Regional deformation from the 2004 Macquarie Ridge great earthquake, Australia-Pacific plate boundary zone
Burgette, RJ and Watson, CS and Tregoning, P, Regional deformation from the 2004 Macquarie Ridge great earthquake, Australia-Pacific plate boundary zone, AGU Papers and Supplements, Fall Meeting , 14-18 December 2009, San Francisco, USA (2009) [Conference Extract]
The transpressional Australia-Pacific plate boundary south of New Zealand has produced some of the largest strike-slip earthquakes in the instrumental record, including the 23 December 2004 Mw ~ 8.1 earthquake. The oceanic setting of this plate boundary limits terrestrial GPS measurements to sites on Macquarie Island (the only subaerial portion of Macquarie Ridge), southeastern Australia, and New Zealand. We investigate coseismic and postseismic deformation from the 2004 earthquake by analyzing GPS data at 16 sites and compare observed GPS vertical velocity with a relative sea level dataset that spans 96 years. Horizontal coseismic offsets decrease systematically with distance from the earthquake epicenter: 24 mm at Macquarie Island (530 km south of the epicenter) to < 2 mm at mainland Australia and the North Island of New Zealand (> 1800 km from epicenter). We invert the horizontal displacements for a best fitting set of fault parameters using an elastic half-space dislocation model. The modeling results are similar to those reported from seismological techniques: predominantly lateral slip occurred on a fault within the Australian plate, west of the main plate boundary, with a moment magnitude of ~8.0. Due to the symmetry of far-field elastic deformation, the geodetic data cannot discriminate between the two possible nodal planes. We prefer left-lateral slip on a NNW-striking fault based on the orientation of the aftershock pattern and fracture zones. We observe transient postseismic horizontal velocity changes at all of the GPS sites with significant coseismic displacements. Postseismic site velocities are significantly different from the pre-earthquake tectonic velocities. Preliminary modeling suggests that most of the postseismic deformation results from viscoelastic relaxation rather than afterslip. In the four years following the earthquake, the total postseismic deformation is approximately equal in magnitude to the coseismic offset observed at each site, highlighting the broad spatial and temporal scales of effects from this earthquake. This scale of deformation has important implications for the maintenance of a mm-level geodetic reference frame. Vertical coseismic deformation from the earthquake is within the noise of the GPS time series. The 14 year GPS record at Macquarie Island (IGS site: MAC1) shows a preseismic trend of subsidence of 2.2 ± 0.4 mm/a. This rate is corroborated by a 96 year duration relative sea level trend of 4.8 ± 0.6 mm/a. In contrast, late Pleistocene marine terraces suggest a long-term uplift rate of ~0.8 mm/a for Macquarie Island. Pre-earthquake horizontal velocities at MAC1 show movement at 11 mm/a relative to the stable Pacific plate, approximately 1/3 the magnitude, and ~20% more convergent than the relative AUS-PAC plate motion vector. Taken together, these geodetic observations suggest that the Macquarie segment of the AUS-PAC boundary is accumulating strain that will be released in earthquakes with coseismic contraction, and emergence that exceeds the interseismic subsidence.