Acoustic waveform tomography was applied to Ocean Bottom seismograph data acquired in the seismogenic Nankai subduction zone. To reduce artifacts arising from elastic effects, the inversion focused on early arrivals, and was conducted in two steps; i) fitting phase information, and ii) the incorporation of amplitude information. Image sources were used to simulate the source ghost. Our image successfully retrieved detailed structures, including seaward continuation of the mega-splay fault responsible for past earthquakes.


In the last decade, waveform inversion has proven to be a powerful imaging tool for subsurface structures, with significantly higher resolution than travel time tomography. Among varieties of waveform inversion techniques, acoustic waveform tomography (a combination of traveltime tomography with waveform inversion) has been a popular choice because of its simple formulation and modest computational costs (e.g. Pratt et al., 1998, Ravaut et al., 2004, Operto et al. 2006, Brenders and Pratt, 2007). However, the acoustic method ignores some elastic effects, and there remain concerns about reliability, especially since P-S conversions are not properly accounted for, leading to incorrect modeling of reflectivity coefficients, amplitudeversus- offset effects and focusing/defocusing. A number of synthetic tests have been conducted, but the results are mingled with positive and negative conclusions. With marine data, Brenders and Pratt (2007) and Brossier et al. (2008) demonstrated the imaging ability of acoustic implementation in both crustal and exploration scale, but Barnes and Charara (2008) pointed out the underestimation of sharp velocity contrasts by the acoustic assumption in the presence of S-waves. This study aims to demonstrate the practical validity of the acoustic implementation for wide-angle Ocean Bottom Seismograph (OBS) data acquired in the central part of the seismogenic Nankai subduction zone, Japan (Nakanishi et al., 2008). The inversion delineates complex fault structures, and provides quantitative estimates of significant velocity variations. We investigate strategies to mitigate elastic effects, non-linearity and source ghost effects. Source ghosts introduce directivity to wavefields. While a free surface boundary condition handles the effects automatically (Shipp and Singh, 2002), the incident-angle dependency can alternatively be simulated by the combination of the absorbing boundary condition and image sources. The technique is appealing for the acoustic implementation, since free-surface multiples are typically removed in conjunction with S-wave arrivals during preprocessing. We apply the image source technique, and evaluate the significance.


In seismogenic subduction zones, high resolution quantitative imaging plays a key role in understanding fault structures and associated fluid migration mechanisms. In the central Nankai trough, a mega splay fault was identified using the prestack/poststack depth migration section from a 2D/3D multi component reflection survey, and it has been suggested the fault had a potential role in earthquake displacement propagation and Tsunami generation (Park et al. 2002, Moore et al., 2009). The splay fault branches from the plate boundary, and is overlaid by complex accretionary prisms and a fore-arc basin. The strong reverse polarity of the fault reflection indicates a low velocity layer beneath the mega splay fault, which may be associated with fluid migration or weakened structures.

This content is only available via PDF.
You can access this article if you purchase or spend a download.