Desmond J. Darby

Subduction zone coupling and tectonic block rotations in the North Island, New Zealand

[1] The GPS velocity field in the North Island of New Zealand is dominated by the long-term tectonic rotation of the eastern North Island and elastic strain from stress buildup on the subduction zone thrust fault. We simultaneously invert GPS velocities, earthquake slip vectors, and geological fault slip rates in the North Island for the angular velocities of elastic crustal blocks and the spatially variable degree of coupling on faults separating the blocks. This approach allows us to estimate the distribution of interseismic coupling on the subduction zone interface beneath the North Island and the kinematics of the tectonic block rotations. In agreement with previous studies we find that the subduction zone interface beneath the southern North Island has a high slip rate deficit during the interseismic period, and the slip rate deficit decreases northward along the margin. Much of the North Island is rotating as several, distinct tectonic blocks (clockwise at 0.5-3.8 deg Myr -1 ) about nearby axes relative to the Australian Plate. This rotation accommodates much of the margin-parallel component of motion between the Pacific and Australian plates. On the basis of our estimation of the block kinematics we suggest that rotation of the eastern North Island occurs because of the southward increasing thickness of the subducting Hikurangi Plateau. These results have implications for our understanding of convergent margin plate boundary zones around the world, particularly with regard to our knowledge of mechanisms for rapid tectonic block rotations at convergent margins and the role of block rotations in the slip partitioning process.

Crustal deformation during 1994–1998 due to oblique continental collision in the central Southern Alps, New Zealand, and implications for seismic potential of the Alpine fault

The positions of 115 ground marks in a 150 × 100 km area of oblique continental collision in the central Southern Alps, New Zealand, have been measured by Global Positioning System (GPS) two to four times between 1994 and 1998. Contemporary velocity and strain rate fields derived from these observations are largely invariant along the northeasterly strike of the mountains and Alpine fault. Across strike, more than 60% of the strain occurs within a band from 5 km NW to 20 km SE of the Alpine fault, but significant strain continues at least a further 60 km SE to near the edge of the Southern Alps foothills. Projections of the fault-parallel and fault-normal components of velocity onto an Alpine faultnormal profile show that about 85% of the NUVEL-1A model relative plate motion is observed within the GPS network. The surface displacements in the high strain rate region are well fit by a model in which stable slip or shearing is occurring at 50–70% of the relative plate rate in a region deeper than about 5–8 km on the down-dip extension of the SE dipping Alpine fault. Material shallower than this is behaving elastically and thus storing elastic strain in the region of the Alpine fault. The longer-wavelength displacements can be modeled either as distributed deformation beneath the Southern Alps, or by localization of elastic strain around the upper end of a discrete NW dipping fault or shear zone that is slipping stably below about 30 km depth and would outcrop near the SE boundary of the mountains if extrapolated to the surface. Strain determined from a small-scale survey network crossing the Alpine fault indicates no significant near-surface aseismic fault slip on the central Alpine fault over the past 25 years. Our results are consistent with independent geological evidence that the central section of the Alpine fault is capable of producing large to great earthquakes.

Crustal deformation of the Marlborough Fault Zone in the South Island of New Zealand: Geodetic constraints over the interval 1982–1994

Crustal deformation across the Marlborough fault zone in the South Island of New Zealand has been investigated by resurveying with the Global Positioning System (GPS) a triangulation and trilateration network across part of the zone. The principal strain rates, which mostly have errors between 5% and 10%, vary systematically across the region. The principal axis of horizontal contraction gradually swings from SE-NW southeast of the Hope fault to almost east-west in the west. The component of velocity parallel to the strike of the faults can account for 95% of the relative motion between the Australian and Pacific plates. This component exhibits little variation west of the Alpine (Wairau) fault, and shows an almost linear variation from the Wairau fault to the east coast. The horizontal strain rate northwest of the Wairau fault is a small east-west uniaxial contraction (∼0.1 ppm yr−1). If it is assumed that the crustal blocks bounded by the major faults are primarily driven by basal shear tractions, with stresses on the faults being relatively insignificant, then the slip rates expected on the faults may be derived from the observed velocity variations across the fault zone. These predicted slip rates are in good agreement with geologically observed slip rates. The larger slip rates on the Hope fault are related to the observation that the strain field extends southeast of the fault over a distance large compared with the spacing between the faults. The agreement between predicted and geological slip rates lends support to the idea that the observed pattern of deformation reflects that in the lower lithosphere.

Terrestrial and GPS measurements of deformation across the Taupo back arc and Hikurangi forearc regions in New Zealand

Geodetic measurements of deformation across the Taupo back arc and Hikurangi forearc regions of New Zealand are derived from Global Positioning System (GPS) measurements made in 1990 and 1991 and triangulation observations made in the 1920s, 1950s, and 1970s. The GPS horizontal coordinate differences have precisions of 5–6 mm and the triangulation observations have precisions of 0.7–1.2 arc sec. These different kinds of observation allow simultaneous estimation of strain parameters and single–epoch coordinates for a total of 184 stations, at 68 of which GPS observations were made. Under the assumption that no shift of the scale or orientation of the GPS reference frame is embodied in the data from 13 GPS stations within the Taupo Volcanic Zone (TVZ) common to the 1990 and 1991 surveys, the rotational, dilatational and shear components of the deformation rate tensor within the TVZ can be estimated for that 1-year interval. The principal extension rate from 1990 to 1991 is 0.21±0.09 × 10−6/yr (68% confidence) at an azimuth of 124±13°, corresponding to 8±4 mm/yr extension over 40 km. Neither the orthogonal principal extension rate, 0.02±0.03 × 10−6/yr, nor the rotation rate, 0.06±0.04 μrad/yr, is significant. The dilatation rate of 0.23±0.09 × 10−6/yr is therefore produced by uniaxial extension. The data from 75 stations distributed from behind the back arc region and well into the forearc region, which were surveyed in at least two of the triangulation or GPS epochs between the 1920s and 1991, allow the spatial variation of the shear components of the deformation rate tensor to be estimated. The maximum engineering shear rates lie in the range 0.1–0.2 × 10−6/yr within the TVZ and maintain these values, with similar orientations, in the forearc region to the east. The western and southern margins of active extension are reasonably well determined by changes in orientation and magnitude of the shear component. These results provide confirmation of the previously less well-determined deformation field, and in comparison with subduction models indicate that the surface deformation is reflecting variation from north to south of coupling of the subduction plate interface, and that both trench suction and gravitational collapse probably contribute to the extension in the forearc region.

Evidence from GPS measurements for contemporary interplate coupling on the southern Hikurangi subduction thrust and for partitioning of strain in the upper plate

The spatial extent and strength of contemporary coupling between the Pacific and Australian plates on the Hikurangi subduction interface beneath the Wellington region of New Zealand are not well known. Consequently, the potential for the plate interface to produce great thrust earthquakes, or to load the major Wellington and Wairarapa faults in the overlying plate, is also unknown. To address these issues, a GPS network across the southern North Island has been partially or wholly surveyed six times between 1992 and 1999 and yields well-determined relative velocities between 82 points. The relative velocity of the east coast with respect to the west coast across the 80-km width of the southern end of the North Island is ∼14 mm yr -1 at an azimuth of 258°. This is about one third of, and closely parallel to, the NUVEL-1A relative velocity of the Pacific with respect to the Australian Plate (39.4 mm yr -1 at 261°). Strain rate solutions indicate partitioning of contemporary strain, with a 35° change in principal axis orientations across the network. Using elastic dislocation theory, an inversion of the observed point velocities and their full variance-covariance matrix determines the variation with depth of contemporary coupling at the subduction interface. If uniform coupling is assumed, then 100% coupling occurs from very shallow depths down to ∼29 ± 1 km. However, if variable coupling is assumed, it reaches 100% at depths between 16 and 24 km. At greater depths, there is a transition zone at least 20 km wide, down to a depth of at least 40 km, over which contemporary coupling diminishes to zero. These depth ranges correlate moderately well with low-angle thrust focal mechanisms recently obtained by other workers and indicate that the downdip width of the interface, over which a great thrust earthquake could occur, exceeds 100 km. Elastic dislocation models also explain the observed partitioning of strain, with Coulomb stress calculations indicating that the Wairarapa fault is being loaded at seismogenic depths.

Possible source models for the 1855 Wairarapa Earthquake, New Zealand

The magnitude 8 Wairarapa, New Zealand, earthquake of 1855 was associated with surface rupture along the Wairarapa fault and regional uplift of the southwest of the North Island. Forward elastic dislocation modelling shows that movement on a steeply dipping Wairarapa fault alone cannot account for the recorded deformation data. Modelling of movement on the subduction interface that underlies the Wellington region as well as the Wairarapa fault also fails to produce a satisfactory ill to the data. Although a complex Wairarapa fault model may be able to explain the deformation pattern if its location, subsurface geometry, and slip distribution could be independently constrained, the best effort supported by available data, a flexed model incorporating a left side step of 8 km at the surface, incorrectly locates the deformation. The best fit to the data is obtained from a listric Wairarapa fault model involving rupture on 0 to 50 km width of the deeper part of the subduction interface. The shallower pan of the subduction interface, east of the Wairarapa fault, apparently did not rupture in 1855, and the uplift mechanism for the overlying Aorangi Range remains unexplained. Partitioning of strike-slip and dip-slip components of the relative plate motions may involve separate earthquakes. Seismological verification of listric fault rupture mechanisms is required to determine the plausibility of the listric model presented here, because its implications arc that the 1855 earthquake did not completely account for the relative plate motion in the region.

Geodetic measurement of deformation in the Taupo Volcanic Zone, New Zealand: The north Taupo network revisited

Abstract A previous geodetic estimate of 18 mm/yr horizontal extension for the Taupo Volcanic Zone (TVZ) immediately north of Lake Taupo for the period 1949–86 is re‐examined for several reasons: this rate has not been confirmed by GPS surveys in the 1990s; newly compiled precise levelling data now allow us to estimate the extent of non‐tectonic deformation attributable to the Wairakei geothermal field; and the precise levelling and lake‐levelling data reveal a spatial variation in tectonic subsidence that casts doubt on the earlier assumption of homogeneous horizontal strain. We use the vertical and horizontal data to derive a Mogi point source model for the geothermal field, and this model allows us to correct the observed horizontal velocities of survey points. Statistical analysis of the corrected horizontal velocities shows that the strain across the TVZ is not homogeneous. When these factors are accounted for, an extension rate of 8 ± 2 mm/yr (1 SE) can be applicable for both 1949–86 and 1986–97. This is about half the previous estimate, which we now consider to be incorrect. The distribution of deformation differs between these periods, and the seismicity of the region shows temporal variations on a similar time‐scale (decades). The extension rate is much greater than can be accounted for by seismic strain release, and the occurrence of historical earthquakes up to M = 6 indicates that a significant part of the measured extension represents seismic strain accumulation. The spatial heterogeneity of the strain partitions the region identically to that derived from geological studies of fault activity. In particular, there is a spatial concentration of extension and tilt about the Whangamata fault system.

Strain distribution across the Australian-Pacific plate boundary in the central South Island, New Zealand, from 1992 GPS and earlier terrestrial observations

As part of a geodetic experiment aimed at understanding the deformation associated with the Australian-Pacific plate boundary in the South Island of New Zealand, in 1992 we reoccupied using Global Positioning System (GPS) techniques a first-order triangulation and trilateration network established in 1978 between Christchurch on the east coast and Hokitika on the west. The network crosses the South Island a few tens of kilometers southwest of the region where the plate boundary changes from a single, throughgoing oblique slip fault, the Alpine fault, to a series of subparallel strike-slip faults, the Marlborough faults. The GPS data have been analyzed as daily network solutions, with about 12 stations in each solution. RMS repeatabilities for stations with multiple occupations are 4 mm, 8 mm, and 15 mm in the north, east, and vertical components, respectively. The observed strain across the network is consistent with the entire NUVEL-IA Pacific-Australia plate velocity being accommodated on land across this part of the South Island. Shear strain rates derived from the GPS and terrestrial data show that the highest strain rate (> 0.4 μrad/yr) occurs in the region of the Southern Alps and Alpine fault. This rate is about two thirds of the rate predicted from the NUVEL-IA plate velocity model assuming that all the plate boundary deformation occurs across this region, implying that about two thirds of the plate motion is accommodated in the vicinity of the Alpine fault. Significant shear strain rates greater than 0.2 μrad/yr are also observed farther east, particularly in the region of the Porters Pass-Amberley fault zone, demonstrating that this zone accommodates a significant part of the plate boundary deformation. Using dislocation modeling, we show that the variation in fault-parallel shear strain rates across the Alpine fault is reasonably consistent with a nonoptimized dislocation model involving a 50° dipping fault that is presently locked to a depth of ∼12 km. One explanation for this observation is that the upper ∼12 km of the crust in this area is storing most of the right-lateral shear component of the relative plate motion as elastic energy that will be released in a future major earthquake. The variation in the other component of shear strain, which may be interpreted as fault-normal relative contraction, is not explained by a simple dislocation model, implying that the normal component of plate motion is taken up in other ways.

Spatial and temporal strain rate variations at the northern Hikurangi margin, New Zealand

We have calculated shear strain rates in the northeastern part of the North Island of New Zealand, using geodetic data collected in the mid-1920s, and 1995. In addition we use observations from a part of the network across Raukumara Peninsula measured in 1976, to look for temporal variation in shear strain rates in that area. The strain rates were estimated using a simultaneous reduction method, assuming uniform strain in space and time within each subnetwork. We find the largest spatial variation in shear strain rates from the mid-1920s to 1995 occurs across Raukumara Peninsula, increasing from 0.06 ± 0.04 ppm/yr in the west to 0.18 ± 0.03 ppm/yr in the east (1σ confidence). The orientation of the azimuth of relative extension shows significant temporal variation across Raukumara Peninsula. From the mid-1920s to 1976 the maximum shear strain rate was 0.22 ± 0.05 ppm/yr with the azimuth of relative extension oriented at 129° ± 7°. From 1976 to 1995 the maximum shear strain rate was 0.14 ± 0.05 ppm/yr with the azimuth of relative extension oriented at 13° ± 12°. We show that a model where the subduction interface is locked down to 15 or 30 km depth, and slipping at NUVEL-IA plate rates below that depth, can explain the strain rates and orientation observed during 1976–1995. To fit the observations from the mid-1920s to 1976, we suggest that additional dip slip occurred on the shallow part of the subduction zone, above 30 km. From this we conclude that the plate boundary below Raukumara Peninsula is locked down to 15 or 30 km depth and the aseismic shallow slip episode that occurred during the mid-1920s–1976 was an isolated event, possibly induced by M ≥ 7 historical earthquakes in the area.

Earthquake triggering in the Hawke's Bay, New Zealand, region from 1931 to 1934 as inferred from elastic dislocation and static stress modeling

During the 1930s the Hawkes Bay region of New Zealand experienced four large earthquakes, Napier (MW 7.6) and Hawke Bay (MW 7.3) in 1931, Wairoa (MW 6.9) in 1932, and Pahiatua (MW 7.4) in 1934. We address the question of whether these comprise a triggered sequence of events. There are significant difficulties in dealing with earthquakes that were recorded 70 years ago as fault parameters are difficult to obtain. With the exception of the Pahiatua earthquake, no primary surface fault ruptures were identified, and locations for the other three events may be in error by tens of kilometers. However, geodetic data were collected before and after the Napier and Wairoa earthquakes, and regions of uplift and subsidence from the former have been mapped from low-order leveling data. This information helps to constrain the fault parameters for the first of these events through elastostatic modeling. Results from recent teleseismic body wave modeling have been used to determine fault parameters for the Hawke Bay event. Our analysis of the induced static stresses with the Coulomb failure criterion shows that the Napier earthquake triggered both the Hawke Bay and Wairoa earthquakes but that the Hawke Bay earthquake probably delayed the Wairoa earthquake. We also conclude that these three events did not trigger the Pahiatua earthquake.