A Microearthquake Study of the Indian/Pacific Plate Boundary, Central Hawkes Bay, New Zealand

Abstract

Microearthquakes recorded during a 15-day survey with an 11-station array in central Hawkes Bay provide sufficient good quality data to enable the region to be modelled by ray-tracinq and inversion technigues. A simple but effective method has been developed for tracing seismic rays in earth models with dipping-layers. By matching the observed Sp-P and S-Sp intervals to the calculated intervals in a model with a dippinq-interface to represent a discontinuity in the subducted plate, it is possible to locate a set of plane interfaces to describe the discontinuity. The results reveal two discontinuities in the subducted plate that are nearly parallel, with both dips changing from 16° to 22° roughly midway between the coast and the Ruahine Range. The strike of both interfaces is arbitrarily fixed at N 40°E, and the depth separating them is found to be about 4.5 km. The depth of the lower discontinuity beneath the coast is about 24.5 km. Critically refracted P-arrivals from three shallow earthquakes, located more than 110 km from the array to various pairs of stations in the array yield a Pn-velocity of 8.6±0.1 km/sec beneath the lower discontinuity. This value agrees with the P-velocity determined separately, using a specially formulated inversion procedure for the simultaneous least-squares estimation of velocity and hypocentre parameters from an ensemble of P-arrivals. P-velocities of 5.88±0.20 and 6.70±0.13 km/sec for the upper and lower crusts respectively in the overlying plate, are rather higher than those of Garrick (1958) for the Wellington region. Comparison of the locations calculated using a horizontally-layered model and a dippinq-layer model for 20 events inside the array reveal little difference between the two sets of locations. But the model with dipping layers produces an average reduction of about 21% in RMS travel-time residuals. Inversion has also been performed on the S arrival-times from the 20 events while keeping constant the hypocentre parameters obtained from the inversion of the P arrival-times. The Poisson's ratios estimated for the upper and lower crusts in the overlying plate, and the subducted oceanic crust are all about 0.27. The poisson's ratios estimated for the uppermost mantle beneath the two subducted lithospheric discontinuities M1 and M2 are about 0.21 and 0.25 respectively. Three orthogonal projections of the microearthquake locations reveal that the seismicity is largely confined to the top 40 km of the subducted plate. The lack of microseismicity in the Indian plate can be ascribed to the plates being unlocked in the region. In the subducted plate there is a higher concentration of activity where the dip changes from l6° to 22°. Composite focal mechanisms of events in the vicinity indicate stresses associated with a downward bendinq of the plate. Composite focal mechanisms of events lying seaward and landward of the bend, respectively, suggest that the Pacific plate is sinking under its own weight. A detail revealed in the seismicity section along the strike supports the existence of a major fault striking down the dip of the subducted plate beneath the Madden Canyon.