Imaging of intrabasalt and subbasalt structure with full wavefield seismic tomography

Abstract

[1] The imaging of subsurface structure using conventional seismic reflection methods is problematic across areas where high-velocity basaltic rocks are intercalated with sedimentary layers of low velocity. In this study we use an approach that combines wavefield inversion of both normal-incidence and wide-aperture seismic data to constrain velocity structure within and beneath the basalt. Our field data are a two-ship long-offset profile acquired in the Faeroe Basin. A conventional traveltime inversion of the wideangle data is performed to obtain a starting model for the wavefield inversion. Synthetic tests are used to explore resolution and to determine a strategy for inversion of the field data. Our final velocity model shows a high-velocity basaltic layer underlain by an intermediate-velocity layer that is presumably sedimentary. The basalts appear highly heterogeneous in velocity, and the degree of heterogeneity increases significantly with depth within the layer. A prestack depth migration, using the final velocity model, reveals a number of new features. Despite the velocity heterogeneity within the basalts, strong reflections are seen within and immediately beneath the basaltic section, and these reflectors are more clearly imaged on the final depth-migrated section than hitherto. The transition from sediments to basalt does not appear to be sharp, uniquely defined, or locally subhorizontal. The data suggest that velocity heterogeneity, and the complexity of the transition from basalt to sediment, may explain the difficulties of imaging beneath basalts west of Britain. Wavefield tomography can resolve this complexity and hence improve both imaging and confidence of interpretation within and beneath basalts.

Inversion of Full Acoustic Wavefield in Local Helioseismology: A Study With Synthetic Data

“…Note that the relatively smooth models obtained from travel-timebased methods (including time-distance helioseismology) are likely to be the result of smoothing and relatively sparse model parameterization, which do not allow an accurate determination of any potentially abrupt structural changes. Our results are therefore compatible with the observations made in terrestrial seismology that full waveform inversions produce betterresolved and more realistic images than travel-time tomography (e.g., Chironi et al 2006).…”

“…Research efforts have been made to address the estimation of noise levels of data in time-distance helioseismology (e.g., Gizon & Birch 2004) and future applications of geophysical wavefield techniques to helioseismic data (Tong 2005) will need to take this into consideration. The inversion of real helioseismic data will then be able to confirm the expected improvement in model resolution with wavefield inversion as demonstrated in previous geophysical studies (e.g., Pratt 1999;Chironi et al 2006). In our inversion, we assume that the acoustic source functions are constant.…”

“…In the helioseismic context, Tong et al (2003c) show, by establishing the sensitivity of the internal structure of helioseismic signals to sound-speed variations, that amplitude modeling is feasible. Such full waveform inversion has already been developed and successfully applied in terrestrial exploration seismology (e.g., Pratt et al 1996;Chironi et al 2006;Brenders & Pratt 2007) and studies based on cross-correlated signals from ambient terrestrial seismic noise (Shapiro et al 2005). Wavefield inversion of cross-correlated signals from terrestrial ambient seismic noise is based on the following principle: cross-correlation of a random isotropic wavefield obtained between two points on the surface gives rise to a waveform that differs only by a factor compared with the Green function between the two points (Shapiro et al 2005, and references therein).…”

“…Sound-speed perturbation model in the wavefield tomographic study. Perturbation is expressed in terms of percentage change to the one-dimensional sound-speed structure as described by Model S (Christensen-Dalsgaard et al 1996). (A color version of this figure is available in the online journal.…”

Seismic waveform tomography with multicomponent data at a groundwater contamination site

“…A companion paper (Pratt and Shipp, 1999) describes one of the first applications to field data, from a crosswell experiment in a layered sedimentary environment. Recently, waveform tomography has also been applied to surface seismic data sets in various environments; e.g., a realistic crustal-scale synthetic model (Brenders and Pratt, 2007), a reflection/refraction survey of the shallow crust across the San Andreas Fault (Bleibinhaus et al, 2007), wide-aperture land data sets over thrust belts (Jaiswal et al, 2008;Ravaut et al, 2004), an ocean-bottom seismic data set at a subduction zone (Operto et al, 2006), a long-offset marine reflection/refraction data set for sub-basalt imaging (Chironi et al, 2006), and a marine reflection data set over a gas deposit (Hicks and Pratt, 2001). To the best of the author's knowledge, the only previous applications to near-surface (i.e., upper 20 m of the subsurface) data used vertical seismic profile (VSP) and surface surveys at the current study site (Gao et al, 2006;Gao et al, 2007).…”