Abstract. In this study we evaluate teleseismically determined focal mechanisms and epicenters for earthquakes along the Macquarie Ridge Complex (MRC) from 45øS to 61 øS and 155øE to 168øE, a region characterized by some previous investigators as undergoing subduction initiation. From 65 centroid moment tensors reported by Harvard, we develop statistical guidelines for choosing 26 which represent better determined, more reliable focal mechanisms for tectonic analysis. Although thrust mechanisms occur in the north, near Fiordland, elsewhere along the MRC the better determined mechanisms virtually all indicate that present-day motion along most of the MRC is strike-slip. This is consistent with sidescan sonar and multichannel reflection data collected between 50øS and 57øS on 1994 and 1996 cruises; the active plate boundar3, zone appears to be quite narrow (< 5 km wide), and no active compressional features can be observed on the seafloor. If we determine a rotation pole for plate boundary motion using only slip vectors from better determined Harvard mechanisms along the MRC, the best fitting "instantaneous" pole is at 57.4øS, 179.4øE, about 2.5 ø north of the NUVEL-I Australian-Pacific pole, which averages motion over the last 3.0 my. If the MRC pole was formerly farther south than at present, this could explain the existence of relict features associated with crustal shortening, such as bathymetric highs and troughs; yet, the absence of active features such as thrust faults, etc., suggests ongoing compression or subduction initiation. We also carefully read arrival times for P phases for 53 earthquakes at 16 teleseismic stations, selected to represent a range of azimuths surrounding the earthquakes; we relocated these earthquakes using standard joint epicentral determination (JED) methods. While most of the better quality relocations lie on or very close to the Australia-Pacific boundar3, as determined on the 1994 cruise, a few epicenters occur well away from the boundar3,, apparently on Cenozoic fracture zones. Thus, on the Macquarie Ridge Complex and other major strike-slip boundaries (e.g., in California), it appears that the very largest earthquakes occur on the principal plate boundar3, fault but that other earthquakes, some quite large, may occur away from the boundary along zones of preexisting weakness.

Large intraplate earthquakes have an almost universal property of occurring in unexpected places, and on March 30, 1986, the Marryat Creek earthquake confirmed this behavioral pattern. The earthquake took place in central Australia (26.2°S, 132.8°E) in a region that was cratonized in the Early to Middle Proterozoic and that has no record of significant previous seismic activity. The main earthquake had a magnitude of Ms ∼5.8 and was associated with a surface fault scarp that is in the shape of a boomerang (convex to the northeast) ∼13 km long, with a maximum displacement of 0.8 m. A preliminary fault plane solution for the earthquake indicates that the mechanism was a combination of thrust and strike‐slip faulting, with the pressure axis striking at N 220°E.

SUMMARY On 1988 January 22 three earthquakes of Ms 6.3‐6.7 occurred in a 12‐hr period near Tennant Creek (Northern Territory) in the Proterozoic North Australian Craton and produced 32 km of surface rupture on two main scarps. the Lake Surprise scarp is shaped like a boomerang with the east arm trending ESE and the west arm, WSW. the Kunayungku scarp lies 7 km to the northwest and trends ESE. Locations of 211 aftershocks determined using data from portable seismograph arrays define a 10‐km by 40‐km aftershock zone elongated parallel to the trend of the surface ruptures. In the western and eastern part of the fault zone, aftershocks occur only south of the Kunayungku and eastern Lake Surprise scarps, respectively. In the central section, in contrast, aftershocks lie primarily to the north of the western arm of the Lake Surprise scarp. Focal depths range from the surface to 8km. In all sections the shallowest earthquakes lie closest to the scarps, whereas deeper events lie further away. In cross‐section, the aftershocks lie in three distinct inclined zones, which delineate fault surfaces ruptured by the main shocks. Aftershocks are concentrated near the edges of the inferred fault surfaces suggesting that stress was relieved in the centre of the faults by rupture during the main shocks. the inferred fault‐plane dips are: 45° to the SSW for the Kunayungku fault, 55° to the NNW for the western Lake Surprise fault and 35° to the SSW for the eastern Lake Surprise fault. These dip directions are consistent with the sense of thrust movement inferred from surface deformation. the reversal of fault dip and sense of movement between the central and end segments requires a complicated fault geometry. the aftershock data would allow a second, south‐dipping plane in the central section that does not intersect the surface. This blind fault may be responsible for much of the moment release interpreted from the modelling of broad‐band teleseismic waveforms (Choy & Bowman 1989). Rupture of the north‐dipping plane may have been controlled by a pre‐existing fault marked at the surface by a quartz ridge 5‐km long and may be of secondary importance. Although the quartz ridge suggests previous faulting, the absence of significant topographic relief together with estimates of erosion rates suggest an average repeat time in excess of 104yr. Preliminary interpretation of water withdrawal histories and water‐level data from a borefield adjacent to the eastern Lake Surprise scarp does not support a causal relationship between the borefield and the occurrence of the earthquakes. The nearby Warramunga (WRA) seismic array provides a record of local earthquake activity from 1965 to the present. the fault area was essentially aseismic until two small earthquakes in 1986 February and a series of six magnitude 4‐5 earthquakes in 1987 January, which were followed by more than 1000 aftershocks. the frequency of occurrence of aftershocks declined very slowly in 1987, with p= 0.2 in the modified Omori formula, whereas for the 1988 aftershock sequence,...

No abstract

No abstract