No abstract

We have used body-wave modelling to determine the source parameters of 22 moderate to large earthquakes that have occurred along the Hikurangi subduction margin and elsewhere in the North Island of New Zealand since 1964. We have also included events from the Harvard CMT catalogue since 1977 as that catalogue contains smaller events than can be modelled with body waves. We have found that shallow earthquakes occurring in the back-arc Taupo Volcanic Zone and its extension to the north show predominantly normal faulting with nodal planes parallel to the regional fabric and T -axes indicating extension in a direction in agreement with geodetic measurements. A normal faulting solution for the 1974 Opunake earthquake is compatible with the mapping of active faults in the offshore Taranaki region, and the initiation depth of 17 km is close to the expected brittle-ductile transition for that area.Earthquakes within the subducted plate at the southwestern end of the margin are dominated by down-dip tension that is rotated towards the deepest part of the subducting plate. Further north, the events in the upper part of the Wadati-Benioff zone (<200 km) show pure down-dip tension, while this association is less obvious for the deeper events, suggesting that the aseismic part of the subducting slab inferred from plate reconstructions has reached the 670 km discontinuity. Normal faulting earthquakes associated with plate bending below the region of interplate contact occur along the length of the margin, while shallow normal faulting events also occur in the trench, indicating that some slab pull propagates to that region. We infer that large interplate earthquakes are not imminent because in other similar areas they are preceded by outer-rise compressional events that have not been recorded in the region in historical times.Slip vectors of interplate thrust earthquakes indicate that the oblique plate motion across the margin is fully partitioned; that is, the plate interface accommodates very little transcurrent motion, which is instead accommodated by faults in the overlying Australian Plate. Extensional geodetic strain has been measured across the margin in the northeast, which has been regarded as unusual in view of the plate convergence. We suggest that the topography at this part of the margin is too steep to be supported by the current horizontal forces due to friction on the plate interface. This may have occurred due to a reduction in friction because of the subduction of seamounts and the subsequent entrapment of sediments, this also being consistent with the idea of underplating at this part of the margin, or due to the subduction of less buoyant crust. Other margins, such as Japan and Mid-America, where there is known to be subduction of sediments, also show extension in the forearc. Compressional strain observed across the margin in the southwest is more usual for a convergent margin, but it does not show the partitioning we see in focal mechanisms. These data might only agree after the long-term compressional st...

The upgrade of the New Zealand National Seismograph Network in the late 1980s has enabled more accurate earthquake locations to be determined. The catalogue data for events occurring from January 1990 until the end of February 1993 show some new patterns that have not been identified in previous observation periods, and also confirm the persistence of some phenomena observed previously, such as the aseismic corridor through the Nelson region. The deep seismicity data show spatial patterns remarkably similar to those for higher magnitude events recognised by Reyners in 1989. The Hikurangi Benioff zone is marked by intense seismic activity at depths between 150 and 200 km beneath the Central Volcanic Region; it has a sharp discontinuity beneath northwest Nelson and it extends as far southwest as Westport. The Fiordland Benioff zone is distinctly more seismically active in its northern block, and activity is noticeably concentrated in a zone to the west of Lake Te Anau.Shallow earthquakes (depth <15 km) for the period 1990 to February 1993 outline the active eastern boundary of the Central Volcanic Region and an east-west band running from Mt Ruapehu to Mt Taranaki. Earthquakes in this latter group also extend to deep crustal levels and they, and some recently recognised faults in this area, are probably related to a crustal discontinuity in this region. The Cape Egmont Fault Zone has been particularly active during this observation period. In the South Island, the 1990 Tennyson earthquake appears to have triggered activity along the Awatere Fault. The Alpine Fault is seismically quiet in the section from Harihari to Jackson Bay. A band of earthquakes lying to the east of the fault north of Harihari may represent activity associated with the Alpine Fault at depth although this cannot be confirmed by existing data. At the southern end of the Alpine Fault, two subparallel lineaments appear to form the boundaries of the western end of the Otago Range and Basin Province, but they are not associated with any known tectonic feature. In the Wellington region, the earthquakes shallower than 15 km are concentrated in the area between the Wellington and Wairarapa Faults. A group of earthquakes near Carterton possibly represents the interaction between two fault systems of different trends in

b The plate motion model NUVEL‐1 predicts oblique convergence between the Pacific and Australian plates in the South Island of New Zealand. We used P and SH body waveform analysis to constrain the focal mechanisms of the 15 largest earthquakes (Ms > 5.8) that have occurred in this region since 1964, in order to see how the plate motion is accommodated. At the southern end of the Alpine Fault, convergence is achieved by oblique slip movement along a concentrated zone of deformation. In the southern offshore region one event may be related to thrusting of the Australian plate beneath the Pacific plate, and another strike‐slip event probably demonstrates movement on an active strike‐slip fault system parallel to, but offset from, the southern limit of the Alpine Fault. This geometry provides a possible mechanism for the rapid uplift of the Fiordland region. Deformation in the northern South Island is more distributed. In the south‐west Marlborough region partitioning occurs between strike‐slip faulting in the SE and reverse faulting farther NW in the Buller region. We suggest that the partitioning developed as a consequence of an increasing component of shortening that was accommodated by slip on reactivated pre‐existing normal faults in the Buller region. Shortening in the Buller region may have deflected the NE end of the Alpine Fault towards the NW, forming the prominent bend. The Marlborough Fault System, with its youngest and most active faults to the SE, probably developed in an attempt to maintain a through‐going strike‐slip structure as each of the strike‐slip faults was transported towards the north‐west. Partitioning of the opposite polarity (with reverse faulting SE of the strike‐slip faulting) occurs in north‐east Marlborough. The boundary between the two different styles of partitioning in NE and SW Marlborough appears to coincide with a change in the nature of the downgoing slab and a change in strike of faults of the Marlborough Fault System. A normal faulting earthquake on the northern edge of the Chatham rise probably results from a complex interaction of the buoyant continental crust in that region with the subduction zone and the overlying Marlborough Fault System.

SUMMARY We have studied 18 earthquakes of Mw > 5.5 within the North Island, New Zealand region to determine fault parameters for earthquake hazards and regional tectonic studies. Since most large (Mw > 6.5) earthquakes in the North Island occurred prior to 1961, these events provide important information to supplement studies of more recent seismicity. Our results indicate that no large plate interface earthquakes have occurred in the central and southern North Island for the past 80 years, although six strike‐slip earthquakes of Mw≥ 6.8 occurred within the Australian (upper) Plate during an extremely active 25 year period that began in 1917. Only the Mw > 7.2 strike‐slip events were associated with surface faulting, indicating the difficulties that may arise in attempting to identify active faults within this region. Two earthquakes (Mw= 6.9–7.1) off the northeastern North Island in 1947 appear to have occurred along the plate interface and were associated with local tsunamis having runup heights of up to 10 m. Two Mw= 6.8 events also occurred within the Pacific (lower) Plate, highlighting the hazards related to intraslab events. Slip vectors for the earthquakes studied suggest that the majority of transcurrent motion along the plate margin is accommodated within the Australian Plate, similar to the results obtained from studies of more recent, smaller earthquakes. Pure thrusting occurs along the plate interface and T axes of intraslab events indicate downdip tension in the Pacific Plate.

We present the results of body waveform modelling studies for 17 earthquakes of Mw ≥5.7 occurring in the South Island, New Zealand region between 1918 and 1962, including the 1929 Ms = 7.8 Buller earthquake, the largest earthquake to have occurred in the South Island this century. These studies confirm the concept of slip partitioning in the northern South Island between strike‐slip faulting in southwestern Marlborough and reverse and strike‐slip faulting in the Buller region, but indicate that the zone of reverse faulting is quite localized. In the central South Island, all historical earthquakes appear to be associated with strike‐slip faulting, although recent (post‐1991) reverse faulting events suggest that slip partitioning also occurs within this region. The difference between historical and recent seismicity in the central South Island may also reflect stress readjustment occurring in response to the 1717 ad rupture along the Alpine fault. Within the Fiordland region (southwestern South Island) none of the historical earthquakes appears to have occurred along the Australian/Pacific plate interface, but rather they are associated with complex deformation of the subducting plate as well as with deformation of the upper (Pacific) plate. Two earthquakes in the Puysegur Bank region south of the South Island suggest that strike‐slip deformation east of the Puysegur Trench is playing a major role in the tectonics of the region.

Abstract.Horizontal and vertical rotational components of teleseismic surface and body waves are detected by large ring laser gyroscopes. This is illustrated with records from magnitudes 7.0 and 7.3 events at distances of 31 deg. and 42.6 deg. respectively. Phase comparisons with synchronous linear seismometer records confirm the gyroscopic coupling.

The 1994 Arthur's Pass earthquake (MW6.7) is the largest in a recent sequence of earthquakes in the central South Island, New Zealand. No surface rupture was observed, the aftershock distribution was complex, and routine methods of obtaining the faulting orientation of this earthquake proved contradictory. We use a range of data and techniques to obtain our preferred solution, which has a centroid depth of 5 km, M0=1.3 × 1019 N m, and a strike, dip, and rake of 221°, 47°, 112°, respectively. Discrepancies between this solution and the Harvard centroid moment tensor, together with the Global Positioning System (GPS) observations and unusual aftershock distribution, suggest that the rupture may not have occurred on a planar fault. A second, strike slip, subevent on a more northerly striking plane is suggested by these data but neither the body wave modeling nor regional broadband recordings show any complexity or late subevents. We relocate the aftershocks using both one‐dimensional and three‐dimensional velocity inversions. The depth range of the aftershocks (1–10 km) agrees well with the preferred mainshock centroid depth. The aftershocks near the hypocenter suggest a structure dipping toward the NW, which we interpret to be the mainshock fault plane. This structure and the Harper fault, ∼15 km to the south, appear to have acted as boundaries to the extensive aftershock zone trending NNW‐SSE. Most of the ML ≥ 5 aftershocks, including the two largest (ML6.1 and ML5.7), clustered near the Harper fault and have strike slip mechanisms consistent with motion on this fault and its conjugates. Forward modeling of the GPS data suggests that a reverse slip mainshock, combined with strike slip aftershock faulting in the south, is able to match the observed displacements. The occurrence of this earthquake sequence implies that the level of seismic hazard in the central South Island is greater than previous estimates.