Active faults, paleoseismology and historical fault rupture in northern Wairarapa, North
Island, New Zealand
E.R. Schermer (Western Washington University, Bellingham, WA 98225 USA); R. Van
Dissen and K.R. Berryman (Institute of Geological & Nuclear Sciences, Lower Hutt, New
Zealand), H.M. Kelsey, S. M. Cashman, (Dept. Of Geology, Humboldt State University, Arcata,
CA 95521 USA)
Abstract
Active faulting in the upper plate of the Hikurangi subduction zone, North Island, New
Zealand, represents a significant seismic hazard that is not yet well understood. In northern
Wairarapa, the geometry and kinematics of active faults, and the Quaternary and historical
surface-rupture record, have not previously been studied in detail. We present the results of
mapping and paleoseismicity studies on faults in the northern Wairarapa region to document the
characteristics of active faults and the timing of earthquakes. We focus on evidence for surface
rupture in the 1855 Wairarapa (Mw8.2) and 1934 Pahiatua (Mw7.4) earthquakes, two of New
Zealand’s largest historical earthquakes.
The Dreyers Rock, Alfredton, Saunders Road, Waitawhiti, and Waipukaka Faults form a
northeast-trending, east-stepping array of faults. Detailed mapping of offset geomorphic features
shows the rupture lengths vary from c. 7 to 20 km and single-event displacements range from 3
to 7 m, suggesting the faults are capable of generating M>7 earthquakes. Trenching results show
that two earthquakes have occurred on the Alfredton Fault since c.2900 cal BP. The most recent
event probably occurred during the 1855 Wairarapa earthquake as slip propagated northward
from the Wairarapa Fault and across a 6 km wide step. Waipukaka Fault trenches show that at
least three surface-rupturing earthquakes have occurred since 8290-7880 cal BP. Analysis of
stratigraphic and historical evidence suggests the most recent rupture occurred during the 1934
Pahiatua earthquake.
1
Estimates of slip rates provided by these data suggest that a larger component of strike slip
than previously suspected is occurring within the upper plate and that the faults accommodate a
significant proportion of the dextral component of oblique subduction. Assessment of seismic
hazard is difficult because the known fault scarp lengths appear too short to have accommodated
the estimated single-event displacements. Faults in the region are highly segmented,
disconnected, and probably structurally immature, which implies that apparent geometric
discontinuities at the surface may not be significant barriers to rupture propagation at depth and
that the surface rupture record significantly under-represents the seismic slip on faults in the
region.
Keywords: active faulting, paleoseismology, Hikurangi margin, oblique subduction, seismic
hazard
INTRODUCTION
In the North Island of New Zealand, oblique subduction of the Pacific plate beneath the
Australian plate is in part accommodated by upper-plate strike-slip and reverse faulting that
represents a significant seismic hazard, as indicated by seismologic and geodetic data (Walcott
1978; 1987; Darby & Meertens 1995; Webb & Anderson 1998), fault slip rates (Van Dissen &
Berryman 1996; Barnes & Mercier de Lépinay 1997; Beanland 1995; Beanland & Haines 1998)
and a series of large (M>6.5) historical earthquakes (Webb & Anderson 1998) (Fig. 1). These
data suggest an increase in coupling of the upper and lower plates from north to south, and an
accompanying decrease in the importance of megathrust events on the subduction interface
versus upper plate fault events (Walcott 1987; Beanland 1995; Reyners 1998; Darby & Beavan
2001). As yet, the slip rates and deformation pattern are poorly constrained except in the
southernmost part of the North Island. Understanding both the nature of the plate interaction and
the seismic hazard requires accurate data on the pattern of faulting, slip rates, and earthquake
history. Recent progress in understanding the seismology and surface deformation in historical
2
earthquakes on the North Island (Webb & Anderson 1998; McGinty et al. 2002; Doser & Webb
2003) can be augmented by paleoseismic investigations of the causative faults, especially
because no large surface-rupture earthquakes have occurred in the last several decades.
Furthermore, in northern Wairarapa (Fig. 1), limited mapping and analysis of active faults and
relatively late settlement and clearing of the area have hampered understanding of active faulting
and the relationship of fault scarps to historical events.
We present field and paleoseismologic evidence from northern Wairarapa that documents
the location, geometry, and kinematics of major active faults. We assess the likelihood of
historical surface rupture on the faults and describe the rupture parameters of the major surfacerupture earthquakes in the region. Questions we address include: how much upper plate strain is
represented by the faults in this region; what are the surface-rupture characteristics of a newly
developing zone of faults; and are established techniques to estimate magnitude from surface
fault geometry (e.g., Wells & Coppersmith 1994) valid in this tectonic setting. Quantification of
the geometry, kinematics, and slip rates of active faults will also lead to better understanding of
active strain partitioning in the Australian plate (e.g., Cashman et al. 1992; Beanland & Haines
1998), and will better constrain studies of the possible triggering relationships among forearc
earthquakes (e.g., Downes & McGinty 2001, McGinty et al. 2002).
Our study area includes the rupture zones of two of New Zealand’s largest earthquakes, the
1855 Wairarapa (Mw 8.2) and 1934 Pahiatua (Ms 7.6) earthquakes. Although surface faulting
was well documented for the 1855 earthquake (Grapes & Wellman 1988), the northern end of the
rupture was not established. The magnitude and relatively shallow depth (12±5 km, Doser &
Webb 2003) of the 1934 earthquake (Table 1) suggests a high likelihood of surface rupture, but
none was reported at the time. We focused our study of active faults in the area within the
highest (MM9) isoseismal of the 1934 earthquake (Downes et al. 1999) (Fig. 2). This area also
lies within the inner isoseismals of several other earthquakes (including 1904, 1931, and 1942;
Downes 1995), so the possibility for multiple historical ruptures can be considered. Lensen
(1969), Berryman & Cowan (1993) and Beanland (1995) documented active fault scarps in this
3
area, but little detailed work had been done and the age of recent ruptures on the faults were
unknown. We conducted an airphoto survey and compiled a map of new and existing
observations (e.g., Lensen 1969; Neef 1974, 1984; 1992; 1997a,b; Beanland 1995; Kelsey et al.
1995) to identify known and likely active faults (Fig. 2). Details of the location and
geomorphology of faults in the area can be found in Schermer et al. (1998). We examined the
earliest airphotos available, taken in 1943 and 1944, to assess fault scarps formed prior to that
time (Fig. 3). Field work examined several of the more prominent fault scarps, with geomorphic
and structural mapping designed to quantify the length, type of faulting, and offset
characteristics, and to assess the relative recency of faulting by a general examination of scarp
morphology and continuity. Two faults, the Alfredton Fault and the Waipukaka Fault (Figs. 2,
3), were identified as having extremely fresh scarps and these were chosen for detailed mapping,
trenching and 14C dating. Trenches were logged with a 1 m2 grid and lithologic units, soil
properties, and faults were identified. We used European settlement data and the field notes of
geologist M. Ongley to constrain dates more recent than obtainable by radiocarbon (14C) dating.
Estimates of rupture lengths, single-event slip, and slip rate of the faults are used to discuss the
implications for seismic hazard assessment in New Zealand and active tectonics of the Hikurangi
margin.
Geological and geodynamic setting of the North Island
The Hikurangi margin is the southern part of the Tonga-Kermadec subduction zone, where
the Pacific plate is being subducted beneath the Australian plate (Fig. 1). The area presently
exposed in the forearc region of the North Island and south of the present-day extent of the
volcanic arc has evolved throughout the Cenozoic from predominantly subsidence and tectonic
quiescence during the Miocene to uplift and deformation in Quaternary time. Numerous studies
have focused on the stratigraphic and structural evolution of the margin during the Tertiary, and
have shown that the area comprised a broad submarine forearc basin with an accretionary prism
to the east from early Miocene onward (e.g. Lewis 1980; Pettinga 1982; (Rait et al. 1991; Kelsey
4
et al. 1995; Delteil et al. 1996; Neef 1997a, 1997b). Emergence of the forearc did not occur until
latest Neogene or Quaternary (Lewis 1980; Pettinga 1982).
From north the south along the Hikurangi margin, the style and rate of deformation changes,
but in much of the onshore area east and south of the volcanic arc there is a combination of
reverse faulting and dextral strike-slip faulting, a region termed the North Island Dextral Fault
Belt (NIDFB; Fig 1, Walcott 1987; (Cashman et al. 1992; Beanland 1995). Geologically
determined dextral slip rate on the main faults in the North Island Dextral Fault Belt increases
from c.2 mm/yr in the north to c.21 mm/yr in the south (Fig 1, Beanland 1995; Van Dissen &
Berryman 1996, Barnes et al. 1998). In much of the NIDFB, the slip rate on faults is not well
constrained, particularly in the study area (Fig. 2), where the integrated dextral slip rate is
estimated at c.8 mm/yr (Beanland 1995). Dextral faulting began in the southern Hikurangi
margin within the last c.0.5 m.y., evolving from an earlier, predominantly reverse faulting
environment (Beanland 1995; Kelsey et al. 1995; Beanland et al. 1998). Although the late
Neogene reverse faulting extended from the Wellington Fault to the present east coast, many of
the reverse faults are now inactive, and the strike-slip faults in general reactivated only some of
the older structures (Kelsey et al. 1995). Major faults of the North Island Dextral Fault Belt
comprise two main active dextral fault zones, the Wellington-Mohaka-Ruahine Fault system in
the west, and the Wairarapa-Alfredton-Makuri-Poukawa Fault system in the east (Fig. 1). The
eastern part of the belt is characterised by a wide zone of discontinuous fault traces, indicative of
a young fault system whose traces have not yet integrated into a throughgoing major fault. The
distributed faulting makes it difficult to determine the seismic hazard without full
characterization of the fault system..
Historical earthquakes
The southern North Island has been the site of several major historical earthquakes,
including three of the four largest earthquakes in the recorded history of New Zealand (Fig. 1,
Table 1). Throughout this work "historical" events are considered as occurring after 1840, the
5
time of earliest European settlement of the North Island. Data on surface rupture for the 1855
Wairarapa (Mw 8.2) and 1931 Hawke’s Bay (Ms 7.8) earthquakes have been compiled
(Henderson 1933; Grapes & Wellman 1988; Hull 1990), but before this study there was no
evidence for surface faulting associated with other earthquakes, including the 1934 Pahiatua (Ms
7.6) earthquake, New Zealand's fourth-largest historical earthquake. Study immediately
following the 1942 (Ms 7.2,5.3,7.0) earthquake series suggested a short rupture (Ongley 1943b),
which has been reinterpreted as a landslide related to strong ground shaking (Lensen 1969;
Schermer et al. 1998; Downes et al. 2001). In this paper we present evidence for surface rupture
in the 1934 earthquake and document the northern extent of the 1855 earthquake.
GEOMORPHOLOGY, STRUCTURAL GEOLOGY,
AND PALEOSEISMOLOGY OF THE MAJOR FAULT ZONES
The Wairarapa fault branches northward to the Pa Valley-Makuri fault zone in the west, and
the Dreyers Rock Fault-Alfredton Fault in the east (Figs. 1, 2). The most prominent fault scarps
occur on the eastern branch, and the Alfredton Fault appears to be part of a right-stepping en
echelon fault zone that includes the Saunders Road Fault and the Waipukaka fault. Other
prominent faults in the region include the Waitawhiti Fault, the East Puketoi Faults, and the
Weber Fault (Fig. 2). The major fault zones will be described from southwest to northeast.
General Geomorphology
Scarps in the study area are underlain by bedrock siltstone, sandstone, or mudstone overlain
by a thin (<0.5 m) soil. Stream terraces are not well developed, and few of the mapped fault
traces cross major terraces. The abundance of landslides recognised from airphoto surveys
suggests rapid change of hillslope morphology, such that nearly all preserved fault scarps
probably are postglacial (<c.14 ka; Pillans et al. 1993) even where the age of the surface is not
known precisely. Hillslope erosion has likely increased significantly following deforestation
during settlement of the area, and the preservation of scarps is in part related to the age, method,
and intensity of clearing and cultivation. As a qualitative estimate of relative age of faulting, we
6
divide the scarps into classes similar to those used by Lensen (1969), recognising that our age
estimates could be in error by at least a factor of 2-3 (e.g., scarps interpreted to be <10,000 years
old are probably not much more than 20,000 years old if they cut post-glacial features such as
stream terraces). The faults shown as bold, unbroken lines (Figs. 2, 4, 8) typically have wellpreserved scarps with moderately steep faces (>25-30°) and continuity across steep, landslideprone slopes (scarps we term “fresh”, e.g., Fig. 3), but locally they have subdued scarps and
somewhat more erosion by streams and landslides. We interpret these scarps to have been active
within the last c.1000-10,000 yr; these would fall into Lensen's Class I (repeated movement in
the last 500,000 yr or single movement in the last 5000 yr). The faults shown as dashed lines are
highly eroded or appear as traces on elevated or dissected Quaternary surfaces; these would fall
into Lensen's Class II (single movement in last 5000-50,000 yr and/or repeated movement in
50,000-500,000 yr).
Alfredton and Dreyers Rock fault zones
Geology and geomorphology
Deformation in the Alfredton region (Figs. 2, 4) changed from reverse faulting to dextral
strike-slip faulting before late Pleistocene time (Kelsey et al. 1995). The Alfredton and Pa
Valley Faults are in part reactivated structures and in part new traces that have stepped east or
west from the older trace. Faults are interpreted to be subvertical based on straight traces across
topography; dips at the surface are both southeast and northwest (Kelsey et al. 1995). Lamarche
et al. (1995) inferred from seismic reflection profiles near Ihuraua (Fig. 4) that the Alfredton
Fault is composed of several strands that converge at depth. At the southern end of the Alfredton
Fault the Dreyers Rock Fault zone connects the Alfredton Fault with the Wairarapa Fault (Fig.
2), and consists of several short north-, east- and northeast-striking faults with dextral-normal
slip (Kelsey et al. 1995). Based on bedrock mapping, Kelsey et al., (1995) showed that the north
end of the Alfredton Fault merges westward with the Pa Valley Fault, and from there northward
with the Makuri Fault near Makuri (Figs. 2, 4). However, as shown in Kelsey et al. (1995, their
7
Figure 6) and in Figures 2 and 4, the Holocene trace of the fault may instead step northeast to
connect with the Saunders Road Fault. Cumulative Quaternary strike slip on Pa Valley and
Alfredton Faults is less than a few hundred metres (Kelsey et al. 1995). Lensen (1969) mapped
several locations along the southern Alfredton Fault where Quaternary features are offset
dextrally (Fig. 4), however none of these features are of known age. Lensen (1969) noted that
postglacial features are offset an average of 40 m and that historical displacement was likely
along part of the fault because the scarps are so fresh. Measurements of offset geomorphic
features compiled from Lensen (1969), Beanland (1995), and this study are shown in Figure 5A.
The data cluster at offset values of c. 4-7 m, 8-12 m, and 18-19 m, suggesting a single-event
displacement of 4-7 m. The juxtaposition of different basement units suggests significant longterm cumulative vertical displacement (c.1 km, Kelsey et al. 1995). Some of this vertical
displacement comes from reverse slip of an earlier phase of faulting and the remainder comes
from the vertical component of oblique slip from the later, contemporary phase of faulting
(Kelsey et al. 1995).
From Ihuraua northward, the Alfredton Fault strikes c.040° and is composed of two late
Quaternary traces in the south and one in the north (Figs. 3A, 4). In several locations along its
length the Alfredton Fault consists of two or more subparallel faults; locally these features form
graben, in other places the upthrown side changes along strike of one or both faults (Figs. 3A,
3B, 4). The western strand typically has a larger dextral offset than the eastern strands. The
eastern strands are subparallel to bedding and may represent flexural slip faults related to folding
in the footwall of the main Alfredton Fault (Fig. 3A).
Based on similar morphology and continuity of structure, we estimate the length of surface
rupture in the most recent earthquake. Scarps in the Dreyers Rock Fault zone appear similar to
those in Alfredton Fault zone, making the total length of relatively continuous distinct scarps 1720 km (Figs. 2, 4). The Dreyers Rock Fault Zone consists of c.4 km of fresh normal and dextralnormal fault scarps (Kelsey et al. 1995). Dextral fault scarps in the c.4 km long segment of the
Alfredton fault from Ihuraua northward to Bartons Line (Figs. 3A, 4) have face angles typically
8
>40° and locally subvertical. There is a segment c.1 km long to the northeast of Barton’s line
where the scarp is obscured, and another c.8 km of relatively fresh scarp farther to the northeast
that cuts stream terraces of probable Holocene age (c.7 ka, McCallion 1996) (Fig. 3B). Within
this segment the fault bends westward and accommodates an oblique reverse component in the
restraining bend (McCallion 1996). North of the Tiraumea River (Figs. 2, 4), the scarp becomes
increasingly difficult to identify although the fault can be recognised by the juxtaposition of
different Tertiary bedrock units, and is mapped as merging with the Pa Valley and Makuri faults
near Makuri (Kelsey et al. 1995) (Fig. 2).
The good preservation, steep scarp faces, and apparently young age of offset features along
the Alfredton and Dreyers Rock Faults can be interpreted to reflect recent, possibly historical,
surface rupture based on geomorphologic estimates alone. Photographs of scarps formed during
the 1855 earthquake on the Wairarapa Fault (Ongley 1943a; Grapes & Wellman 1988) show
similar morphologic characteristics.
The entire length of the fault zone from the southern end of the Dreyers Rock Fault Zone
near Mauriceville to the northernmost mapped trace of the Alfredton fault near Makuri, is 30 km
(Fig. 2), which would represent a maximum rupture length for an earthquake that was confined
to the Alfredton and Dreyers Rock faults. The fault scarp is very difficult to find between the
northernmost scarp and where it crosses the Tiraumea River, however, and the northernmost
scarp appears to have been freshened by recent landslides. The scarp is also obscured for several
kilometres near the southern end near Mauriceville (Fig. 2); thus we believe it is unlikely that the
most recent surface rupture exceeded c.20 km in length. We infer a minimum rupture length of
17 km during most recent earthquake based on similar scarp morphology. North of the Tiraumea
River near Nikatea (Fig. 2), Lensen (1969) mapped and interpreted a fresh trace c.500 m long as
a possible historic rupture; if this trace also ruptured in the most recent earthquake, it would
extend the rupture length to 25 km and to within c.3 km of the Pa Valley Fault.
9
Paleoseismic investigation
Trenching at two sites along the Alfredton Fault reveals evidence for the age of
displacement during the last two earthquakes. Two trenches were excavated in 1991 along the
fault near Alfredton (golf course trenches, Figs. 2, 3B), and a third trench was excavated in 1998
along a southern section of the fault west of Bartons Line (Percy trench, Figs. 2, 3A).
Golf Course Trench sites: The golf course trenches were dug c.120 m apart, perpendicular
to the fault across an uphill-facing scarp segment where two laterally offset channels of 12-14
and 8.5-13 m occur, and ponding of the drainage has formed swampy areas along the hillside.
The scarp trends 040° and has a fairly constant height of c.1.5 m, northwest side up, where not
obscured by landslides or artificial fill (Fig. 3B).
Trench ALF-1 (Fig. 6) contains a single peaty mud layer 10-20 cm thick deposited above
mottled grey silt which in turn lies above weathered yellow-brown siltstone interpreted as
bedrock. In trench ALF-2 (Fig. 6) a peat horizon of similar thickness lies above silt interpreted
either as weathered bedrock or alluvial silt. The peat layer in both trenches is deformed adjacent
to the fault. Both trenches have a mixed unit of sheared clay, silt, and mud along the fault, and in
ALF-1, a sliver of peat occurs along the fault plane, and a piece of wood (sample 1/3) occurs at
the base of the mixed unit between the peat and the bedrock. The fault plane in both trenches
dips southeast, and has a normal component of slip. Above the peat layer is a thin silt horizon,
followed by topsoil that is seen to overlap the fault in ALF-2.
The two trenches can be interpreted to record the last two surface-rupturing events on the
Alfredton Fault: (1) the initial deposit of the peat records ponding of the channel at the scarp by
the earliest recorded earthquake, and (2) an event that deformed the peat. The most recent
rupture probably postdates the youngest peat date (<200 BP, Table 2), but this date could be
contaminated by modern roots. The penultimate rupture occurred after 2960-2780 cal BP which
is the age of wood caught in the fault plane, and before the oldest peat age of 510-690 cal BP
(Table 2). The five peat samples yield a fairly small range of dates <200-690 cal B.P.,
suggesting that the ages, although minima, reflect peat formation not much prior to 690 cal B.P.
10
If this interpretation is correct, the wood caught in the fault would be considerably older than any
event (or events) that incorporated it into a fault sliver.
The trenches record two earthquakes and the channels are offset by 8.5-14 m (Fig. 3B), so
the single-event horizontal slip on this segment of the Alfredton Fault ranges from 4.25-7 m.
This value is consistent with the smallest lateral displacement observed at a number of sites
along the fault trace, and is consistent with the cluster of observations at 4-7 m shown in Fig. 5A.
Percy trench: A third trench was excavated c.5 km south of the golf course site across a
steep section of a 4 m-high scarp in a ponded area on the western strand of the Alfredton Fault,
where the fault consists of two active traces (Figs. 2, 3A). The western trace contains evidence
for dextrally offset features whereas the eastern strand appears to have mainly a dip-slip
component and was interpreted by Neef (1976) as a bedding-plane parallel fault. Lensen (1969)
observed a terrace riser offset dextrally 4.6 m a few metres north of the trench site, but this riser
has been destroyed by road building. A larger stream offset (28-39m; reported by Lensen (1969)
as 32 m) occurs just south of the trench site (Figs. 3A, 4).
The most important features of the trench log (Fig. 7) are the presence of different bedrock
units in the hangingwall and footwall and the presence of two gravelly silt units (units 7 and 8),
interpreted as colluvial wedges, deposited above massive silt on the footwall. Unit 7 is not cut
by faults and part of unit 8 is cut by fault A. There are two layers of concentrated charcoal, with
the younger layer at the base of unit 12 and an older layer along the top of unit 9 and the base of
unit 13. Two buried soils (developed on units 9 and 11) are present in the footwall. Three faults
(A, B, and C) were observed; fault A cuts nearly up into the topsoil, which itself is fractured.
Fault B appears to be overlapped by the unit 8 colluvial wedge but the presence of disrupted soil
blocks at the apparent base of the wedge makes it difficult to see the upward extent of this fault.
Fault C is confined to bedrock in the hangingwall of the fault.
At least two surface-rupture earthquakes are interpreted at the Percy trench. The most
recent event is represented by movement along fault A and deposition of colluvial wedge unit 7,
11
which appears to be ongoing as remobilisation of similar material in unit 2. Colluvial wedge unit
8 may have formed during the same earthquake as unit 7, or possibly an earlier one. Unit 8
overlaps fault B and is partially cut by fault A; it is possible that this unit was deposited and
collapsed during movement on fault B, and then was cut through by fault A as slip propagated
upward in the same event. Dating of the lower charcoal layer and material within wedge unit 8
would then suggest this event occurred after 540-310 cal B.P (sample 1/3) and possibly after
470-0 cal BP (sample 1/6; Fig. 7). The eastern thin edge of wedge unit 8 overlaps a charcoal
horizon, but this part of the wedge likely represents somewhat later reworking of the wedge
material and the charcoal ages (samples 1/2, 1/6) do not directly constrain the event age. The
oldest event in the trench is represented by the wedge-shaped silts of units 9-10 burying the soil
of unit 11, but no dateable material was recovered from unit 11.
It is also possible that faults A and B slipped during two separate events; thus the trench
would contain evidence for three events. In this interpretation, fault A would postdate unit 8 but
slip decreased upward on that fault during the earthquake that formed unit 7. The earlier event
would be constrained to ≥540-310 cal BP (sample 1/3). There is no constraint on the later event
except that it overlaps the charcoal horizon of sample 1/6 and thus postdates 470-0 cal BP.
Geologic and historical evidence constrain the last earthquake to have occurred since the
bush was cleared, possibly less than 200 yr ago. Radiocarbon ages also permit this
interpretation. The charcoal horizons probably represent land clearing by both Maori and
Europeans, as burning was the common method of clearing bush, although natural fires are also a
possibility. The history of the Alfredton area and northern Wairarapa (Bagnall 1976; Edmonds
1987) suggests that there were Maori present in the area around Alfredton township before the
arrival of the first European settlers in the 1860s, but how long they had been there is unknown.
The first European settlers in the study area west of Bartons line were the Percys, ancestors of
the current landowners, who report having to clear the bush in the area in the late 1870s-80s
(Edmonds 1987), but do not mention specific localities that might have already been cleared. A
small hand-dug trench c.100 m north of the Percy trench along the Alfredton Fault showed the
12
fault cuts a prominent charcoal horizon, and is consistent with the evidence from the larger
trench that the last earthquake postdates land clearing. The 14C dates on samples 1/2 and 1/6
from the lower charcoal layer (Fig. 7, Table 2) suggest that there was some prehistoric burning at
the site, and put an upper limit on the timing of the last earthquake at c.AD 1450.
Alfredton fault slip rate
Calculation of a slip rate on the Alfredton Fault is important for consideration of regional
strain because it is the largest fault in the region besides the Wellington Fault, and is estimated to
have accommodated 1-3mm/yr slip out of a total of 8mm/yr dextral slip across this portion of the
NIDFZ (Beanland 1995). Although we do not have data on either several (>2) well-dated offsets
or a large offset that reflects slip over several thousand years, we can propose minimum and
maximum slip rates on the Alfredton fault from the golf course trench data. We consider that
8.5-14 m strike slip occurred in two events over a maximum time of 2960 yr (the maximum age
of sample Alf 1/3), or a minimum time of 510 yr (the minimum age of the oldest peat sample Alf
2/11) giving a lateral slip rate of 2.9-27 mm/yr. The minimum slip rate is consistent with the
estimate of McCallion (1996) who trenched the fault c.1 km north of the golf course trenches.
At McCallion’s site, terrace risers are offset laterally by 20 ± 2 m and gravels underlying the
terrace surface are dated at 7 ka. He calculated a minimum net slip rate of 2.6 mm/yr when
corrected for the vertical component and the strike of the fault (McCallion 1996). Two
earthquakes are recorded in his trench; faulting of the topsoil in both trenches suggests fairly
recent rupture but the soils are not dated. Although these estimates are first approximations, a
c.3 mm/yr minimum slip rate is larger than previously estimated (1-3 mm/yr; Beanland 1995)
and represents a significant proportion of the integrated dextral slip rate in this region (8 mm/yr,
Beanland 1995). Furthermore, if cumulative Quaternary offset is less than a few hundred metres
as suggested by Kelsey et al. (1995), the data would require that the slip increased to this rate
within the last few hundred thousand years.
13
Saunders Road Fault
The Saunders Road Fault is a dextral fault that occurs as a series of discontinuous en
echelon scarps c.1-4 km long that trend more easterly (050-065°) than the Alfredton Fault and
extend for at least 13 km northeast from the northern end of the Alfredton Fault (Fig. 2). The
faults appear to dip steeply as they have straight trends across ridges and valleys. Scarps have
fairly steep (20-40°) grassy faces and are composed of bedrock with a thin soil cover. Most
traces are eroded by the larger modern streams, gullies, and landslides, but are preserved as
uphill-facing segments across steep slopes in between the gullies and on ridge crests. Some
individual scarp segments appear very fresh where small (c.30 cm high), or very steep (up to
60°) scarps and complex and fine-scale fault features (e.g., bifurcations, open fractures, sharp
curves) are preserved.
Few offset features are preserved along the Saunders Road Fault, but the smallest dextral
offsets are 3.5-6 m, with a vertical component of 0.5-1 m, typically with northwest side
upthrown (Schermer et al. 1998). One strand of the fault at Kaitawa (Fig. 4) displaces a terrace
riser by 9-13 m dextral, 2-2.5 m vertically; the scarp on the lowest terrace is 1-1.5 m high, but no
dextral offset can be measured there. Thus, we interpret the older terrace offset to be the result
of at least two events. If the terrace is equivalent in age (c.7 ka) to that near Alfredton
(McCallion 1996), we can calculate an approximate maximum single-event lateral slip of 4-6 m,
a dextral slip rate of c.1.3-2 mm/yr, and an estimated recurrence interval of 7000/2=3500 yr. The
Saunders Road Fault juxtaposes identical bedrock types of the Eketahuna Group (Saunders
Siltstone; Neef 1974; Kelsey et al. 1995) and thus has little cumulative vertical displacement.
The locally fresh fault scarps suggest the possibility that part of the Saunders Road Fault
ruptured in recent or historical time, but we cannot constrain the timing or magnitude of any
causative earthquake. The overall somewhat eroded nature of the trace suggests the last rupture
was within the last few thousand years rather than a few hundred. On the other hand, it is
intriguing that the compilation of 1934 earthquake damage (Downes et al. 1999) lists particularly
severe cracking on the Alfredton-Pongaroa road at Kaitawa, where it is crossed by one trace of
14
the Saunders Road Fault (Fig. 4). The Saunders Road school closed after the earthquake because
damage from chimney collapse was too severe to repair (Edmonds 1987). Without further
detailed study of the Saunders Road Fault, however, we cannot distinguish between cracking due
to shaking of poorly consolidated road fill and fault displacement, or whether any strands of the
fault might have ruptured in historical times. There are no constraints on the rupture length other
than a maximum of 25 km from the mapped length of the fault. It is difficult to tell whether the
short length of individual scarps may be in part a real feature of the Saunders Road Fault or
whether this characteristic is due to erosion.
Waitawhiti Fault
The Waitawhiti Fault (Figs. 2, 3D) exhibits the longest surface trace in the southeastern part
of the study area, with a trend of 040-060° and a length of 5-6 km. One exposure of the fault
has an orientation of 058/77° NW; however, the trace across topography indicates subvertical
and locally steep southeast dips. The upthrown side changes along strike, with the southeast side
typically upthrown in the central section, but northwest side up at the ends. The fault dies out to
the south where it splays into two short, subdued scarps across terrace surfaces c.5 m above
Waitawhiti stream (Fig. 2) but does not cut the lowest two terraces. The scarp is similarly
subdued and splayed at the northeast end. The trace of the northern half of the fault, although
difficult to follow, can be located at track and ridge crossings.
In the central section of the Waitawhiti Fault, the scarp face, developed on mudstone with
thin soil, is fairly fresh, with a scarp-face angle of up to 45°, typically c.20-25°, and shows
minimal erosion by streams and landslides. Although no dates are available, these characteristics
suggest the last rupture was within the last few thousand years. The scarp face is somewhat
lower angle and more degraded than scarps along the Alfredton Fault and the Waipukaka Fault,
despite being in similar materials on similarly steep slopes. Scarps are rarely preserved on
northeast-facing sides of ridges in the area, where dextral offset would produce a downhill-facing
scarp or a smaller uphill-facing scarp.
15
Measurements of offsets are limited despite careful field study. There are several locations
where ridgelines or gullies offsets of c.4 and c.8 m are observed, and one location where a gully
is offset 4±2 m, (Fig. 5B), thus the single-event lateral slip appears to be c.4 m. The lateral to
vertical component is ≥4:1 on all measured offsets. The largest geomorphic offset is a pair of
streams laterally offset by 50 m (Fig. 3D); this is the same as the measured displacement of the
contact between two Miocene bedrock units (Schermer et al,. 1998). These observations suggest
that the fault has accumulated very little net slip, and thus has formed very recently.
Berryman & Cowan (1993) noted that the geomorphic expression of the fault suggested the
possibility of rupture within the last few hundred years. However, our comparison of the
morphology of the Waitawhiti Fault with other faults in the region suggests that the Alfredton
and Waipukaka Faults were more likely to have had historical surface rupture. The fairly large
single-event displacement of c.4 m suggests the fault is capable of producing large earthquakes,
and further detailed work would be useful in characterising the importance of this fault in a
regional seismic hazard analysis.
East Puketoi Faults
In the Pongaroa area (Fig. 2), the absence of strata younger than Miocene makes it difficult
to interpret the late Neogene-Quaternary history. Bedrock consists of Miocene strata and Late
Cretaceous Whangai Formation (Kingma 1962, 1967; Delteil et al. 1996). Faults in the area are
inferred to have early Miocene movement (Ridd 1967; Delteil et al. 1996). However, the
mapped relationships (Kingma 1967), also indicate a younger period of reverse or dextralreverse movement in post-early Miocene time.
The East Puketoi faults comprise a zone of distributed short fault traces east of the Puketoi
range front. Although many of the traces appear to have a normal component of slip (Berryman
& Cowan 1993), the largest fault in the zone, the Waipukaka Fault (new name; Figs. 2, 3C, 8)
has a reverse component. The faults trend NNE to northeast and locally appear subparallel to
bedding in the underlying Whangai Formation, which dips moderately to steeply northwest in
much of the zone (Kingma 1967). Most of the active fault traces trend more northerly than either
16
the bedding or the older (Miocene) faults (Delteil et al 1996; Kingma 1967). The Miocene
faults juxtapose Whangai Formation against Miocene units and thus have significant cumulative
slip; however, none of the Quaternary traces appear to reactivate the older faults. The Quaternary
traces appear to have limited (<100 m) cumulative slip because they do not juxtapose different
bedrock units and do not significantly offset the contacts of bedrock units (as mapped by Kingma
1967). In some areas, earlier Quaternary fault trends, which are more easterly (Fig. 3C; dashed
lines, Fig. 8), appear to be cut through by the more northerly trending Waipukaka Fault. Seismic
reflection data in the Dannevirke region just to the north of the East Puketoi Fault Zone suggest
that faults commonly dip 40-70°NW and are dominantly reverse faults that have initiated within
the last c.1 m.y. (Beanland et al. 1998).
Several of the freshest traces of the East Puketoi zone were examined in this study. The
northwest-dipping Oporae Road faults (Fig. 8) have a dominant normal component of
displacement. Scarps along the Oporae Road Faults are well preserved, with moderate face
angles of 20-30°, but they do not cut terraces at the major stream crossings. There are no age
constraints on rupture along these faults so we can only surmise activity within the last few
thousand years based on the relative freshness of the scarps. The Weber Fault (Fig. 2), another
fault possibly related to the East Puketoi zone, appears to reactivate an older fault and has
measured dextral offsets of 6 and 13 m of channels cut into a landslide (Berryman & Beanland
1990; Berryman & Cowan 1993). Scarps on this fault are substantially more subdued than those
on Oporae Road Faults or the Waipukaka Fault, but Berryman & Beanland (1990) reported a 1-2
m high scarp on a Pleistocene terrace 30 m above river level, indicating Quaternary activity.
Waipukaka Fault
The Waipukaka Fault is the longest, most continuous fault in the East Puketoi zone. The
active trace extends along a NE-NNE trend, but several right and left steps are marked by zones
of shorter en echelon traces (Figs. 3C, 8). The fault dips west at c. 25-45° in the central section
but appears to be subvertical at the northern end. Between Waipukaka and Waihi Streams, the
youngest trace is a 25° west-dipping thrust that appears to have propagated east from an older,
17
straighter, and presumably subvertical fault (Fig. 8). The mapped extent of the active trace of the
Waipukaka Fault is c.7-10 km, with the larger value including fresh traces southwest of Korora.
The southern portion is obscured by slips or cultivated terraces near Korora and the northern end
dies out gradually north of Horoeka (Fig. 8).
Scarps along the Waipukaka Fault appear extremely fresh and well preserved even where
downhill facing and where the fault crosses the lowest stream terraces. Only a few offset
features were observed (Figs. 3C, 5, 8), and most offset features are incised gullies with slumped
walls. Therefore, the streams could be in part deflected rather than tectonically offset.
Nevertheless, offset features point to a combination of dextral and reverse (locally normal) slip,
with a lateral to vertical ratio of c.5:1 to 2:1, somewhat lower than that for faults farther south.
Minimum dextral slip of c.3 m can be inferred from the geomorphic offsets (Fig. 5).
Two trenches c.35 m apart were excavated perpendicular to the Waipukaka Fault south of
Horoeka, where a 2-4 m high scarp cuts a young alluvial surface (Figs. 3C, 8). Vertical
displacement along the fault dammed a small stream to produce a swamp; the stream has cut
through the scarp. No lateral offset markers are apparent at the site.
Trench logs from the Waipukaka Fault show evidence for multiple slip events on shallowly
west-dipping faults (Fig. 9). The important features of Hendricksen trench 1 (Fig. 9A) are
juxtaposition of Cretaceous bedrock over Quaternary deposits along two low-angle thrust faults
and the presence of three gravelly silt units (units 4, 6, 7) that thicken and coarsen toward the
fault and are not present on the upthrown block except at the scarp face. A soil is developed on
units 8a and 8b, and units 6 and 7 are overlain by a layer of concentrated charcoal dated at <200
BP (sample Hend 1/5, Table 2) The thrust faults exhibit oblique striae (rake 65-75°SW)
consistent with dextral-reverse slip. In Hendricksen trench 2 (Fig. 9B), similar features are
represented, including: three thrust faults juxtaposing bedrock over Quaternary units and three
gravelly silt units that thicken and coarsen toward the fault (units 1, 3, 5), with a fourth gravelly
silt, unit 4, near the fault that does not thicken or coarsen westward. The organic-rich horizon at
the top of unit 8 is interpreted as a buried soil; another buried soil with a concentrated layer of
18
charcoal and historical artifacts (base of unit 0) is developed on unit 6. The artifacts include
glass bottles, glass and china fragments, and the leather sole of a shoe. The eastern tip of unit 1
appears to overlie and interfinger with the sediment that contains the charcoal and artifacts at its
base. The thrust faults in both trenches dip at 16-33°NW, and appear to steepen downward.
We interpret the evidence from both trenches to record at least the last three, and possibly as
many as five earthquakes on the Waipukaka Fault, with the most recent rupture being after
European settlement of the area. A comparison of inferred event stratigraphy in the two trenches
and alternative interpretations of trench 2 are shown in Table 3, based on the radiocarbon ages
(Table 2), correlation of the uppermost colluvial wedges in the two trenches, and the
interpretation that these wedges represent the most recent surface rupture.
Trench 1 has the most straightforward interpretation. The three gravelly silt units in trench 1
(units 4, 6, 7) are interpreted as colluvial wedges formed in separate events because each
overlaps a buried soil or significant thickness of alluvial sediment, and the two lower wedges are
cut by fault B (Fig. 9A). The most recent event is represented by unit 4 and postdates the
“modern” (<200 cal BP) date of sample Hend 1/5. The penultimate event, represented by wedge
unit 6 burying c.80 cm of alluvial silt of unit 7, is constrained to be older than sample Hend 1/3
(310-0 cal BP, Table 2); a root that grew into the wedge material and may have been cut off by
fault B. An earlier event is represented by the gravelly silt at the western end of unit 7 and the
buried soil at the top of unit 8a. There is also a suggestion of a fourth event in trench 1 that did
not produce a colluvial wedge visible in the trench, where fissured bedrock is overlain and
infilled by alluvial sediment, and a faulted or possibly tilted tree root (unit 9) is present at the top
of the bedrock (Table 3, interpretation B). The age of the oldest unit sampled indicates that the
three youngest earthquakes recorded in trench 1 postdate 8290-7880 cal BP (sample Hend 1/4,
Table 2, Fig. 9A). The oldest event would be >8 ka but its maximum age is unconstrained.
Trench 2 provides additional age constraints and evidence of possible additional events (Fig
9B, Table 3). The most recent event is represented by unit 1, which overlaps and grades into the
artifacts layer. The penultimate event, represented by wedge unit 3, is constrained only to be
19
younger than 2.8 ka (2950-2750 cal yr BP, sample Hend 2/2, Table 2). An earlier event is
represented by the buried soil of unit 8 and colluvial wedge unit 5, which would thus shortly
post-date 2.8 ka. The oldest event may be interpreted from the silts and clayey silts in the
western part of unit 8 and sand unit 7, which could be produced from faulting and initial
damming of the stream. A possible fifth event could be interpreted from gravelly silt unit 4 in
trench 2; however, it has a channel rather than a wedge geometry, and has a fine-grained silt
layer within it. We prefer to interpret unit 4 as an alluvial unit cut into unit 5 and overlapped by
wedge 3. It is plausible that units 3, 4, and 5 formed during the same event, as slip propagated
upward toward the surface and lower parts of the scarp collapsed, so there is a minimum of three
events in trench 2. However, the interpretation of unit 4 as an alluvial deposit cut into unit 5
suggests significant time elapsed between units 3 and 5, and correlation with trench 1 suggests
these represent two separate events. In summary, our preferred interpretation (Table 3) is that
four events are recorded in both trenches: the oldest event at >8 ka, three events since 2.8 ka, and
one since AD1750.
Evidence from the trenches indicates a historical age for the most recent rupture. We
correlate the charcoal layers (unit 5 in trench 1 and the base of unit 0 in trench 2) in the two
trenches. These charcoal layers are interpreted to represent the time the earliest settlers cleared
the forest by burning. This interpretation is supported by the artifacts found within the charcoalrich layer in trench 2 and the < AD1750 age of the charcoal layer in trench 1. In trench 1, the
most recent colluvial wedge overlies the charcoal layer, thus constraining the most recent event
to postdate forest clearing and occupation. Further consistency can be provided by the young
(310-0 cal BP) age of sample 1/3, if it is interpreted as a root that post-dates the penultimate
event and was sheared off by faulting during the most recent event. Historical evidence for the
age of the charcoal layer, the time of European settlement and land clearing in the Horoeka area
is provided in Wilson (1976) and in a series of unpublished school centennial histories and
manuscripts held in the Pongaroa district historical society. This evidence suggests the forest
20
was burned and the site was occupied between about 1874 and the 1920s (Schermer et al. 1998),
thus suggesting the most recent earthquake postdates c.1900.
The minimum slip in each event can be constrained by restoring the bedrock fault slivers
along each fault back to the western edge of the underlying colluvial wedge and back to level of
the bedrock in the footwall (Fig. 9). This reconstruction indicates minimum thrust slip of 2 m in
trench 1 and 1 m in trench 2 in the most recent rupture, 0.7 m in both trenches in the penultimate
event (assuming fault B was reactivated in trench 1), and 2 m in trench 1 and 2.5 m in trench 2 in
the antepenultimate (and possibly one prior) event. If the lateral to vertical ratio is similar to that
of locations a few kilometres to the south along the Waipukaka River where the strike of the
fault and the scarp heights are similar (Figs. 3C, 8), we can estimate 3-6 m lateral slip per event
(Figs. 5, 9). Combining the geomorphic evidence with the trench measurements gives a net
(oblique) slip range of 3-7 m per event, while using the oblique striae and dip-slip displacement
measured in the trenches gives an average minimum net slip of c.2 m per event. Since the
trenched segment of the fault trends more northerly than the overall fault, we would expect a
greater than average thrust component.
Waipukaka Fault slip rate
Along the Waipukaka Fault, a slip rate can be estimated from the trench data and the
estimated single-event displacement. Three to four earthquakes in the last 8290-7880 yr with net
(dextral-reverse) slip 3-7 m each (strike slip 3-6 m), gives a range of oblique slip rates from 1.1
to 3.6 mm/yr with a strike-slip component of 1.1-3.0 mm/yr. Calculation of the rate since 27502950 yr (2 or 3 events) yields 2.0-7.6 mm/yr oblique slip with a strike-slip component of 2.06.5mm/yr. Even the minimum slip rate represents a significant addition to the known dextral
shear in the region (Fig. 1); (Beanland 1995).
DISCUSSION
The evidence from mapping and trenching in northern Wairarapa suggests that significant
surface-rupture earthquakes have occurred in historic and prehistoric time. In this section we
21
analyse which faults are most likely to have ruptured in known historical events. We combine
the data on rupture parameters and timing of earthquakes to discuss the seismic hazard in the
region.
Historical surface ruptures
The single-event displacements on the Alfredton and Waipukaka Faults are compatible with
large-magnitude earthquakes, and the geomorphologic and paleoseismic evidence from both
fault zones supports a recent or historical event. We examine the record of large historical
earthquakes in the North Island to assess which, if any, events are likely to have occurred on
faults in the study area. Doser & Webb (2003) suggest that, in the North Island, earthquakes of
Mw 7.2-7.3 or less are unlikely to break the surface. Because radiocarbon age determinations
will not distinguish between events in the last few hundred years, historical evidence provides
the means to establish the timing of rupture. Historical earthquakes near enough and large
enough to be considered as surface rupture candidates for this study include 1855, 1904, 1917,
1931, 1934, and 1942 (Fig. 1, Table 1). Slip on the faults in the study area during the 1931
Hawke’s Bay events is ruled out by the extensive studies on surface rupture and geodetic
displacements (Henderson 1933; Hull 1990), which show the southern extent of the rupture at
Poukawa, 70 km north of the northern end of the Waipukaka Fault Zone. The 1904 Cape
Turnagain earthquake (Fig. 1, Table 1) appears to have an epicentre offshore and a magnitude of
Ms 6.7 (Downes 1995), thus, we consider it unlikely to have produced onshore surface rupture.
Similarly, the 1917 earthquake, although poorly constrained by instrumental recordings, appears
to be somewhat too small (Mw 6.8) and too deep (20 km; Doser & Webb 2003) to have caused
surface rupture. Therefore, we consider the possibilities for rupture in 1855 (Mw 8.2), 1934 (Mw
7.2-7.4) and 1942 (Mw 6.9-7.2), because of the locations of these earthquakes in Wairarapa and
because the area of the Alfredton and Waipukaka Faults was not examined for surface rupture in
the immediate aftermath of these earthquakes. In 1855 and 1942 study focused on the southern
parts of the Wairarapa region. Although Ongley (Ongley 1934; Ongley 1934 (unpub.); Ongley
22
1935) visited the 1934 epicentral region to look for fault scarps immediately after the earthquake,
his field notes indicate that he did not go to the area of the Waipukaka or Alfredton Faults. He
did return to the Alfredton Fault in December, 1935 (Henderson 1936); we discuss those
observations in the following section.
Probable Historical Earthquake on the Alfredton Fault
Historical evidence is permissive of rupture on the Alfredton Fault in the 1855 earthquake,
but does not unequivocally require it. Grapes & Wellman (1988) and Grapes & Downes (1997)
summarise the geological and historical evidence for rupture along the Wairarapa Fault in 1855.
The rupture is conventionally regarded to have had a northward termination at Mauriceville (Fig.
2), where the fault branches into the Pa Valley-Makuri Fault system and the the Dreyers RockAlfredton Fault system. However, the slip observations summarised by Grapes & Wellman
(1988) show that although the vertical component decreases northward to become nearly zero at
Mauriceville, the strike-slip component remains at c.10 m (compared to a maximum of 13 m),
suggesting there may be a northward extension of the rupture with decreasing dextral slip.
Inasmuch as the area was unoccupied and covered by thick forest at the time, any rupture would
have been extremely difficult to find for several decades after the earthquake. Grapes &
Wellman (1988) include on their map the sites along the Alfredton Fault that Lensen (1969) had
recorded as extremely fresh, but the age of displacement as unknown. A contemporaneous
settler's report of a continuous rupture from 97 to 145 km from Palliser Bay cited by Lyell
(1856), which would include the Alfredton Fault, is regarded by Grapes & Downes (1997) as
suspicious because the same report contained other errors in distance measurements. However,
Ongley (1943a) showed the 1855 rupture extending north of Mauriceville as discontinous scarps
to near Alfredton; he visited scarps that were found by early settlers in the area and interpreted to
be 1855 traces, but he had neither independent dating of the scarps nor contemporaneous
accounts.
23
The June (Mw6.9-7.2) and August (Mw6.8) 1942 Wairarapa earthquakes caused considerable
damage in Wairarapa (up to MM8; Downes et al. 2001). Recent analysis suggests the June event
was a strike-slip earthquake in the Australian plate, while the August event was a normal fault
earthquake in the downgoing Pacific Plate (Doser & Webb 2003). The great depth (40 km) of
the August earthquake suggests it was unlikely to cause surface rupture. Neef (1976) noted that
the fault scarp east of the main Alfredton Fault appeared extremely fresh on airphotos taken in
1944, and suggested that this c.2 km segment of fault moved as a bedding-plane-parallel fault in
the August earthquake, with c.1-2 m of reverse displacement and ≤0.6 m dextral displacement.
However, our observations are that the main Alfredton Fault trace is at least as fresh (Fig. 3B),
and probably moved at the same time with greater slip. Contemporaneous reports of rupture in
the 1942 earthquakes are limited to some extremely short scarps northeast of Masterton (labeled
1942 surface break on Fig. 2); (Ongley 1943b). In this study, we examined airphotos from 194344 in the northern part of the area of 1942 ground disturbance and rupture reported by Ongley.
Our airphoto analysis and recent fieldwork by Downes et al. (2001) suggests that the scarp
features are due to landslides rather than faults.
Photographs taken by Ongley in December of 1935 (some published in Ongley (1943a),
others in Institute of Geological and Nuclear Sciences (IGNS) archives) and by Combes in 1947
(IGNS archives) at several sites along the Alfredton Fault show fairly fresh scarps, but the
amount of vegetation and scarp degradation are comparable to that at other sites along the 1855
trace (Ongley 1943a). This photographic evidence does not suggest rupture in the previous few
years (e.g., in 1934 or 1942). Thus it appears that 1855 is the most likely date of the most recent
earthquake along the Alfredton Fault. We cannot, however, completely rule out rupture during
other earthquakes between c.AD 1750 and the time of New Zealand’s first instrumentally
recorded earthquakes (1906), or the 1917 Mw 6.8 earthquake.
If our interpretation of rupture in 1855 is correct, it extends the rupture length by c.20 km
added to the 97 km previously recognised south of Mauriceville (Grapes & Wellman 1988), and
suggests the rupture had to propagate across a c.6 km wide right-step, the Dreyers Rock Fault
24
Zone (Fig. 10). The actual surface slip north of Mauriceville may have been heterogeneous,
however, as there are several kilometres where the scarp is subdued or difficult to locate on
airphotos taken in 1944, and probably reflects the decreasing slip at the northern end of the 1855
rupture. Our estimate of 4-7 m maximum lateral slip on the Alfredton Fault, compared with 1213 m on the Wairarapa Fault (Grapes & Wellman 1988), is consistent with this inference.
Likely Historical Earthquake on the Waipukaka Fault
Along the Waipukaka Fault Zone we can rule out 1855 rupture because the most recent
earthquake rupture exposed in the trenches (Fig. 9) postdates European settlement, which did not
occur in this area until c. 1896. Furthermore, the fault is nearly 150 km north of the interpreted
epicentre of the 1855 earthquake (Fig. 1) and c. 30 km from the northernmost Alfredton Fault
rupture (Fig. 10). Isoseismals for the 1942 earthquake suggest that damage was too slight near
Horoeka for that event to involve local fault rupture (MM6-7; Downes 1995). Shaking and
extensive chimney and building damage is reported at the time of the 1934 earthquake (Whitta
1990; Downes et al. 1999), but no fault or surface features were reported, and we have been
unable to contact any of the people who lived along the fault zone at the time. Ongley's 1934
field notes indicate that he surveyed the area between Pongaroa and the coast after the 1934
earthquake, but did not go to the Horoeka area. Thus, we conclude that 1934 is the most likely
time of an historical event near Horoeka. This finding is consistent with the newly determined
epicentre location of Downes et al., (1999; Fig. 2), which is only 7 km (±20 km) east of the
Waipukaka Fault.
Possible Historical Earthquakes on the Saunders Road Fault?
The gap in fresh scarp morphology between the northern end of the Alfredton Fault and the
southern end of the Waipukaka Fault is c.30 km and coincides approximately with the known
trace of the Saunders Road Fault (Fig. 10). Geomorphologic characteristics suggest that the last
rupture on the fault zone as a whole was older than that to the north and south. It is possible,
however, that the fault did rupture in 1855 or 1934 because some short (1-2 km long) segments
25
of the fault have scarps that appear to be as fresh as those we consider historical, and the intense
road cracking in 1934 occurred along one of the fault segments (G. Downes, pers. comm.. 1998).
The largest aftershock in 1934 (Ml 5.8) also occurred near the Saunders Road Fault (Fig. 2).
Recent work shows that the epicentre of a small, shallow, strike-slip earthquake in 1942 (Mw 5.6)
is also near the Saunders Road fault (Downes et al. 2001, Doser & Webb 2003). These two
small earthquakes are unlikely to have produced surface faulting.
Implications for seismic hazard analysis
We estimate the magnitude of likely earthquakes on the Alfredton and Waipukaka Faults
using the regressions of Wells & Coppersmith (1994), and Stirling et al. (2002), and calculation
of moment magnitude (Hanks & Kanamori 1979) (Table 4). For these calculations, we need to
estimate rupture area and slip. As described previously, the most recent surface rupture on the
Alfredton Fault occurred along a 17-20 km long segment of the fault, and the single-event lateral
displacement was 4-7 m. On the Waipukaka Fault, the surface rupture length is interpreted to be
7-10 km and the single-event oblique slip is estimated at 3-7 m. We can estimate fault area by
using the fault length and average dip together with a range of earthquake depths of 12-23 km
provided by the body-wave modelling of earthquakes in the upper crust of the North Island
(Webb & Anderson 1998; Doser & Webb 2003). We consider the possibility that each fault
ruptured independently, although it is likely that the Alfredton Fault was a small segment of the
much larger 1855 rupture. We also consider the scenario, yet to be tested, that during the
Waipukaka Fault earthquake, coeval rupture also occurred in the 25-30 km long gap that
coincides with the Saunders Road Fault, as is suggested by the map continuity of the faults (Fig.
10).
As can be seen in Table 4, using the regressions of Wells & Coppersmith (1994) for rupture
length and area predicts a much smaller magnitude than that calculated for the regression for slip,
or the calculation based on seismic moment (Hanks & Kanamori 1979). Updating and reanalysis
26
of the Wells & Coppersmith regressions by Stirling et al. (2002) showed that a different set of
regression parameters provides a better fit for events with rupture lengths >10 km and average
surface displacements >2m. Magnitudes based on the Stirling et al. (2002) regressions are
shown in Table 4 to be closer to the slip-based magnitude estimates, however the range of
magnitudes is large due to the combined uncertainties in regression parameters and in depth, and
the corresponding uncertainty in fault area.
Even using the Stirling et al. (2002) regression, there is a mismatch between magnitudes
calculated from slip and those calculated from length of the individual faults, especially for the
Waipukaka fault (Table 4). The regression of surface displacement on rupture length shows a
range of predicted slips from 1-3 m rather than our observed range of 3-7 m. The faults are
shorter than would be expected for their slip, and short faults typically do not generate large
earthquakes. The calculated magnitude is more consistent with the slip-based estimate if a
combined rupture length (32-40 km) for the Waipukaka and Saunders Road Faults is used. The
>50 km length of the MM9 isoseismal for the 1934 earthquake (Fig. 2), the location of the
epicentre and surface rupture at the northern end of the isoseismal, and the long unilateral rupture
length (70 km) and multiple pulses of slip recognised in the waveform analysis of Doser & Webb
(2003), suggest that significant slip occurred southwest of the Waipukaka Fault trace. The
average fault slip of 4.3 m calculated by Doser & Webb (2003) agrees with our observations but
their rupture length is far greater than that observed, even if the Saunders Road fault is included.
We cannot constrain whether slip in 1934 occurred as surface slip on the Saunders Road Fault or
as subsurface slip on the Saunders Road or other faults; however the lack of continuous fresh
scarps southwest of Korora (Fig. 8), suggests that the rupture did not break the surface along
most of its length.
The surface rupture characteristics and geometry of the Waipukaka Fault Zone present
problems for seismic hazard analysis. The fault zone is composed of many small segments
distributed over a wide (c. 1 km) zone (Fig. 8); until this study, the faults were not recognised as
constituting either a continuous structure or a significant seismic source. The incompatibility
27
between surface slip and length of the fault (Table 4) may be due to the relative immaturity of
the structure. If the Waipukaka Fault is a newly developing fault, a short surface rupture length
relative to slip might be expected until a system of through-going faults can connect along strike
and at depth, as has been argued for strike-slip faults by Wesnousky, (1988). Alternatively, the
oblique-reverse character of the fault zone shows some similarities to reverse faults that ruptured
in large earthquakes, such as those in 1932 Chang Ma, China 1980 Spitak, Armenia 1971 San
Fernando, California, and 1992 Suusamyr, Kyrgyzstan (Rubin 1996; Ghose et al. 1997). These
earthquakes exhibited complex and highly segmented surface rupture patterns, with large (kmscale) gaps between surface rupture segments, suggesting this may be a characteristic of large
intracontinental reverse fault earthquakes (Rubin 1996). Ghose et al (1997) argued that the short
surface rupture in the 1992 Suusamyr (Ms 7.3) earthquake is related to redistribution of slip at
depth on secondary splays and variable surface slip.
Although we can place some constraints on the maximum magnitude of an earthquake on
both the Alfredton and Waipukaka Faults, it is considerably more difficult to evaluate recurrence
intervals. Because only the last two earthquakes on the Alfredton Fault could be dated, and
because we have poor constraints on slip rate, we cannot calculate a long-term recurrence
interval, or evaluate whether the fault moves in a time-predictable manner. Minimum recurrence
intervals can be estimated from the golf course sites, where two earthquakes occurred in the last
700-2900 yr, and the Percy site, where two earthquakes occurred in the last >540 yr. However,
if one of the earthquakes reflects the northern end of the 1855 earthquake on the Wairarapa
Fault, any apparent recurrence interval does not apply to the Alfredton Fault alone.
The likely surface rupture on the Alfredton Fault in 1855 has implications for assessing
seismic hazard along the Wairarapa and Alfredton Faults. In addition to extending the rupture
length by c.20 km, the rupture had to propagate across a c.6 km wide right step, the Dreyers
Rock Fault Zone. This stepover is the largest geometric discontinuity in the Wairarapa Fault
zone with the exception of one similarly-sized left step c.20 km north of Palliser Bay (Grapes &
Wellman 1988). Such a stepover would ordinarily be considered a barrier to rupture propagation
28
in an analysis of seismic hazard based on mapped fault characteristics. Fault-segment rupture
lengths are usually measured from one mapped geometric discontinuity (e.g., stepover, gap,
bend, or intersection with other faults) to the next (e.g., Schwartz & Sibson 1989). This
approach has proved useful in analysis of some strike-slip faults such as the Imperial Fault of
California (e.g., Sieh 1996), and the North Anatolian Fault in Turkey (e.g., Barka & KadinskyCade 1988), but would not have been able to describe the 1992 (Mw 7.3) Landers earthquake in
southern California, which ruptured across several pre-existing faults of different orientations
and had significant stepovers and gaps along the surface rupture trace (Sieh et al. 1993). In the
case of the Wairarapa Fault, the earthquake magnitude is increased by only 0.1 magnitude units
by adding in the rupture length of the Alfredton Fault. However, if the Alfredton Fault slips in
conjunction with the Wairarapa Fault, the maximum earthquake in the Alfredton area increases
by c.0.6-0.9 magnitude units if the entire length is considered, or by c.0.3-0.4 magnitude units if
the maximum slip of 12 m on the Wairarapa Fault is used. Paleoseismic data summarised by
Van Dissen & Berryman (1996) suggest that the recurrence interval of earthquakes on the
Wairarapa Fault is 1160-1880 yr (with a shortest measured inter-event time of 1200 yr). The
maximum recurrence interval on the Alfredton Fault (two events in 2900 yr), is consistent with
this estimate. However, if the interpretation of the earlier event as only somewhat predating the
oldest peat age (c.700 yr) is correct, the apparently shorter inter-event time on the Alfredton
Fault would suggest that earthquakes are smaller and more frequent on the Alfredton Fault than
on the Wairarapa Fault. Although our data do not constrain the timing, it is possible that the
Alfredton Fault has ruptured independently of the Wairarapa Fault (e.g., the penultimate
earthquake on the Alfredton Fault), as well as at the same time as the Wairarapa Fault (e.g.,
1855).
The Waipukaka Fault trenches contain evidence for the last three, probably four,
earthquakes (Table 3). The average recurrence interval since 2.8 ka ranges from 710 to 1400 yr
depending on the interpretation of trench 2. The interevent time for between the oldest two
earthquakes, however, is from >8ka to 2.8 ka (5200 yr), suggesting that the recurrence interval is
29
not regular, unless there are earthquakes between 8 and 2.8 ka that are not recorded in the
trenches. The range of calculated recurrence intervals and the recency of the last event (1934)
makes it unlikely that another earthquake will occur on this fault in the near future.
Slip rates of faults in the forearc region
Although we are unable to put precise constraints on the slip rates of faults in this study,
estimates of minimum slip rates suggest these faults contribute significantly to the active
deformation. Beanland (1995) compiled slip rate data for faults in the North Island Dextral Fault
Belt (Fig. 1), and estimated a cumulative dextral slip rate of c.21 mm/yr across the southernmost
North Island, a rate confirmed by Van Dissen & Berryman (1996). Faults in the forearc north of
the Wellington area have been less intensely studied, and there are fewer data; however,
Beanland (1995) suggested decreasing dextral slip northward from c.8 mm/yr across the
Alfredton region to <2 mm/yr at the north coast, with most of that slip occurring on the
Wellington-Mohaka-Ruahine system. The Alfredton Fault had estimated rates of 1-3 mm/yr
(Beanland 1995) and 2.5 mm/yr (Kelsey et al. 1995) and the entire East Puketoi Fault Zone
(including the Waipukaka Fault) was estimated at <1 mm/yr (Beanland 1995). The minimum
rates estimated in this study--3 mm/yr on the Alfredton Fault and 1 mm/yr on the Waipukaka
Fault--represent significant additions to the slip rate budget in the region, and indicate that
further work needs to be done to understand the distribution of slip rates on the many faults of
the forearc.
A newly formed fault in the forearc
The Waipukaka Fault may be a newly forming fault in a region in which older faults are no
longer favorably oriented. Cumulative strike-slip on the faults in this study does not exceed a
few hundred metres, suggesting recent initiation and/or low slip rates. Wesnousky (1988)
described a seismological evolution of strike-slip faults which suggests that faults with more
cumulative offset have fewer geometric irregularities. Kelsey et al. (1995) presented evidence
for the young nature of strike-slip displacement on the Alfredton Fault (<0.5 Ma) when other
30
faults to the west were abandoned; they suggested this change in deformation is related to >10°
clockwise rotation of the Wairarapa region with respect to more westerly parts of the Australian
plate. The Waipukaka Fault and other faults in the East Puketoi zone strike more northerly than
older northeast-striking faults, suggesting that in this area a vertical-axis rotation may be
important in causing misorientation. There is also paleomagnetic and geodetic evidence for
clockwise rotation at a rate of c.7-8°/m.y. of the Hikurangi margin in Quaternary time (Walcott
1984; Lamb 1988; Beanland 1995). Perhaps the continual rotation prevents long, continuous
faults from forming, and that there has not been enough time for the currently active faults to
have coalesced since the older faults were abandoned.
CONCLUSIONS
Fieldwork and paleoseismicity studies on several of the major faults within the 1934 MM9
isoseismal region indicate that the major faults in the area, including the Alfredton, Saunders
Road, Waitawhiti, and Waipukaka Faults, are dextral-reverse faults with single-event slip
ranging c.3-7 m and are capable of generating surface-rupture earthquakes of M≥7.
Geomorphologic characteristics suggest surface rupture occurred on each of these faults within
the last few thousand years, and along the Alfredton and Waipukaka Faults in historical time.
Estimated minimum lateral slip rates indicate the Alfredton and Waipukaka Faults represent a
significant component of upper plate dextral shear.
Paleoseismology studies using trenching and 14C dating, combined with historical data,
provide insight into the rupture history of the Alfredton and Waipukaka Faults. Along the
Alfredton Fault, data from trenches reveal evidence for the last two surface rupture earthquakes
in the last c.2900 yr, with the most recent event occurring after AD1750. Analysis of published
and unpublished historical evidence suggests the Alfredton Fault ruptured in 1855 during the M
8.2 Wairarapa earthquake, and extends the c.100 km rupture of the Wairarapa Fault during this
earthquake north and east across a 6 km wide right step for an additional c.20 km. Along the
Waipukaka Fault, two trenches reveal evidence for 3 surface-rupturing earthquakes in the last
31
2800 yr, and one event at >8000 yr. The youngest event is interpreted to postdate a charcoal
horizon that contains European historical artifacts. Analysis of historical evidence suggests the
most recent earthquake occurred in 1934.
Future earthquakes of M>7 are possible in the region, but recurrence intervals are not
constrained. However, the recency of events combined with estimates of slip rate suggest long
recurrence intervals (thousands of years) and a fairly low probability of an earthquake within the
next century, at least on the Alfredton and Waipukaka Faults. It is important to determine
whether the Saunders Road Fault has experienced historic (or young) surface rupture because
this fault could have a slip deficit relative to nearby faults that might indicate increased seismic
hazard. Estimation of seismic hazard is made more difficult in the region because the known
fault scarp lengths appear too short to have accommodated the estimated single-event
displacements. Faults in the region are highly segmented, disconnected, and probably
structurally and geometrically immature. A component of seismic slip may not result in surface
rupture, and geometric discontinuities at the surface may not be significant barriers to rupture
propagation.
ACKNOWLEDGEMENTS
New Zealand EQC Research Foundation Project 97-320 provided financial support for this
research. We thank Gaye Downes, Diane Doser, and Terry Webb for providing unpublished
data, and helpful discussion, and Pilar Villamor for help with trench logging. We also thank the
Percys and Hendricksens for permission to trench on their properties. ERS also wishes to thank
the Percys for providing lodging and hospitality during her stay in Wairarapa, and gratefully
acknowledges the University of Otago Geology Department for providing a William Evans
Visiting Fellowship to partially support her travel to and stay in New Zealand.
32
REFERENCES CITED
Bagnall, A. G. 1976: Wairarapa - an historical excursion. Masterton, New Zealand, Hedley's
Bookshop, Ltd.
Barka, A. A.; Kadinsky-Cade, K. 1988: Strike-slip fault geometry in Turkey and its influence on
earthquake activity. Tectonics 7: 663-684.
Barnes, P. M.; Mercier de Lépinay, B. 1997: Rates and mechanics of rapid frontal accretion
along the very obliquely convergent southern Hikurangi margin, New Zealand. Journal of
Geophysical Research 102: 24931-24952.
Barnes, P. M.; Mercier de Lépinay, B.; Collot, J.-Y.; Delteil, J.; Audru, J.-C. 1998: Strain
partitioning in the transition area between oblique subduction and continental collision,
Hikurangi margin, New Zealand. Tectonics 17: 534-557.
Beanland, S. 1995: The North Island Dextral Fault Belt, Hikurangi Subduction Margin, New
Zealand. Unpublished Ph.D. Thesis, lodged in the library, Victoria University of Wellington,
Wellington, New Zealand.
Beanland, S.; Haines, J. 1998: A kinematic model of active deformation in the North Island, New
Zealand, determined from geological strain rates. New Zealand Journal of Geology and
Geophysics 41: 311-324.
Beanland, S.; Melhuish, A.; Nicol, A.; Ravens, J. 1998: Structure and deformational history of
the inner forearc region, Hikurangi subduction margin, New Zealand. New Zealand Journal of
Geology and Geophysics 41: 325-342.
Berryman, K. R.; Beanland, S. 1990: Reconnaissance of epicentral area of May 13, 1990 "Weber
2" earthquake, Dannevirke Distric. DSIR New Zealand Geological Survey EDS Immediate
report 90/5: 1-12.
Berryman, K. R.; Cowan, H. 1993: Manawatu-Wanganui Region Earthquake Hazard Analysis.
Institute of Geological and Nuclear Sciences Client Report 333902/A: 30 p.
Cashman, S. M.; Kelsey, H. M.; Erdman, C. F.; Cutten, H. N. C.; Berryman, K. R. 1992: Strain
partitioning between structural domains in the forearc of the Hikurangi Subduction Zone,
New Zealand. Tectonics 11: 242-257.
Darby, D.; Beaven, J. 2001: Evidence from GPS measurements for contemporary interplate
coupling on the southern Hikurangi subduction thrust and for partitioning of strain in the
upper plate. Journal of Geophysical Research 106: 30,881-30,892.
Darby, D. J.; Meertens, C. M. 1995: Terrestrial and GPS measurements of deformation across
the Taupo back arc and Hikurangi forearc regions in New Zealand. Journal of Geophysical
Research 100: 8221-8232.
33
Delteil, J.; Morgans, H. E. G.; Raine, J. I.; Field, B. D.; Cutten, H. N. C. 1996: Early Miocene
thin-skinned tectonics and wrench faulting in the Pongaroa district, Hikurangi margin, New
Zealand. New Zealand Journal of Geology and Geophysics 39: 271-282.
DeMets, C.; Gordon, R. G.; Argus, D. F.; Stein, S. 1994: Effects of recent revisions to the
geomagnetic reversal time scale on estimates of current plate motion. Geophysical Research
Letters 21: 2191-2194.
Doser, D. I.; Webb, T. H. 2003: Source parameters of large historical (1917-1961) earthquakes,
North Island, New Zealand. Geophysical Journal International 152: 795-832.
Downes, G. L. 1995: Atlas of isoseismal maps of New Zealand earthquakes. Institute of
Geological and Nuclear Sciences Monograph 11: Lower Hutt, New Zealand. Institute of
Geological and Nuclear Sciences Limited. 304 p.
Downes, G.; Dowrick, D.; Smith, E.; Berryman, K. 1999: The 1934 Pahiatua earthquake
sequence: analysis of observational and instrumental data. Bulletin of the New Zealand
National Society for Earthquake Engineering 32: 221-245.
Downes, G. L.; Dowrick, D. J.; Van Dissen, R. J.; Taber, J. J.; Hancox, G. T.; Smith, E. G. C.
2001: The 1942 Wairarapa, New Zealand, earthquakes: Analysis of Observational and
instrumental data. Bulletin of the New Zealand National Society for Earthquake Engineering
34: 125-157.
Downes, G. L.; McGinty, P. 2001: New insights on old New Zealand earthquakes: Was the 1942
Mw 7.2 Wairarapa earthquake triggered by the 1934 Mw 7.4 Pahiatua earthquake?
Seismological Research Letters 72: 266-267.
Dowrick, D. J.; Rhoades, D. A. 1998: Magnitudes of New Zealand earthquakes, 1901-1993.
Bulletin of the New Zealand National Society for Earthquake Engineering 31: 260-280.
Dowrick, D. J.; Smith, E. C. G. 1990: Surface wave magnitudes of some New Zealand
earthquakes 1901-1988. Bulletin of the New Zealand National Society for Earthquake
Engineering 23: 190-204.
Edmonds, J. 1987: Alfredton, The School and The People. Carterton, Roydhouse Publishing (for
the Alfredton School Commitee). 347 p.
Ghose, S.; Mellors, R. J.; Korjenkov, A. M.; Hamburger, M. W.; Pavlis, T. L.; Pavlis, G. L.;
Omuraliev, M.; Mamyrov, E.; Muraliev, A. R. 1997: The Ms=7.3 1992 Suusamyr,
Kyrgyzstan, earthquake in the Tien Shan; 2. Aftershock focal mechanisms and surface
deformation. Bulletin Seismological Society of America 87: 23-38.
34
Grapes, R.; Downes, G. 1997: The 1855 Wairarapa, New Zealand, earthquake-Analysis of
historical data. Bulletin of the New Zealand National Society for Earthquake Engineering 40:
271-368.
Grapes, R.; Wellman, H. 1988: The Wairarapa fault. Wellington, N.Z., Research School of Earth
Sciences, Victoria University of Wellington. Geology Board of Studies Publication 4.
Hanks, T.; Kanamori, H. 1979: A moment-magnitude scale. Journal of Geophysical Research
84: 2348-2352.
Henderson, J. 1933: Geological aspects of the Hawke's Bay earthquakes. New Zealand Journal
of Science and Technology 15: 38-75.
Henderson, J. 1936: Report of Director. Extract from the Annual Report of the Department of
Scientific and Industrial Research, 1935-36, 30th Annual Report: 1.
Hull, A. G. 1990: Tectonics of the 1931 Hawke's Bay earthquake. New Zealand Journal of
Geology and Geophysics 33: 309-320.
Kelsey, H. M.; Cashman, S. M.; Beanland, S.; Berryman, K. R. 1995: Structural evolution along
the inner forearc of the obliquely convergent Hikurangi margin, New Zealand. Tectonics 14:
1-18.
Kingma, J. T. 1962: Sheet 11--Dannevirke. Geological map of New Zealand. 1:250,000.
Wellington, New Zealand. Department of Scientific and Industrial Research.
Kingma, J. T. 1967: Sheet 12--Wellington. Geological map of New Zealand. 1:250,000.
Wellington, New Zealand. Department of Scientific and Industrial Research.
Lamarche, G.; Beanland, S.; Ravens, J. 1995: Deformation style and history of the Eketahuna
region, Hikurangi forearc, New Zealand, from shallow seismic reflection data. New Zealand
Journal of Geology and Geophysics 38: 105-115.
Lamb, S. H. 1988: Tectonic rotations about vertical axes during the last 4 Ma in part of the New
Zealand plate-boundary zone. Journal of Structural Geology 10: 875-893.
Lee, J. M.; Begg, J. G. 2002: Geology of the Wairarapa area. Institute of Geological and Nuclear
Sciences 1:250,000 geological map 11. 1 sheet +66 p. Lower Hutt, New Zealand. Institute of
Geological and Nuclear Sciences Limited.
Lensen, G. J. 1969: Sheet N153--Eketahuna. Late Quaternary Tectonic map of New Zealand.
1:63,360. Wellington, New Zealand. Department of Scientific and Industrial Research.
Lyell, S. C. 1856: Sur les éffects du tremblement de terre du 23 Janvier, 1855 à la Nouvelle
Zélande. Société Géologique de France, Bulletin Série 2, 13: 661-667.
35
McCallion, M. D. 1996: Late Quaternary displacements and tectonic geomorphology of the
Alfredton fault zone east of Eketahuna, north Wairarapa. Unpublished BSc. (Hons) thesis,
lodged in the library, University of Canterbury, Christchurch, New Zealand.
McCormac, F. G.; Hogg, A. G.; Higham, T. F. G.; Baillie, M. G. L.; Palmer, J. G.; Xiong, L.;
Pilcher, J. R.; Brown, D.; Hoper, S. T. 1998: Variations of radiocarbon in tree rings: southern
hemisphere offset preliminary results. Radiocarbon 40: 1153-1159.
McGinty, P.; Darby, D.; Haines, J. 2001: Earthquake triggering in the Hawke's Bay, New
Zealand region from 1931 to 1934 as inferred from elastic dislocation and static stress
modelling. Journal of Geophysical Research 106: 26,593-26,604.
Neef, G. 1974: Sheet N153--Eketahuna. Geological Map of New Zealand. 1:63,360. Wellington,
New Zealand. Department of Scientific and Industrial Research.
Neef, G. 1976: Probable faulting on a minor, reverse, bedding fault adjacent to the Alfredton
fault, north Wairarapa, during the 1 August 1942 earthquake (note). New Zealand Journal of
Geology and Geophysics 19: 737-742.
Neef, G. 1984: Late Cenozoic and Early Quaternary stratigraphy of the Eketahuna District
(N153). New Zealand Geological Survey Bulletin 96: 101.
Neef, G. 1992: Geology of the Akitio area (1:50,000 metric sheet U25 BD, east), northeastern
Wairarapa, New Zealand. New Zealand Journal of Geology and Geophysics 35: 533-548.
Neef, G. 1997a: Stratigraphy and structure of an outboard part of the forearc of the Hikurangi
margin, North Wairarapa, New Zealand. Journal and Proceedings of the Royal Society of
New South Wales 130: 1-24.
Neef, G. 1997b: Stratigraphy, structural evolution, and tectonics of the northern part of the
Tawhero Basin and adjacent areas, northern Wairarapa, North Island, New Zealand. New
Zealand Journal of Geology and Geophysics 40: 335-358.
Ongley, M. 1934: Eketahuna Subdivision. Extract from the Annual Report of the Department of
Scientific and Industrial Research, 1933-34, 28th Annual Report: 2-6.
Ongley, M. 1934: Field notes on excursions to the area of the 1934 Pahiatua earthquake.
Unpublished notes, Institute of Geological and Nuclear Sciences, New Zealand.
Ongley, M. 1935: Eketahuna Subdivision. Annual Report of the Department of Scientific and
Industrial Research, 1934-35, 29th Annual Report: 1-6.
Ongley, M. 1943a: Surface trace of the 1855 Earthquake. Trans. Roy. Soc. New Zealand 73: 8489.
36
Ongley, M. 1943b: Wairarapa earthquake of 24th June, 1942, together with map showing surface
traces of faults recently active. New Zealand Journal of Science and Technology 67-78.
Pillans, B.; McGlone, M.; Palmer, A.; Mildenhall, D.; Alloway, B.; Berger, G. 1993: The last
glacial maximum in central and southern North Island, New Zealand: A paleo-environmental
reconstruction using the Kawakawa tephra Formation as a chronostratigraphic marker.
Palaeogeography, Palaeoclimatology, Palaeoecology 101: 283-304.
Reyners, M. 1998: Plate coupling and the hazard of large subduction thrust earthquakes at the
Hikurangi subduction zone, New Zealand. New Zealand Journal of Geology and Geophysics
41: 343-354.
Ridd, M. F. 1967: Miocene transcurrent movement on the Pongaroa Fault, Wairarapa, New
Zealand. New Zealand Journal of Geology and Geophysics 10: 209-216.
Rubin, C. M. 1996: Systematic underestimation of earthquake magnitudes from large
intracontinental reverse faults: Historical ruptures break across segment boundaries. Geology
24: 989-992.
Schermer, E. R.; Van Dissen, R.; Berryman, K. R. 1998: In search of the source of the 1934
Pahiatua earthquake. Final report, Earthquake Commission research project 97/320: 53 p.
Schwartz, D. P.; Sibson, R. H. 1989: Introduction to workshop on fault segmentation and
controls of rupture initiation and termination. In: Schwartz, D. P.; Sibson, R. H. ed. Fault
segmentation and controls of rupture initiation and termination. U.S. Geological Survey
Open-File Report 89-314.
Sieh, K. 1996: The repetition of large earthquake ruptures. National Academy of Science
Proceedings 93: 3764-3771.
Sieh, K.; Jones, L.; Hauksson, E.; Hudnut, K.; Eberhart-Phillips, D.; Heaton, T.; Hough, S.;
Hutton, K.; Kanamori, H.; Lilje, A.; Lindvall, S.; McGill, S.; Mori, J.; Rubin, C.; Spotila, J.;
Stock, J.; Thio, H. K.; Treiman, J.; Wernicke, B.; Zachariasen, J. 1993: Near-field
investigations of the Landers earthquake sequence, April to July 1992. Science 260: 171-176.
Smith, E. G. C.; Downes, G. L. 1997: Epicentres of large east coast North Island earthquakes,
1931-42. Pp. 25-27 In: Research Report '97, Wellington, New Zealand, Victoria Univ. of
Wellington.
Stirling, M.; Rhoades, D.; Berryman, K. 2002: Comparison of earthquake scaling relations
derived from data of the instrumental and preinstrumental era. Bulletin Seismological Society
of America 92: 812-830.
Stuiver, M.; Reimer, P. J. 1993: Extended 14C database and revised CALIB radiocarbon
calibration program. Radiocarbon 35: 215-230.
37
Van Dissen, R. J.; Berryman, K. R. 1996: Surface rupture earthquakes over the last ~1000 years
in the Wellington region, New Zealand, and implications for ground shaking hazard. Journal
of Geophysical Research 101: 5999-6019.
Walcott, R. I. 1978: Geodetic strains and large earthquakes in the axial tectonic belt of North
Island, New Zealand. Journal of Geophysical Research 83: 4419-4429.
Walcott, R. I. 1984: The kinematics of the plate boundary zone through New Zealand: a
comparison of short- and long-term deformations. Geophysical Journal Royal Astronomical
Society 79: 613-633.
Walcott, R. I. 1987: Geodetic strain and the deformational history of the North Island of New
Zealand during the late Cainozoic. Philosophical Transactions of the Royal Society of
London, Series A: Mathematical and Physical Sciences 321: 163-181.
Walcott, R. I. 1989: Paleomagnetically observed rotations along the Hikurangi margin of New
Zealand. Pp. 459-471 In: Kissel, C.; Laj, C. ed. Paleomagnetic rotations and continental
deformation, Dordrecht-Boston, Netherlands, D. Reidel Publishing Company.
Webb, T. H.; Anderson, H. 1998: Focal mechanisms of large earthquakes in the North Island of
New Zealand: slip partitioning at an oblique active margin. Geophysical Journal
International 134: 40-86.
Wells, D. L.; Coppersmith, K. J. 1994: New empirical relations among magnitude, rupture
length, rupture width, rupture area, and surface displacement. Bulletin Seismological Society
of America 84: 974-1002.
Wesnousky, S. G. 1988: Seismological and structural evolution of strike-slip faults. Nature 335:
340-343.
Whitta, B. 1990: Horoeka/Pukehinau/Maku Schools Jubilee 1900-1990. unpublished; held at
Pongaroa Historical Society.
Wilson, J. G. 1976: A history of Hawke's Bay. Wellington, New Zealand, A.H. & A.W. Reed
Publishing.
38
Table 1 Historical North Island earthquakes, M≥6.5, depth ≤25km
*
Locality
Wairarapa 1
Cape Turnagain
Wairarapa 2
Hawke’s Bay
Hawke’s Bay
Wairoa
Pahiatua
Wairarapa 3
Date
yyyymmdd
lat
18550123
19040808
19170805
19310202
19310213
19320915
19340305
19420624
41.4††
40.6
-40.8
-39.69
-39.43
-38.9
-40.51
-40.96
Epicentre†
long
175.0††
176.8§§
176.0
176.73##
177.48##
177.6
176.29***
175.67†††
Depth§
Ms#
25
LC
20±9
15±3
16±3
8±4
12±5
12±6
6.7
6.6
7.8
7.3
6.9
7.6
7.2
Mw**
8.2
6.8-6.9
7.6
7.3-7.5
6.9
7.2-7.4
6.9-7.2
*
List of earthquakes interpreted to be in the Australian plate.
†
Epicentres from Doser & Webb (2003), except †† Darby & Beanland (1992); §§ from intensity data (Downes
1995); ##Smith & Downes (1997), *** Downes et al. (1999); †††Downes et al. (2001). Typical errors on epicentre
location are ±20km (Downes et al. 1999, 2001).
§
Depth from Doser & Webb (2003), except for 1855 (Darby & Beanland,1992) and 1904 (Downes 1995);
LC=lower crustal.
#
Ms Surface wave magnitude from Dowrick & Rhoades (1998) except 1904 from Dowrick & Smith (1990).
**
Mw Moment magnitude from Doser & Webb (2003) except 1855 from Darby & Beanland (1992).
39
Location,
sample no.*
Table 2. Summary of Radiocarbon Ages from trenches
Calibrated age§ 
13
14C age†
Lab#
cal
BP)
C
%
(yr BP)
Sample
material
1range
2range
Alfredton Fault
Golf course trenches
Alf 1/3
2822±34
Alf 1/2
330±65
Alf 1/1
483±38
Alf 2/11
642±91
Alf 2/2
modern
Alf 2/10
370±50
2950-2850
460-290
520-500
660-540
<200
480-310
2960-2780
510-0
540-470
690-510
<200
510-300
Percy trench
Percy 1/2
Percy 1/3
Percy 1/6
460-290
520-340
310-0
510-0
540-310
470-0
-25.5
-26.4
-29.4
NZA 9094 wood (burned)
NZA 9095 wood
Wk-6261 wood (burned)
Waipukaka Fault
Hendricksen trenches
Hend 1/2
710±130
Hend 1/4
7300±70
740-540
8170-7980
930-460
8290-7880
-27.4
-24.9
Wk-6262
Wk-6263
Hend 1/5
Hend 1/3
Modern
195±57
<200
290-0
<200
310-0
-26.8
-23.6
Hend 2/2
2762±57
2920-2780
2950-2750
-27.3
324±72
450±62
240±70
NZ 7896
NZ 7897
NZ 7899
NZ 7913
NZ 7898
Beta
54176
wood in fault sliver
peat
peat
peat
peat
peat
wood (root)
wood (root or large
branch)
Wk-6264 wood (burnt)
NZA 9096 wood (root) with
bark
NZA 9097 wood (fine
material in organic
soil)
* For sample descriptions and locations, see text and trench logs.
† Conventional radiocarbon age before present (AD 1950) calculated using a Libby half-life of 5568 yr, and
corrected to 13C of -25%. Quoted error is ±1.
§ Calendar years before present (AD 1950) using calibration of Stuiver & Reimer (1993), and references
therein), southern hemisphere correction of -27 radiocarbon years (McCormac et al. 1998), lab error multiplier of
1.217 for Wk samples and 1.0 for NZ and NZA samples.
# Laboratory: Wk, University of Waikato Radiocarbon Dating Laboratory; NZ and NZA, Institute of
Geological and Nuclear Sciences Rafter Radiocarbon Laboratory; Beta, Beta Analytic dating laboratory.
40
Table 3. Interpretation of surface rupture events on the Waipukaka Fault
Event
Date
Trench 1
Date
Trench 2
designation (trench 1,
units, faults
(trench 2,
units, faults
cal BP)
cal BP)
Interpretation A (min. # events)
a, most recent <200
Unit 4, fault B
<100 yr
Unit 1, faults B,
C, bury artifacts
b,
>200
Unit 6, fault B?
<2750-2950 cal Units 3, 4, 5,
penultimate
(reactivated in event
BP
fault A, bury
a)
sample 2/2
c
≤7880Unit 7, Fault A
>2750-2950 cal ?Units 7, 8, 9
8290
BP
(basement
subsidence)
Interpretation B (preferred # events)
a, most recent <200
Unit 4, fault B
<100 yr
Unit 1, fault C,
bury artifacts
b,
>200
Unit 6, fault B?
Unit 3, fault B
penultimate
(reactivated in event
a)
c
≤7880Unit 7, Fault A
c.2750-2950 cal Unit 5, fault A,
8290
BP
bury sample 2/2
d
>7880Units 8a, 8b, 9
Units 7, 8
8290
(basement subsidence,
(reworked
bedrock fissuring)
colluvial
wedge,
basement
subsidence)
41
Table 4. Calculation of moment magnitude (Mw) from inferred rupture characteristics
Fault
Rupture Max
Slip
Mw
Mw
Mw
Mw
Mw from
length slip
from
from
from slip
from
from
moment
(km
(m)
length
length,
length area (S)§ (M0)#
(WC) †
fresh
(S) §
area
(S)§
scarp)
m
(WC)*
Alfredton
17-20
4-7
6.2-6.8
7.2-7.6
6.6-7.2
6.4-7.4
6.9-7.3
1.1-3.4
Waipukaka
7-10
3-7
5.8-6.7
7.1-7.6
6.3-7.0
6.2-7.3
6.6-7.2
1.0-2.9
Waipukaka
+ Saunders
Rd.
32-40
3-7
6.5-7.3
7.1-7.6
6.8-7.5
6.6-7.8
7.0-7.6
1.2-4.1
* Calculated using the rupture length and area given a depth range of 12-23 km from locations of historical
earthquakes, and regressions of Wells & Coppersmith (1994) (WC) for strike-slip faults; for Waipukaka Fault, a
range of average fault dips from 45-90 was used to calculate area.
†Calculated from maximum dextral slip using regressions of Wells & Coppersmith (1994) (WC).
§Calculated from rupture length and area and regressions of Stirling et al. (2002) (S).
#Calculated from M0=µAs, and Mw=(log M0-16.05)/1.5, with µ=3x1011dyne-cm-2, A=area; s=slip (Hanks &
Kanamori 1979).
42
Figure Captions
Fig. 1 A, The Pacific-Australian plate boundary in New Zealand, with relative plate motion from
DeMets et al. (1994). Shaded box shows the area of main figure B. B, Major active faults of the
North Island shown as bold lines, with total dextral slip rate in mm/yr (bold numbers) for several
NW-SE transects across North Island Dextral Fault Belt (from Beanland 1995). Location
abbreviations: W-Wellington, PB-Palliser Bay, M-Masterton, D-Dannevirke, HB-Hawke Bay.
Fault abbreviations: WE-Wellington, WA-Wairarapa, A-Alfredton, MA-Makuri, P-Poukawa,
MO-Mohaka, RU-Ruahine. Hikurangi deformation front shown with teeth on upper plate of
subduction zone. Shaded area indicates Wairarapa region. Historical earthquakes with Mw ≥6.5
and depth <25km from Downes et al. (1999, 2001); epicentre location errors are ± c. 20 km. See
Table 1 for details of earthquakes.
Fig. 2 Map of major Holocene fault traces in the study area. Dash-dot line shows the MM9
isoseismal for the 1934 “Pahiatua” earthquake from Downes et al. (1999). Boxes indicate
location of more detailed maps (Figs. 4, 8). Grid indicates NZMS 260 sheet boundaries.
Fig. 3 Aerial photographs of faults. White arrows show scarp locations. A Portion of photo
894/26, located on Fig. 4, Alfredton Fault, Percy trench area. B Portion of photo 892/35, located
on Fig. 4, Alfredton Fault, golf course area. C Portion of photo 883/46, located on Fig. 8,
Waipukaka Fault trench area. D Portion of photo 892/57, located on Fig. 2, Waitawhiti fault.
Note the more subdued nature of the scarp compared to those in A-C.
Fig. 4 Detailed map of Alfredton Fault with slip observations measured in this study or
confirmed from previous work (Lensen 1969; Beanland 1995; McCallion 1996). As discussed in
text, scarps shown as solid lines are likely Holocene in age. Field measurements of offset
features are typically accurate to 1-2 m horizontal and 0.5 –1 m vertical , see Schermer et al.
(1998) for details. Location of trench sites shown by black boxes. Grid from NZMS 260 sheet
T25, showing 5 km spacing.
43
Fig. 5 Histograms of dextral offsets. Data from Lensen (1969), Beanland (1995), McCallion
(1996), and this study. Errors on measurements are typically 1-2m horizontal and 0.5-1m
vertical. The approximate single-event displacement on each fault is inferred from the smallest
measured offset that produces a peak on the histogram, e.g. 4-7 m on the Alfredton Fault and 4-5
m on the Waitawhiti fault.
Fig. 6 Trench logs of Alfredton Fault at the golf course site; see Figs. 3b, 4 for location. Logs
shows NE wall of ALF-1 and SW wall of ALF-2. Trench 1 is 120 m south of trench 2. Sample
ages are 2 range in calibrated years before 1950 (cal BP). Sample 2/11 was collected from the
opposite (NE) wall of ALF-2.
Fig. 7 Percy trench log. Sample ages are 2 range in calibrated years before 1950 (cal BP).
Fig. 8 Detailed map of Waipukaka Fault and other nearby East Puketoi faults. As discussed in
text, scarps shown as solid lines are likely Holocene in age. Field measurements of offset
features are typically accurate to 1-2 m horizontal and 0.5 –1 m vertical , see Schermer et al.
(1998) for details. Location of trenches shown in Fig. 3c and Fig. 9 (box). Grid from NZMS
260 sheet U24.
Fig. 9A Log of north wall of trench Hendricksen 1. Sample ages are 2 range in calibrated
years before 1950 (cal BP).
Fig. 9B Log of south wall of trench Hendricksen 2. Sample ages are 2 range in calibrated
years before 1950 (cal BP).
Fig. 10. Map showing historical ruptures in the study area. Thin lines show faults mapped in
bedrock (Neef 1997b; Lee & Begg, 2002) thick black lines show Holocene fault scarps, and
thick grey line shows inferred historical rupture extent. Other symbols as in Fig. 2.
44