Ambient Seismic Noise Tomography in the Auckland Volcanic Field

Abstract

Auckland, New Zealand's most populous region (1.4 million), is situated over a monogenetic volcanic field of at least 53 volcanoes. The most recent eruption in the Auckland Volcanic Field (AVF) was 600 years ago, therefore the AVF is considered active. The risk and hazard of a volcanic eruption in Auckland are difficult to assess. It has been difficult to produce a high resolution model of the subsurface structure of Auckland due to its low seismicity. However, because of New Zealand's long coastlines, Auckland experiences high energy ocean wave generated seismic noise. This study is the rst to use ocean noise to infer the structure of the Auckland subsurface. We successfully incorporated a University of Auckland borehole seismometer from Rangitoto Island (RBAZ) into an existing network of GeoNet seismometers in Auckland. We were able to correct for the different instrument responses of the 12 seismometers. The vertical components of 200 days of noisy wavefields from 66 possible station pairs were cross-correlated to estimate the impulse response; this way, the resulting time series is as if one of the stations is a seismic source, and the other a receiver. Using these 66 impulse responses, we estimated Rayleigh wave group and phase velocity dispersion relations by multiple filter analysis. Frequency-dependent surface wave speeds with periods between 3 and 10 seconds provided robust information about the subsurface. We inverted group and phase velocity dispersion curves for shear velocities near the surface to approximately 25 km depth. We found that shear velocity variations correlate well with the crust type and surface geology. The distribution of low-velocity zones shows good spatial correlation with the Murihiku and Waipapa Terranes (semi-continental crust). Models with monotonically increasing shear wave velocity mainly associated with the Dun Mountain Ophiolite Belt, of the Maitai Terrane (oceanic crust). Near the surface (0-1.5 km), basement rock exposures correlated well with higher shear wave velocities, while soft cover rock and poorly consolidated sediments correlated with lower velocities. The average of our 16 models also largely agrees with a shear velocity model obtained by joint teleseismic receiver functions and seismic surface wave inversion techniques by Horspool et al. (2006).