24 March, 2010 (16:00 GMT), 5 Merrion Square, Dublin 2.

Speaker: Dr Bill Fry, GNS, New Zealand.
Title: Passive seismic imaging of an earthquake prone sedimentary basin.

Abstract:

We passively image shear-wave velocity structure in an area of dense urbanization. Wellington, the capital of New Zealand is a densely populated urban area located at the boundary of the convergent Australian and Pacific plates. It is prone to the effects of both large strike-slip and large subduction earthquakes. The city lies on the edge of a funnel shaped basin and is underlain by the strike-slip Wellington Fault. The fault has average single event displacements of approximately 5 metres for surface-rupturing events. These events represent a significant seismic hazard. In order to accurately model wave propagation and basin resonance resulting from earthquake scenarios derived from geological mapping and calculations of synthetic seismicity, we must accurately know the basin geometry and velocity structure. Geotechnical data in the basin is sparse and insufficient to provide a full 3D basin model. The dense human development precludes the use of active source seismic imaging. We therefore apply noise-based, passive seismic imaging to define velocity structure and basin geometry. We begin by implementing an interferometry approach to isolate the surfacewave compenent of the interstation Green’s Function between a pair of stations. Due to the contaminating presence of anthropogenic sources proximal to stations, only data with shortterm power spectra that are below the site’s long-term average are correlated. Dispersion observations measured from the stacked correlation functions are inverted with a non-linear algorithm to find a suite of likely Vs profiles. To image shallower velocities, we also apply an array-based spatial autocorrelation technique (SPAC). In this method, we find the azimuthally averaged coherency between recordings at a triangular array of seismometers. The
coherency is then directly inverted for shallow velocity structure with a linear perturbation approach. By using both spatial autocorrelation (SPAC) and 2-station interferometry approaches, we map shear wave velocity on scales ranging from metres to a few hundred metres and provide a detailed basin velocity model for subsequent efforts focused on numerical modelling of the wavefields for plausible earthquake scenarios.