An investigation into the characterisation of seismic traces: optimising shear wave signals with source type, orientation and survey geometry

Abstract

Seismic surveys are used to create visual representations of subsurface geology by detecting the reflections and refractions of seismic waves. In most cases, the seismic wave of interest is the P-wave, the easiest wave to identify in a seismic record. The S-wave is more difficult to identify due to the slower velocity which coincides with surface wave arrivals, however the S-wave has the potential to image much more detail than the P-wave due to its shorter wavelength. While the use of horizontal component geophones to detect S-waves has become increasingly common, there remains little published research on the characteristics of seismic traces recorded on these components compared to traditional vertical geophones. Furthermore, there have been few investigations into arranging survey geometries in such a way as to eliminate surface waves from the seismic record, providing easier identification of S-wave arrivals. This thesis aims to characterise and compare the seismic signals received by horizontal and vertical components of geophones, with a focus on S-wave arrivals, along with an investigation into eliminating surface waves from the seismic record to better identify S-waves. Two surveys were conducted at Whataroa Valley on the West Coast of New Zealand: a 1270 m long array with a rolling string of 48 single vertical component geophones spaced at 10 m intervals and a 70 m long array with eight 3C geophones spaced at 10 m intervals. Sources consisted of 42 array-parallel explosive shots at 20 m intervals and an 80 m line of 10 m spaced hammer and plate shots. The vertically mounted “shear source” plate and horizontal plate used allowed the effect of different shot orientations on the different geophone components to be investigated. The data collected from these surveys indicate that the polarity of S-waves recorded on the horizontal components of the 3C geophones are highly dependent on the on the geometry of the shot direction in regards to the vibration direction of the geophone components. Furthermore, it was found that significantly different first arrival velocities were detected on the horizontal components compared to the vertical for the hammer and plate source shots, but not for the explosive shots. Stacked seismic profiles of the data were also produced, detailing the subsurface structure of the Alpine Fault at this location and highlighting sedimentary units offset by cumulative fault rupture. Further investigations were conducted within an abandoned railway tunnel at Chain Hills, Dunedin, to determine whether placing receivers below ground can delay surface waves enough that S-waves become more distinguishable in the seismic record. 48 single component vertical geophones spaced at 5 m intervals were placed within the tunnel while horizontally mounted hammer and plate shots were conducted at various locations above tunnel. Two lines of eight 3C geophones at 5 and 10 m intervals with array-parallel shear source shots were used to collect S-wave velocities on horizontal components. Possible S-wave arrivals were identified in the vertical component data, however attempts to confirm the nature of these arrivals by comparing with data recorded on the 3C horizontal geophones were inconclusive.