Imaging a Subduction Interface With Local-Earthquake Seismograms, Hikurangi Margin, New Zealand

John Louie
Nevada Seismological Lab, University of Nevada, Reno, NV 89557-0141

Stuart Henrys and Stephen Bannister
Institute of Geological and Nuclear Sciences, 69 Gracefield Rd., PO Box 30-368, Lower Hutt, New Zealand

Back-projection reflectivity imaging of local-earthquake seismograms has revealed the seismogenic structure of the Hikurangi subduction interface. Previous recordings near East Cape showed a 2 km-thick plate interface that strongly converted S to P waves. The 19 Feb. 1990 M6.0 Weber I earthquake was a down-to-the-NW normal event that occurred within the downgoing Pacific slab, while the 13 May 1990 M6.2 Weber II event was a shallow thrust in the overriding Australian plate. The data from a 10-station portable array augmenting the New Zealand national network were sufficient to define a 1-d velocity model and gave consistent composite mechanisms for each sequence. We used seismograms from 501 previously relocated, A-quality events in both sequences to directly image the structure and multi-phase reflectivity of the plate interface. Gathers of vertical seismograms from the relocated events, and corresponding synthetics, show reflections and converted waves traveling upward from the 20 km-deep interface. We computed Kirchhoff-summation image sections for P-P, S-P, and S-S scattering from these seismograms. The imaging artifacts due to poor ray coverage we mitigated with an obliquity factor, an antialiasing criterion, and enhancement by resampling statistics. When illuminated from above with back-scattered energy, the interface appears as a 3-5 km-thick zone with 5-degree NW dip, but offset 5 km down-to-the-NW above the Weber I source area. The two reflectors defining the zone show 1.5 km resolution. When illuminated from below with forward-scattered energy, a wedge of the plate interface on the downdip side of the 5-km offset forms a prominent scatterer. The 5-km NW-down normal-fault offset in the plate may be bulldozing a 5 km-thick fluid-filled sedimentary section under the Australian plate. Subduction of such plate topography may explain the Weber I normal faulting below, and the Weber II thrusting above.