Body-wave modelling of earthquakes, Hikurangi Margin, New Zealand

Abstract

Long-period body waves recorded at teleseismic distances have been used to model three moderate-sized earthquakes: the Dannevirke earthquake of 10 June 1975, mb 5.7, the Cape Campbell Earthquake of 18 January 1977, mb 5.9, and the Wairoa earthquake of 19 July 1985, mb 5.6. Synthetic seismograms created using trapezoidal source-time functions and well-determined local velocity models were used to resolve source depth, orientation, and complexity. Both the Dannevirke and Cape Campbell earthquakes were complex, and sources with two discrete moment releases were modelled using a second trapezoid delayed in time. The 10 June 1975 earthquake was modelled using a complex source-time function with the second rupture initiating 5.0 seconds after the first. Due to interaction between the surface reflections of the first and second sources, the depth could not be resolved, with two possible depths for the first source at i) 33 km (strike, dip, and rake of 51.8°, 77.2°, and 284.3°) or ii) 50 km (114.8°, 41.3°, and 75.1°). For both cases the first moment release was normal faulting in the subducting plate with a different focal mechanism for the second. The best fitting waveforms were obtained using a model with both sources located at a 33 km depth and a strike, dip, and rake of 37.2°, 36.3°, and 57.4° for the second source. The 18 January 1977 earthquake was located in the lower crust of the subducting plate at a depth of 36 km with a strike, dip, and rake of 71.8°, 32.0°, and -31.5°. A good fit was obtained between part of the synthetic and observed waveforms, but an arrival occurring about 4 seconds after the P wave and complexity in the latter part of the waveforms were not well fit. Neither a double moment release nor inclusion of a low velocity sediment layer at the top of the subducted Pacific Plate provided a better fit The 19 July 1985 earthquake was a simple earthquake with a strike, dip, and rake of 223.7°, 78.4°, and 273.8°. The fault plane is subparallel to the distribution of aftershocks which extend from depths of 30 to 50 km and outline a plane dipping steeply to the northwest. The mainshock was located at a depth of 31 km placing it in the upper crust of the subducting plate at the top of the aftershock distribution. All three earthquakes show normal faulting within the subducting Pacific plate with two events definitely located in the subducting crust. The similarity in depths and focal mechanisms suggest that the downdip extensional stress regime observed in intermediate depth earthquakes extends to shallow depth along the subducted slab. Source complexities may be caused by the addition of complex localised stresses such as the state of coupling on the interface or the location of the hypocentre relative to the bend in the subducting plate. The fault plane and aftershock distribution of the 19 July 1985 earthquake indicates that rupture continued through the subducted crust possibly indicating the breaking up of the descending plate. Both of the other two earthquakes have a focal plane which would also be in agreement with this interpretation.