Microseismicity of the Central Alpine Fault Region, New Zealand

This study investigates the spatial and temporal patterns in microseismicity along the central section of the Alpine Fault, South Island, New Zealand. This section, between Harihari and Karangarua, has significantly lower seismicity than the regions to the northeast and southwest. Several hypotheses of mechanisms said to contribute to the anomaly have been proposed over the years including locked fault, slow slip, shallow creep and external fluids affecting the thermal regime and brittle-ductile transition. Focussing on the shallow crust, the contrasting seismic character is compared to the northern and southern sections from seismicity behaviour, focal mechanisms and seismogenic depth. A temporal array of eight seismographs (including three broadband instruments) was augmented with three GeoNet stations bounding the array. This provided an average spacing of 14 km and a magnitude cut-off of ML 1.6 compared to the GeoNet national network cut-off of ML 2.6 and station spacing of 80-100 km. The Gutenberg-Richter distribution for the four month time frame analysed defned a b-value of 0.75 plus or minus 0.06 which may indicate a locked, heterogeneous zone under high-stress from fluid pressure or a predominance of thrust mechanisms over the survey period. Seismicity over the deployment was within the average range of the last 15 years. The 'horseshoe' shaped seismicity pattern observed from long-term national catologue data is similar for smaller magnitudes. While the central portion of the Alpine Fault is quieter with unusually low b-value, the region is not aseismic. Neither does it experience the level of microseismicity seen in creeping faults. The brittle-ductile transition varies laterally along the fault and is estimated at up to 15 km for most of the survey region but closer to 10 km for the region associated with the highest orogenic uplift rates which compares well with past studies. A local magnitude scale was developed from direct linear inversion of the pseudoWood-Anderson amplitudes and event-station distances. A linear inversion of data from the standard New Zealand magnitude equation characterised an attenuation parameter of 0.0167 km minus 1; more than double the value used in national local magnitude calculations (of 0.0067 km minus 1). Swarm clustering dominates the seismicity character of the time frame. Utilising the earthquake relocation program HypoDD, a selection of clusters both near the Alpine Fault and away from it resolve to point sources. Those close to the Alpine Fault are located in what may be the footwall of the Fault which may indicate that the velocity model has located the events too far to the northwest