 2012), based mainly on the surface-rupture length. Fault lengths are based on the findings of our study. The single-event displacement was calculated with subsurface length and event moment magnitude (M_W).(1)

Single-event displacements and fault area are used to assess the different fault length scenarios. Also, in conjunction with fault slip-rate, single-event displacements are used to estimate fault rupture recurrence intervals when possible. For fault width, we assume a 15 km thick seismogenic crust, as suggested by previous seismic studies (Hurst & McGinty 1999; Hayes et al. 2004; Stirling et al. 2012).

Results of paleoseismic analysis

Upper Waikato Stream Fault (sections 1 and 2)

The Upper Waikato Stream Fault is located 15.3 km east of Mt Ruapehu’s active crater lake (Figures 1 and 3). It is a low-relief alluvial fan area, representing an important catchment for regional volcanic tephras and Ruapehu-sourced lahars from the last 50 ka to the Late Glacial warm period (Cronin et al. 1996a, 1997b; Lee et al. 2011; Figures 3 and 5).

Section 1

Section 1 of the Upper Waikato Stream Fault is located immediately adjacent to the Desert Road (State Highway 1) (Figures 3 and 5) and extends for 0.4 km along the Upper Waikato Stream. Here, five exposures were surveyed with a TLS (Figures 4–5 and S1). Wall ‘a’ is a NE–SW-striking outcrop, 65 m long and 27 m high; walls ‘b’ and ‘c’ are parallel, NNW-striking outcrops 155 m long and 22 m high; and walls ‘d’ and ‘e’ are parallel, NNE-striking outcrops 62 m long and > 13 m high (Figures 6 and S2). Seven main faults and extensional fractures were analysed in this section (Figure S1).

Figure 4. Terrestrial laser scanning survey. A, Upper Waikato Stream (UWS) 3D point cloud data. B, Oblique view of the digital outcrop model of the wall ‘a’ in the Upper Waikato Stream Fault built with 3DReshaper Application. C, Satellite image (using Google Earth images, 2015) showing the scanning locations and river exposures (walls) for the terrestrial laser scanning survey in the Upper Waikato Stream, section 1 (see location on Figure 3).

Figure 5. Location of the main faults in the Upper Waikato Stream area. A, Hillshade map of the Upper Waikato Stream based on 2 m DSM showing the studied faults in sections 1 (green rectangle) and 2 (blue dots, blue rectangle). Fault exposures (dots) and fault traces (continuous lines) are differentiated from the inferred faults (discontinuous lines). Faults downthrown to the NW are marked with red lines; faults downthrown to the SE are marked with yellow–brown lines. The fault planes are plotted in stereographic (lower hemisphere) projection and superimposed rose diagrams of fault strike frequency (right-hand rule). B, Section 1 of the Upper Waikato Stream (green rectangle in A) showing the seven main fault traces (purple dots), as well as an extensional fracture and two transects t1 and t2 (white lines) used to sum and calculate total offset for the Upper Waikato Stream Fault.

Figure 6. Faults in section 1 in the Upper Waikato Stream. A, Wall ‘a’ main fault exposure, showing fault 1 displacing volcanic deposits older than c. 28 ka (left, field photo; right, interpretation with stratigraphic column). B, Wall ‘b’ showing the location of fault 2 with a 0.3 m ‘fissure fill’ (left, field photo; right, interpretation with stratigraphic column). C, Detailed exposure of fault 4 on wall ‘b’ displacing tephras older than R11 lahars. The numbers show the elevation in m a.s.l. See Figure 2 for further information about stratigraphic units.

Several fault traces have been observed (Figure 5) that we interpret as merging into a single fault plane at depth given their close across-strike proximity (e.g. < 100 m). We also interpreted how these faults splay and merge as we linked them from outcrop to outcrop. The dip direction of the faults was used to assess their continuity from outcrop to outcrop along strike. Based on this interpretation, we infer that the active faulting in the Upper Waikato Stream represents a single fault structure, which we refer to as the Upper Waikato Stream Fault (Figure 5).

All of the faults studied in section 1 are NE-striking normal faults, dipping to the NW; except for fault 5, which is an antithetic fault and is downthrown to the SE (Table 1 and Figure 6). Based on our interpretation that these faults coalesce into a single fault at deeper levels, the data were analysed across two transects, which incorporate most deformation accommodated by the Upper Waikato Stream Fault. Transect 1 includes faults 1–5 and transect 2 includes faults 2–7 (Figure 5). We refer to each individual fault as a strand or splay and the sum of fault strand displacement values represents the total fault displacement. For transect 1, fault 5 is antithetic, so its motion was subtracted from the offset sum of faults 1–4.

Based on progressive displacement increments of individual strands, transect 1 shows evidence for at least nine surface-rupture events between 36.1 ka BP (Hauparu Tephra) and 21.9 ka BP (Okareka Tephra) (Table 1). In Table 1, we display the measured net displacement values together with the analysis of progressive displacement for each individual fault, as well as summed values for faults within the same transect. We use net displacements on the R11 lahar deposits (c. 27 ka; Figure 2; Cronin & Neall, 1997) as our best reference to estimate fault slip-rate and to assess distribution of deformation across the transect (across fault strands), because this marker appears in all outcrops. The total displacement across transect 1 (f1 + f2 + f3 + f4 – f5) on R11 lahars is 6.72 ± 0.62 m (net-slip), and at transect 2 (f6 + f7; note that offsets for f2–5 are missing), the offset is 5.43 ± 0.72 m (Table 1). The difference likely reflects the missing f2–f5 measurements on transect 2 (not exposed). Thus transect 1, which strikes perpendicular to the fault strike, is the best representation of the total deformation across the Upper Waikato Stream Fault, whereas along the riverbank to the north, exposures are sub-parallel to the faults and faulting information is probably missing because displacement on some makers could not be measured for all faults. The lack of events on some fault strands is related to two possible causes: (1) the Upper Waikato Stream Fault is comprised of several parallel fault strands at a single location and thus it is possible that not all strands ruptured at once (see e.g. Paeroa Fault, North Taupo Rift, Berryman et al. 2008) and not all strands are exposed at each location; (2) not all recognisable horizons were exposed at each of the fault exposures, or they were obscured by vegetation.

The progressive displacement analysis of marker R11 lahar deposits may also help us to assess the amount of deformation that might be missing on other chronological markers. It seems that all fault splays in the Upper Waikato Stream Fault could have ruptured at the same time across transect 1 for some individual surface ruptures because progressive displacement is associated with all strands. Assuming that the summed progressive displacement for R11 lahars in transect 1 is the best representation of the Upper Waikato Stream Fault movement, we can assess the percentage of deformation that is accommodated by each individual fault strand, namely 10%, 58%, 12%, 26% and −6% (antithetic fault), for faults 1, 2 3, 4 and 5, respectively. These percentages can be applied to other chronological markers for faults without information in order to assess total deformation at different times. When we use these percentages for other tephra layers, it suggests that we may miss between 30% and almost 100% of the total progressive displacement on those other markers. With this exercise, we are assuming that all faults always rupture during individual earthquakes. This is a conservative (maximum) approach because it has been observed on normal faults that splay into several fault strands at the surface, that surface rupture can occur on different fault splays for different events in a random way (see, e.g. Berryman et al. 2008). However, this approach allows a conservative estimate for displacement increments that can be compared with the sum values obtained using only the information that we have measured on the outcrops (minimum).

Extensional fractures

In section 1, on the NE side of wall ‘a’, there is a > 12 m long narrow fissure, dipping 030/68NW (Figure 7) that cuts through tephra and debris flows which are older than 18–16 ka BP. The fissure is 0.05–0.19 m wide and filled with the Taupo-sourced rhyolitic Ōruanui ignimbrite and associated tephra fall (25.4 cal ka BP; Wilson et al. 2006; Vandergoes et al. 2013). At some levels, the reworked Ōruanui Tephra shows finer grain-size banding and cross-bedding stratified layers, suggesting that water was involved in the transport–depositional process.

Figure 7. Extensional fractures located on wall ‘a’ in the Upper Waikato Stream. A, The 12 m long fissure filled with Ōruanui Tephra (white arrow; > 25.4 cal ka BP; photo taken by S. Donoghue), cutting older tephras and lahar deposits. B, Other fissures were measured on the same wall, cutting R11 lahars. C, These fissures do not show any vertical displacement, but record extension. D, The R11 lahar deposits are abruptly truncated at the fissure margins, with no mixed material along fissure sides. The reworked Ōruanui tephra shows cross-bedding stratified layers. E, All the fissures reflect orthogonal extension and are consistent with the NNE-trending Mt Ruapehu Graben. Rose diagrams for the measured fissures and faults in the Tongariro Volcanic Centre (TgVC), where strike directions (right-hand rule) are plotted.

Just below this fissure, on the same wall, there are > 10 other fissures cutting R11 lahar deposits (27 ka), some also filled with reworked Ōruanui Tephra. They vary slightly in orientation from 350° to 035°, generally matching the preferred direction of the main faults in southern TVZ (027°), downthrown to the NNW with steep dips (80°–90°) (Rowland & Sibson 2001), and similar to the orientations of the faults described here for the Upper Waikato Stream Fault. There is also a normal 170/78W fault that vertically displaces the R11 lahars by 0.18 ± 0.14 m (0.21 ± 0.14 m of net displacement) (Figure 7).

The main Upper Waikato Stream fissure was noted earlier by Donoghue and Neall (2001), while Topping (1973) described two similar structures in the Poutu canal (Figure 1) and another section nearby (60H 5678969.46 m S, 385626.35 m E). The sites described by Topping (1973) are no longer exposed, but the fissures were described as relating to the Ōruanui eruption, and are up to 0.52 m wide and striking at 356, cutting through lapilli and tuff deposits older than Ōruanui ignimbrite.

Section 2

Section 2 of the Upper Waikato Stream Fault starts 0.5 km downstream (NE) from section 1, and continues for 5 km downstream of the Tongariro River from there (Figures 3 and 5). Measurements on these faults were made using conventional field techniques, but only the lower few metres of the riverbanks could be reached, the remainder is bush covered or inaccessible. Eleven fault planes were described along this section, which helped assess fault continuity and kinematics, when we combined results with those of section 1 (Figures 8 and S3).

Figure 8. Main outcrops of the Upper Waikato Stream Fault in section 2. A, 10/4_2: Normal fault displacing the Papakai Formation. B, 10/4_3: Exposure of fault displacing deposits including R13 and older lahars.

Earthquake history

A summary of the fault strike and dip, net-slip offsets and progressive displacement analysis is shown in Table 2. Faults 11/4_1, 11/4_2 and 11/4_3 are not included in Table 2 because the stratigraphic position of the deposits that they display is uncertain. These data confirm that our summed values (see below) are minimum offsets because we might be missing some faults. Fault 11/4_1 crops out at the Tongariro River and is a 075/60S normal fault that cuts the Mesozoic greywacke sequence in the Kaimanawa Mountains (Lee et al. 2011). Fault 11/4_2 includes two normal faults: 042/64SE with > 6 ± 0.3 m of vertical displacement (> 6.9 ± 0.96 m of net-slip); and 035/72SE with 0.5 ± 0.2 m of vertical displacement (0.58 ± 0.22 m of net-slip). Fault 11/4_3 is a small normal fault 060-striking and dipping 76° NW with 0.4 ± 0.2 m of vertical offset (0.46 ± 0.21 m of net-slip). Three faults in section 2 (10/4_3, 10/4_5, 11/4_2; Figure 5) are dipping to the SE (antithetic).

Table 2. Net-slip, single event displacement and timing of fault rupture of the Upper Waikato Stream Fault, at section 2.

Faults	10/4_2	10/4_3	10/4_4	10/4_5	10/4_6	11/4_4 and 11/4_5	Events
Offset (m)	Net-slip	Net-slip	Net-slip	Net-slip	Net-slip	Net-slip	(Total SED/PD)
Taupo pumice (1.72 ka)	?						?
Mangatawai Tephra (3.52 ka)	0.54 ± 0.19						? (0.54 ± 0.19 m)^a
Papakai formation (3–11 ka)	?					?	?
Poutu lapilli (11 ka)			?			0.75 ± 0.2	2 (0.75 ± 0.21 m)^b
Pahoka Tephra (11 ka)			–			?	?
Okupata (11.77 ka)			2.37 ± 0.42				? (2.37 ± 0.42^c m)^b
Waiohau Tephra (13.64 ka)			?				?
Rerewhakaaitu Tephra (17.7 ka)							?
Okareka Tephra (21.9 ka)						?	?
Hokey Pokey eruptive period						0.92 ± 0.2	? (0.17 ± 0.31 m)^b
Ōruanui Tephra (25.4 ka)						2.3 ± 0.42	? (1.38 ± 0.48 m)
R10-R11 lahars (26.45 ka)						?	?
Okaia Tephra (28.6 ka)							?
Omataroa Tephra (28.2 ka)					?		?
R12 lahars			?		–		?
Hauparu Tephra (36.1 ka)		?	–		–		?
Rotoehu Ash (45.16 ka)		–	–		–		?
R13 lahars (>45 ka)		–	–		–		?
Marker unit 3		–	–		1.44 ± 0.3		? (1.44 ± 0.3)^b
R14 lahars (>55 ka)		3 ± 0.52	3.46 ± 0.58	?	–		1 (≥1.1 ± 0.72 m)^b
R15 lahars (<133 ka)		?	?	2.43 ± 0.44	?		? (2.43 ± 0.44)
Total	0.54 ± 0.19	3 ± 0.35	5.83 ± 0.72	2.43 ± 0.44	1.44 ± 0.3	3.98 ± 0.52	8
Fault observations	005-028/77NW	015/72SE	355-020/46-84NW	030/72SE	015-020/60-72NW	010-025/70-76NW
Fault termination	uncertain	<45 ka	uncertain (<11 ka)	uncertain	?<36.1 ka Unconformity	Uncertain (<11 ka?)	Unknown
Mean slip-rate (mm/yr)^d	0.21 ± 0.08 (2.54 ± 0.23 ka)	>0.07 ± 0.02 (>45 ka)	0.13 ± 0.04 (55 ka)	>0.02 ± 0.0 (133 ka)	0.02 ± 0.01 (36.1 ka)	0.07 ± 0.03 (25.4 ± 0.2 ka)	0.08 ± 0.07 (133 ka)
Other comments	Multiple-steps at Papakai F.	10 m-high river exposure	Erosive paleo-channel <55 ka	4-step fault at >133 ka	10 m-high exposure 3-step fault >36.1 ka	4 faults & 2 fractures, tilted deposits

Net-slip, total displacement associated with each chronological maker; SED/PD, single event displacement/progressive displacement; N event, minimum number of rupture events. All values are in metres. All ages are cal ka BP. See Methods section for data accuracy. Vertical displacements and more detailed description of each fault can be seen in Table S1.

^aLast event known, uncertain timing. –, no value; ?, not sure; ___, fault termination.

^bCould have happened after this time.

^cMore than one event.

^dSlip rate based on net-slip.

Based on the analysis of progressive displacement of individual fault strands, we distinguish a minimum of two events in section 2 from 133 to 1.72 cal ka BP. The first event took place between 55 and 11.77 cal ka BP, possibly approximately < 25 ka, and the second event < 11 cal ka BP; this latter event is additional to those found in section 1. A total of 12 events can be extracted for the information analysed in sections 1 and 2. Our data show a gap of seismic activity between 21.9 and 13.64 cal ka BP for the Upper Waikato Stream Fault (Tables 1, 2, 4 and S1).

Wahianoa Fault (section 3)

The Wahianoa Fault clearly displaces lava flows, moraines, tephras and lahar deposits for 10 km along the southeast flanks of Mt Ruapehu (Figures 3, 9 and S4). We traced 10 topographic profiles across the fault on geomorphic surfaces of known age in order to assess the fault’s earthquake history in section 3. We also analysed three outcrops within the Karioi Forest that expose the Wahianoa Fault (Figure 9).

Figure 9. Wahianoa Fault profiles and outcrops. A, Geological map (Leonard et al. 2014) showing the location of the studied field sites along the Wahianoa Fault in section 3; and the 10 topographic profiles (39)pehu. ld photo; B the TgVC (e) traced with their corresponding vertical offset values in metres. The geological map overlies a hillshade created with a 2 m-resolution Digital Surface Model from the SE flank of Mt Ruapehu. B–C, The two main outcrops of the Wahianoa Fault in section 3 showing displaced tephra (field locations shown on Figure 9A). The fault planes are plotted in stereographic (lower hemisphere) projection and superimposed rose diagrams of fault strike frequency (right-hand rule). Most of the tephra belongs to the Bullot Formation from Mt Ruapehu (Donoghue & Neall 2001; Pardo et al. 2012). See Figure 2 for more information about the stratigraphy and Figure 3 for location of section 3.

Earthquake history

All the Wahianoa Fault planes exposed in the outcrops in section 3 are NE-striking, SE-dipping normal faults. At the ‘Tonga 23’ field site, the faults exposed are antithetic to the main fault (Figures 9 and S4). Unfortunately, we could only analyse progressive displacement on exposed fault planes for the antithetic faults (Table 3), which helps us assign timing to some individual rupture events, but not to assess single-event displacements for the main fault. That fault shows that two ruptures occurred between 23 and 11.77 cal ka BP.

Table 3. Net-slip, single-event displacements, progressive displacements and timing of rupture of the Wahianoa Fault, at section 3.

	Outcrops				Profiles		Outcrops + profiles N events
Faults	Tonga 15	Tonga 23	Tonga 24	N event
Offset (m)	Net-slip	Net-slip	Net-slip	(Total PD)	Net-slip	PD	(Total SED/PD)
Taupo pumice (1.72 ka)				?		?	?
Mangatawai Tephra (3.52 ka)		?		?		?	?
Poutu lapilli (11 ka)		–		?	?	?	?
Poronui Tephra (11.2 ka)	?	–		?	6 ± 1.5	6 ± 2.12^a	9 (6 ± 2.12^b)^a
Pahoka Tephra (11 ka)	–	–		?	11.5 ± 4	5.5 ± 5.3	8 (5.5 ± 5.32^b)
Pourahu member (11.77 ka)	–	–		?	–	?	?
Okupata (11.77 ka)	0.69 ± 0.2	1.07 ± 0.2		2(0.69 ± 0.2)^a	–	?	7? (0.69 ± 0.17)^c
Shawcroft Tephra	–	1.44 ± 0.2		?(0.37 ± 0.3)	–	?	(0.37 ± 0.31)
Waiohau Tephra (13.64 ka)	?	1.77 ± 0.3		?(0.32 ± 0.4)	–	?	(0.32 ± 0.36)
Karioi eruptive period (15 ka)		–		?	23 ± 3	12 ± 3.9^c	6 (12 ± 3.9^b)^c
Rerewhakaaitu Tephra(17.7 ka)		–		?	30 ± 4.5	7 ± 6	5 (7 ± 6^b)
Last glacial maximum (>18 ka)		2.04 ± 0.3	?	?(0.28 ± 0.4)^c	–	?	? (0.28 ± 0.41)^c
Hokey Pokey eruptive period		?	2.2 ± 0.32	1(2.22 ± 0.3)^c	–	?	4?(2.22 ± 0.32)^c
Ōruanui Tephra (25.4 ka)			?	?	–	?	?
R10 lahars (27 ka)					–	?	?
R11 lahars (27 ka)					–	?	?
Okaia Tephra (28.6 ka)					–	?	?
Omataroa Tephra (28.2 ka)					–	?	?
Hauparu Tephra (36.1 ka)					–	?	?
R13 lahars (< 45 ka)					32 ± 4.5	2 ± 6^c	3 (2 ± 6)^c
R14 lahars (> 45 ka)					–	?	?
R15 lahars (< 133 ka)					34 ± 3	2 ± 4.24^c	2 (2 ± 4.24)^c
Wahianoa Formation (133 ka)					39 ± 4.5	5 ± 6.36	1 (5 ± 6.36^b)
Total	0.69 ± 0.2	6.33 ± 0.5	2.2 ± 0.3	> 2	175.51 ± 21.3		> 9 events
Fault observations	060-090/80SE	036-098/60-90NNW	020-080/69-85SSE	NE/SE	035-045/SE		NE/SE
Fault termination	Uncertain (< 11.77 ka)	Uncertain (< 11 ka)	Uncertain (< 17.7 ka)	Uncertain (< 11.77 ka)	Uncertain (< 11 ka)		Uncertain (< 11 ka)
Mean slip-rate (mm/yr)^d	0.06 ± 0.0 (11.77 ka)	0.11 ± 0.0 (17.7 ka)	0.09 ± 0.02 (23.6 ka)	0.15 ± 0.31 (> 45 ka)	< 0.3 ± 1.12 (133 ka)		> 0.15 ± 0.31 (133 ka)
Comments	Multiple-step fault	Antithetic faults of Tonga 24	Primary fault of Tonga 23	Good geomorphic expression

Net-slip, total displacement associated with each chronological maker; SED/PD, single event displacement/progressive displacement; N event, minimum number of rupture events. All values are in metres. All ages are cal ka BP. See Methods section for data accuracy. Vertical displacements and more detailed description of each fault can be seen in Table S2.

^aLast event known, uncertain timing. –, no value; ?, not sure; ___, fault termination.

^bMore than one event.

^cCould have happened after this time.

^dSlip rate based on net-slip.

Displacements on geomorphic surfaces show different offsets for different ages (Figure 9). Some measurements give a minimum offset, because the downthrown block is buried by younger deposits. We can account for seven progressive displacement increments for the geomorphic displacements measured along the fault (profiles in Table 3), suggesting that at least seven events occurred. If we combine the outcrop data with the geomorphic data, we add two more ruptures (Table 3). However, values of progressive displacement increments for geomorphic data are too large (2–12 m), suggesting that those values are related to several ruptures (see more on single event displacements below) and that the nine events analysed are a minimum for the last c. 130 kyr.

Discussion

Based on an analysis of the variation in slip-rate on each fault (Figure 10) and their paleoseismic histories, we are able to suggest the role of each fault in the kinematics of the Taupo Rift, the implications for seismic hazard and their possible association with volcanism.

Is the Upper Waikato Stream Fault part of the Rangipo Fault or part of the Wahianoa Fault?

The location of our study area is at the putative intersection between the Rangipo and the Wahianoa faults, hence we tried to assess whether the Upper Waikato Stream Fault (Figures 3 and 5) can be associated with one fault or the other. This has been evaluated by a comparison of fault geometry, event history and fault slip-rate.

Based on the fault geometry, we interpret the extension of the Upper Waikato Stream and suggest that it may have a physical linkage with the Rangipo and the Wahianoa faults (see potential fault linkage models in Figure 11). First, we look at the outcrops of sections 1 and 2 within the Upper Waikato Stream. Faults 10/4_2, 10/4_4, 10/4_6, 11/4_3, 11/4_4 and 11/4_5 in section 2 are the same fault, and equivalent to the Upper Waikato Stream Fault of section 1, because they have similar strike and throw sense (NW–W) (Table 1) and are in close proximity, and thus we interpret that the Upper Waikato Stream Fault extends at least from section 1 to point 11/4_5 (Figure 5). This suggests that the Upper Waikato Stream Fault is at least 15 km long if it is an independent fault (Figure 11A).

Second, and with respect to the linkage with the Rangipo Fault, the Stereonet diagrams in Figure 5 show that section 1 of the Upper Waikato Stream Fault, close to the Rangipo Fault, has a strike that varies from 000 to 030°, similar to the Rangipo Fault. Some of the fault strike measurements from section 2 are also similar to those of the Rangipo Fault and thus we consider that Rangipo Fault might extend northwards into section 2 of the Upper Waikato Stream Fault, and perhaps beyond into the Kaimanawa Fault (see further discussion below). This is represented as a fault linkage model in Figure 11C.

Finally, we look at the possible linkage of the Upper Waikato Stream Fault with the Wahianoa Fault. Faults 10/4_3, 10/4_5, 11/4_1 and 11/4_2 are downthrown to the S–SE. These may be antithetic faults to the Upper Waikato Stream, but also might be part of, or to some extent controlled by, the Wahianoa Fault (Figure 5 and Table 2). This might suggest that the Wahianoa Fault extends to point 11/4_2 in Figure 5, and thus a model in which the faults intersect or merge has been developed (Figure 11B). Section 2 has a greater strike range (000 to 070°E) that can be attributed to the influence of the Wahianoa Fault at that latitude (Figure 5). If we consider them as part of the Wahianoa Fault, they can confirm events 3 and 9 of section 3 (see Table 4).

Table 4. Comparison of fault rupture number and timing of the Upper Waikato Stream, Wahianoa and Rangipo faults.

Fault	Upper Waikato Stream Fault	Wahianoa Fault	Rangipo Fault Villamor et al. 2007
Section	1 and 2	2 and 3
		(Outcrops + profiles)
Offset (m)	N event (Total SED/PD; unknown %)	N event (Total SED/PD)	N event (Total SED/PD)
Taupo pumice (1.72 ka)	?	?	7 (0.1 m)
Mangatawai Tephra (3.52 ka)	12? (0.54 ± 0.2 to 0.86 ± 0.2; 37.3%)^a	?	6 (0.14 m)
Papakai formation (3–11 ka)	?	?	5 (0.8 m)
Poutu lapilli (11 ka)	11 (0.75 ± 0.2 to 1.2 ± 0.2; 37.3%%)^b	?	4 (0.35 m)
Poronui Tephra (11.2 ka)	?	9^c (6 ± 2.12 m)^a	?
Pahoka Tephra (11 ka)	?	8^c (5.5 ± 5.32 m)	?
Pourahu member (11.77 ka)	(2.37 ± 0.4 to 3.78 ± 0.4^c; 37.3%)^b	?	3 (1.2 m) > 25 km
Okupata (11.77 ka)	?	7? (0.69 ± 0.2–1.76 ± 0.3 m)^b	2 (> 0.4 m)
Shawcroft Tephra		(0.37 ± 0.31 m)	?
Waiohau Tephra (13.64 ka)		(0.32 ± 0.36 m)	1 (32 m)
Karioi eruptive period (15 ka)		6^c (12 ± 3.9)^b
Rerewhakaaitu Tephra (17.7 ka)		5^c (7 ± 6)
Last Glacial Maximum (>18 ka)		(0.28 ± 0.41 m)^b
Hokey Pokey eruptive period	10 (0.01 ± 0.2 to 0.17 ± 0.2; 99.8%)	4? (2.22 ± 0.32 m)^b
Ōruanui Tephra (25.4 ka)	9 (1.71 ± 0.4 to 2.2 ± 0.4; 36%)	?
R10 lahars (27 ka)	8 (1.92 ± 0.4 to 2.73 ± 0.4^c; 29.6%)	?
Marker unit 1 (olive tephra)	7 (0.65 ± 0.5 to 0.92 ± 0.5; 29.6%)	?
R11 lahars (27 ka)	6 (1.03 ± 0.7 to 2.76 ± 0.8^c; 0%)	?
Okaia Tephra (28.6 ka)	?	?
Omataroa Tephra (28.2 ka)	5 (0.47 ± 0.5 to 1.24 ± 0.5; 62.2%)^b	?
Orange lapilli	4 (0.46 ± 0.6 to 1.79 ± 0.6; 74.3%)	?
Elephant surge (marker unit 2)	3 (0.58 ± 0.7 to 1.21 ± 0.7; 51.94%)	?
Hauparu Tephra (36.1 ka)	2 (0.58 ± 0.5 to 5.64 ± 0.5^c; 89.7%)	?
R13 lahars (<45 ka)	?	3 (2 ± 6)^a
Marker unit 3	(1.44 ± 0.3 to 2.3 ± 0.3; 37.3%)^a	?
R14 lahars (>45 ka)	1 (1.1 ± 0.7 to 1.75 ± 0.7; 37.3%)	?
R15 lahars (<133 ka)	?	2 (2 ± 4.24)^a
Wahianoa Formation (133 ka)		1^c (5 ± 6.36)
Total	> 12 events	> 9 events	7 events
Fault observations	NE/NW	NE/SE	NNE/NW
Fault termination	Uncertain (< 3.52 ka)	Uncertain (< 11 ka)	Uncertain (< 1.72 ka)
Mean slip-rate (mm/yr)^d	0.5 ± 0.06 (45 ka)	>0.15 ± 0.31 (133 ka)	0.25–0.49 (13.64 ± 0.2 ka)

The dark and medium grey shaded areas show the possible timing for simultaneous rupturing of the Upper Waikato Stream and Wahianoa Faults, while the dark and light grey shaded areas show possible timing for simultaneous rupturing of the Upper Waikato Stream and Rangipo faults. SED/PD, single event displacement/progressive displacement; N event,  minimum number of rupture events; Unknown %, unknown percentage of total PD relative to the 100% displacement measured on R11 lahars. All values are in metres. All ages are cal ka BP. See Methods section for data accuracy. Vertical displacements and more detailed description of each fault can be seen in Tables S1 and S2.

^aLast event known, uncertain timing. –, no value; ?, not sure; ___, fault termination.

^bCould have happened after this time.

^cMore than one event.

^dSlip rate based on net-slip.

Based on earthquake history and timing of events, we observe that the Upper Waikato Stream Fault might have ruptured in conjunction with the Rangipo and Wahianoa faults (Table 4). Comparisons of the rupture history of the Upper Waikato Stream Fault with the Rangipo and Wahianoa faults for the last 11 kyr show that there were two events on the Rangipo (events 4 and 6) and Upper Waikato Stream (events 11 and 12) faults that might overlap in time; one c. 11 cal ka BP and another c. 3.5 cal ka BP. Possibly one event at c. 11 cal ka BP on the Wahianoa Fault (event 8 or 9) also overlaps with the Rangipo and Upper Waikato Stream faults rupture.

Based on a comparison of slip-rate variations in time (Figure 10), the Upper Waikato Stream Fault has a contrasting pattern to that of the Rangipo and Wahianoa faults. Assuming that the summed progressive displacement for R11 lahars represents all of the Upper Waikato Stream Fault movement, the mean maximum slip-rate over the last 45 ka is 0.63 ± 0.06 mm/yr. With only observed data, the minimum slip-rate is 0.3 ± 0.06 mm/yr (Table 4). From past to present on the Upper Waikato Stream Fault, the slip-rate has changed from 0.45 mm/yr to 1.5 mm/yr and to 0.26 mm/yr for the periods c. 45–36.1, c. 36.1–23.65 and 23.65 cal ka BP to the present, respectively (Figure 10).

Figure 10. Fault slip-rate variation from > 45 ka to present day for the Rangipo Fault (

) (Villamor et al. 2007), the Wahianoa Fault (based on profiles,

), and the Upper Waikato Stream Fault (

 minimum values and

 maximum values). Higher slip-rate periods are coincident with the Bullot Formation, the Pahoka–Mangamate eruptions (PM), the Orange lapilli and marker unit 2 (Elephant surge) from Mt Ruapehu, and the Ōruanui eruption from Taupo. Taupo pumice (TP) eruption is indicated at 1.72 cal ka BP (TP). Zero (0) indicates the present day.

The slip-rate on the Wahianoa Fault varied over time: from 45 to 15 cal ka BP it was < 0.1 mm/yr; from 15 to 10 cal ka BP it increased to 5 mm/yr; and from 10 cal ka BP to the present day it decreased to > 0.55 mm/yr (note that the most recent observed data are for 10 cal ka BP, and thus the age of the change in slip-rate and the slip-rate are minimum values). The slip-rate on the Rangipo Fault varied in a different fashion: from > 25.4 to 13.6 cal ka BP it was 1.8 mm/yr and from 13.6 cal ka BP to the present day it decreased to 0.24 mm/yr (Figure 10).

All faults seem to have a period of higher slip-rate, followed by a period of lower slip-rate into the present. However, these periods of higher slip-rate do not coincide with each other: the Rangipo Fault accelerates at 13.6 cal ka BP; the Upper Waikato Stream Fault at 23.65 cal ka BP, although there is a small increase at 11 ka; and the Wahianoa Fault at 10 cal ka BP (or earlier).

From 11 cal ka BP to the present day, all the faults have apparently experienced lower slip-rates, between 0.24 and 0.55 mm/yr, and during this period it was more likely that the Upper Waikato Stream Fault might have ruptured together with either the Rangipo Fault or the Wahianoa Fault. From 11 to 25 ka, the Rangipo Fault had an accelerated period that is not reflected on the Upper Waikato Stream Fault, suggesting that they might have moved independently (Figure 10).

In summary, the Upper Waikato Stream Fault is a newly discovered and potentially independent seismogenic source in the southern termination of the Taupo Rift, but a fault that could also rupture together with the Rangipo or the Wahianoa faults.

The role of the Wahianoa, Upper Waikato Stream and Rangipo faults in the kinematics and evolution of the Taupo Rift

Since rifting and extensive volcanism started in this area approximately > 340 ka ago (Gamble et al. 2003; Villamor & Berryman 2006), the Wahianoa and Rangipo faults formed, contemporaneously and oblique to each other, playing an important role in the kinematics and evolution of the southern termination of the Taupo Rift.

The orientation of the Upper Waikato Stream and Wahianoa faults were probably inherited from basement faults such as the Oligocene–Miocene Kaimanawa Fault (Figure 1; Kear 1993) seen in the Kaimanawa Mountains, east of the Mt Ruapehu Graben. In Figure 1 several NE- to ENE-striking lineaments, sub-parallel to the Kaimanawa, Wahianoa and Upper Waikato Stream faults can be observed in the exposed basement in the Kaimanawa Mountains. The strike of the Rangipo Fault does not seem to be represented in the basement lineaments in the Kaimanawa Mountains, but it is possible that the dominant NNE trend of the faults, just north of this area in the Tongariro domain, is forcing the NNE-strike into the Mt Ruapehu Graben area (Villamor & Berryman 2006b).

Villamor and Berryman (2006a) suggest that the co-existence of active normal faults with very oblique fault trends in the wider southern Taupo Rift region is likely to be explained by a stress tensor with vertical│σ1│>│σ2│≈│σ3│ that will cause radial extension (radial strain) and where, theoretically, any normal fault direction is possible. Townend et al. (2012) also found several orientations of the azimuth of the maximum horizontal compressive stress (SHmax) in their analysis of the shallow focal mechanism in the region of southern Taupo Rift. If radial stresses are affecting the crust in the Mt Ruapehu area and any fault direction is possible, but still we see a preferred NE orientation, it is possible that the NE and ENE fault trends that dominate in our study area are controlled by pre-existing weaknesses in the basement.

Our study shows that in the short term (few thousands of years), the Wahianoa, Rangipo and Upper Waikato Stream faults have been accommodating extension at the same time, but have taken dominant roles during different periods. The Upper Waikato Stream Fault was dominant from 36.1 to 23.65 cal ka BP (higher slip-rate period), whereas the Rangipo Fault was dominant from > 25 to < 15 cal ka BP, and the Wahianoa Fault was dominant from 15 to 10 cal ka BP (Figure 10). These changes in slip-rate over time might represent short periods (c. 5–10 ka) when one extension direction may predominate over others, whereas in the long term (e.g. since the inception of rifting at > 340 ka) the deformation pattern shows extension in different directions. It is also possible that slip variability is simply related to fault interactions (stress transfer among closely spaced faults) as observed globally for other extensional tectonic regimes (Weldon et al. 2004; Nicol et al. 2006; Mouslopoulou et al. 2009; Ferry et al. 2011; Cowie et al. 2012; Benedetti et al. 2013), or that slip-rate variability is a combination of both processes, fault interactions and short-term changes in the extension direction.

The prominent decrease in slip-rate of the Upper Waikato Stream Fault since c. 24 ka could be used as a sign of rift evolution at this latitude. The Upper Waikato Stream and Rangipo faults represent the east margin of the Taupo Rift, at or very near the major lithological boundary between volcanic rocks and basement greywacke. It is known from the volcanic/greywacke contact that the rift narrows to the north of this area (Villamor & Berryman 2006). The decrease in the slip-rate of the Upper Waikato Stream Fault, and the lack of geomorphic expression that suggests low current activity, might be a consequence of the Mt Ruapehu Graben margin shifting closer to the axis at the Tongariro Graben. However, it is possible that the decrease in slip-rate of the Upper Waikato Stream Fault is a consequence of fault interactions and transfer of motion onto other faults. If so, a rift narrowing at this latitude is not necessary.

Seismic hazard from the Upper Waikato Stream and Wahianoa faults

In this section, we evaluate the seismic hazard for the studied faults estimating single-event displacements, earthquake magnitudes and recurrence times for segments of different surface lengths.

Average single-event displacements are derived from fault length using fault-rupture scaling relations (see Methods section) and can be used to assess whether the progressive displacements calculated above represent one or multiple single-event displacements. Fault rupture length is also used to assess future earthquake magnitudes. Several scenarios were considered for rupture of the Upper Waikato Stream Fault (Figure 11). The Upper Waikato Stream Fault was mapped for only 5.5 km given the lack of geomorphic expression but we expect that, as an individual seismogenic source, it may be at least as long as the thickness of the seismogenic crust, that is c. 15 km (Figure 11A). We calculate that a combined rupture of the Wahianoa and Upper Waikato Stream faults will produce c. 27 km of rupture (Figure 11B). Rupture of the Upper Waikato Stream and Rangipo faults was suggested as a possibility by Villamor et al. (2007). They suggested the potential rupture of the entire Rangipo Fault, including the section between the mapped traces and the Kaimanawa Fault, a section that includes the newly defined Upper Waikato Stream Fault (Figure 11C). Our observations support a potential fault rupture of up to 43 km.

Figure 11. Potential physical intersections and rupture models of the Upper Waikato Stream Fault with Rangipo and Wahianoa faults. A, Independent ruptures. B, Potential rupture of the Wahianoa (WF) and Upper Waikato Stream (UWS F) faults (27 km). C, Potential rupture of the Rangipo (RF), Upper Waikato Stream and Kaimanawa (Kai F) faults (43 km).

For a 27 km long surface rupture (Upper Waikato Stream Fault along with Wahianoa Fault) we obtain a M_W6.8 earthquake with an average single-event displacement of 1.26 m (Table 5). For a 43 km long surface rupture (Upper Waikato Stream Fault along with Rangipo Fault), we obtain a M_W7.1 earthquake and a single-event displacement of 2.4 m. These single-event displacement values represent a total-fault rupture. Lower single-event displacement values (< 1 m) suggest a segmented fault rupture.

Table 5. Moment magnitude and single-event displacement for different fault lengths.

Fault	Segment surface length (km)	Average displacement (m)	MW
UWS^a	15	0.8	6.5
Wahianoa + UWS	27	1.5	6.8
Rangipo + UWS + Kaimanawa	43	2.4	7.1

Estimated average single-event displacement in metres and earthquake moment magnitude (M_W) based on Webb’s equation, for a certain segment surface length (km) and fault area (as described in Villamor et al. 2007; Stirling et al. 2012, see Methods section).

^aNote that the Upper Waikato Stream (UWS) Fault individual rupture length is a minimum value.

Based on the estimated average displacements, the maximum progressive displacements for events 1, 3, 4, 5, 9 and 11 on the Upper Waikato Stream Fault could represent a whole-fault rupture, whereas minimum progressive displacements for all the events represent segmented ruptures. Greater values (e.g. 2.73 m on event 8) represent multiple rather than single-event displacements (Tables 1 and 4).

Based on the combined findings, we estimate the minimum recurrence interval for various periods (Table 4). For the Wahianoa Fault, the recurrence interval is 2.4 ka between 25.4 and 11.2 cal ka BP, and 13.6 ka between 133 and 11 cal ka BP (based on sections 2 and 3). For the Upper Waikato Stream Fault, the recurrence interval is 1.6 ka between 36.1 and 21.9 cal ka BP (section 1) and 3.5 ka between 45 and 3.52 cal ka BP (sections 1 and 2).

Earlier studies on the Wahianoa Fault (Villamor & Berryman 2006) estimated mean slip-rates of 0.3 mm/yr. Our data show that the Wahianoa Fault could have a minimum slip-rate of 0.55 mm/yr from 10 ka to the present day, changing the current seismic hazard potential. Consequently, our study provides an update to the National Seismic Hazard Model for New Zealand (Stirling et al. 2012; Litchfield et al. 2014). Our data reinforce the segment surface lengths given to the Rangipo (up to 43 km; Villamor et al. 2007) and Wahianoa (27 km; Stirling et al. 2012) faults.

Potential association of volcanism with periods of accelerated seismic activity

Fault slip-rate variations appear to be associated with times of enhanced volcanism in the TgVC (Villamor et al. 2007). Higher slip-rate periods in the Upper Waikato Stream Fault seem to be coincident with the very large Ōruanui eruption from the Taupo caldera immediately to the north (25.4 ± 0.2 cal ka BP; vent located 50 km from the fault) at the same time that large extension fissures formed in close proximity to the Upper Waikato Stream Fault (Figure 7). This eruption was one of the largest known over the last c. 40 ka, producing 530 km³ of erupted magma (Wilson 2001). An acceleration of the Upper Waikato Stream Fault movement may also coincide with the eruption of the Orange lapilli (Mt Ruapehu; vent located 15 km from the fault) and marker unit 2 (probably Mt Tongariro) at c. 32 ka, as well as with the major Pahoka–Mangamate eruption sequence at c. 11 cal ka BP (vent located 18 km from the fault), an eruptive process during regional extension in the TgVC (Topping 1973; Nairn et al. 1998). In the Rangipo Fault, a higher slip-rate period coincides with the eruptions of the Bullot Formation, a period of frequent major plinian eruptions of Mt Ruapehu from 25 to 11 cal ka BP (Donoghue et al. 1991; Pardo et al. 2012) (vent located 13 km from the Rangipo Fault). In the Wahianoa Fault, a higher slip-rate period coincides with the Pahoka–Mangamate eruption sequence at c. 11 cal ka BP (vent located 18 km from the Wahianoa Fault) (Figure 10).

During these periods of large and explosive volcanic eruptions, the extension across this portion of the rift may be mainly accommodated by dike intrusion, and crustal stresses created by the volcanic intrusions could be influencing the slip-rate changes for some of the faults. The temporal association of Bullot Formation (Mt Ruapehu) and Pahoka–Mangamate (Ngāuruhoe, Tongariro) eruptions with periods of increased activity on the southern faults (Rangipo and Wahianoa) implies a connection between volcanic and tectonic activity, but the faults did not react in the same way. The Rangipo Fault may have been more connected to the Bullot Formation eruptions from Ruapehu Volcano and the Wahianoa Fault to the Pahoka–Mangamate eruptions from Tongariro and Ngāuruhoe volcanoes (Figure 10).

The distance and magnitude of volcanism (e.g. dike opening), together with fault orientations, determine whether the fault will be affected by dike intrusion. The Upper Waikato Stream Fault was strongly affected by the Ōruanui eruption, presenting a higher slip-rate period and extensional fissures. This suggests that Ōruanui eruption was significant enough to alter the crustal stress at distances of > 55 km from the caldera, and perhaps change the kinematics of the rift by allowing the activity to shift from the margin of the rift towards the axis.

Conclusions

This study provides an updated understanding of the eastern boundary of the Taupo Rift close to its termination, extending the mapping of active faults in this area and adding more information on fault characterisation for hazard assessment and suggesting potential volcano–tectonic interactions. The NE-striking Upper Waikato Stream Fault is a newly discovered and potential independent seismogenic earthquake source at the eastern boundary of the rift just north of the NNE-striking Rangipo Fault. Geological evidence suggests that the Upper Waikato Stream Fault might be an independent fault, but might also rupture in conjunction with the Wahianoa and/or Rangipo Fault. The EWE-striking Wahianoa, a known active fault, might be potentially linked with the Upper Waikato Stream Fault, forming a fault intersection pattern similar to those described farther south in the rift.

Paleoseismic and geomorphic analyses demonstrate the occurrence of at least 12 surface-rupturing earthquakes in the last 45 ka on the Upper Waikato Stream Fault, which has a mean slip-rate of 0.3 ± 0.06 to 0.63 ± 0.06 mm/yr. New data on the Wahianoa Fault suggest at least nine surface-rupturing earthquakes in the last 133 ka, and we calculate a long-term mean slip-rate of 0.15 ± 0.31 mm/yr for this fault.

Our study provides new data to update the National Seismic Hazard Model (Stirling et al. 2012; Litchfield et al. 2014) for this region. The new seismic source been defined here, the Upper Waikato Stream Fault (M_W6.5), can be added to the National Seismic Hazard Model. Also, the Wahianoa Fault has a slip-rate of > 0.55 mm/yr from 10 cal ka BP to the present day, faster than in the current National Seismic Hazard Model. The rupture lengths for the Wahianoa and Rangipo faults were confirmed, the Wahianoa Fault might rupture together with the Upper Waikato Stream Fault (27 km), and the Rangipo Fault might rupture together with the Upper Waikato Stream and Kaimanawa faults, giving a total surface-rupture length of 43 km.

The Rangipo Fault might have ruptured at the same time as the Upper Waikato Stream Fault at c. 11 and 3.52 cal ka BP (Table 4), posing a greater potential seismic hazard. For a 27 km long rupture of the Wahianoa Fault together with the Upper Waikato Stream Fault, the expected earthquake would be M_W6.8; and for a 43 km long rupture, involving the Rangipo Fault with the Upper Waikato Stream and the Kaimanawa faults, we would expect a M_W7.1 earthquake.

The slip-rate variability of the Upper Waikato Stream and Wahianoa faults estimated here has been compared with existing data on the Rangipo Fault. Tentative explanations of the slip-rate changes on these three faults are: (1) stress transfer interactions among closely space faults, similar to other observations for the Taupo Rift farther north; (2) short-term (few thousands of years) changes in extension direction, with the Wahianoa, Rangipo and Upper Waikato Stream faults taking dominant roles in different periods; (3) higher slip-rate periods coinciding with periods of major explosive volcanism (e.g. the Ōruanui eruption from Taupo caldera at 25.4 ± 0.2 cal ka BP coinciding with higher slip-rates on the Upper Waikato Stream Fault, the Bullot Formation from Mt Ruapehu at 25–11 ka with the Rangipo Fault, and the Pahoka–Mangamate eruptions from Tongariro at 11 cal ka BP with the Wahianoa Fault);  (4) rift evolution with slip-rate decrease at the rift boundary suggesting shifting of active faulting closer to the rift axis (as has been described farther north); and (5) a combination of all or some of the process above.

Acknowledgements

We wish to thank Kate Arentsen, Anja Moebis and the Volcanic Risk Solutions group in Massey University for their support. We acknowledge the Department of Conservation staff of Tongariro National Park. We also thank the Waiouru Military Camp for providing access to the restricted zones. We thank the reviewers (Dr Vasso Mouslopoulou and Associate Professor Julie Rowland) and editor of the NZJGG (Dr Nick Mortimer) for their comments that greatly improved the presentation of our data.

Associate Editor: Dr Nick Mortimer.

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCiD

SJ Cronin http://orcid.org/0000-0001-7499-603X

G Kereszturi http://orcid.org/0000-0003-4336-2012

Additional information

Funding

Funding was provided by a Massey University Doctoral Scholarship [Grant number 13162875], the Living with Volcanic Risk programme of the NZ Natural Hazards Research Platform (SJC, JP) and GNS Science CORE Funds.