Subsalt velocity analysis by target-oriented wavefield tomography: A 3-D field-data example

introduction

Accurate reflectivity imaging requires an accurate background velocity model. As seismic exploration moves towards structurally complex areas, wavefield-based tomography that better models band-limited wave phenomena becomes necessary for high-quality velocity model building. Wavefield-based tomography, however, is still expensive for industrial-scale applications (Biondi and Sava, 1999; Shen et al., 2005; Albertin et al., 2006; Fei et al., 2009), both because the method uses more expensive wavefield modeling engines, and because it lacks flexibility and usually requires the use of the whole recorded data set for velocity analysis.

To reduce the cost and increase the flexibility of wavefield-based tomography, Biondi (2006); Guerra (2010); Tang and Biondi (2010) originated the idea of target-oriented wavefield tomography. The idea is to synthesize a target-oriented new data set specifically for velocity analysis. The new data set is designed to be much smaller than the original surface-recorded data set, while still containing all necessary information for velocity updating. This strategy allows us to apply the powerful but expensive wavefield-based technique only in areas where it is necessary, such as in subsalt regions with complex overburdens, and leave areas with relatively simple geologies to be handled by conventional velocity-analysis methods, which are sufficient to produce accurate results.

By localizing the computation within a selected target zone, the target-oriented inversion strategy dramatically improves the efficiency and flexibility of wavefield-based tomography. Therefore, it can greatly shorten the cycle time from seismic processing to interpretation, enabling interpretation-driven interactive wavefield-based velocity analysis, where different geological scenarios can be tested in almost real time (Halpert et al., 2008). The high efficiency may also make velocity uncertainty analysis feasible, which requires inverting velocity models multiple times to build the probability distribution (Tarantola, 2005).

Guerra et al. (2009); Guerra (2010); Biondi (2006) synthesize a new data set for local tomography using the concept of prestack exploding-reflector modeling (PERM). PERM, however, generates crosstalk when multiple image events (reflectors) are modeled simultaneously. This limits the number of reflectors it can model. Manual picking and stochastic encoding methods, such as random-phase encoding, are required to mitigate the impact of the crosstalk.

In contrast, Tang and Biondi (2010) formulate the problem under a seismic-data-mapping (SDM) framework (Bleistein and Jaramillo, 2000; Hubral et al., 1996) and use generalized Born wavefield modeling as the mapping operator to synthesize a new data set for velocity analysis. Generalized Born wavefield modeling is extended from conventional Born modeling (Stolt and Benson, 1986) to include modeling of the non-zero subsurface offset images. As shown by Tang and Biondi (2010), the inclusion of the subsurface offset in the modeling process preserves the correct velocity information and is crucial to the success of this method. One advantage of generalized Born wavefield modeling is that it does not require any picking, but picking can be incorporated if it is desired. Another advantage is that it can model arbitary number of reflectors simultaneously and is minimally affected by crosstalk artifacts.

In this paper, we follow the method of Tang and Biondi (2010) and use generalized Born wavefield modeling to generate a target-oriented data set for local velocity analysis. In the subsequent sections, we first review the theory of target-oriented tomography using synthesized Born data. We then apply the methodology to a 3-D field data set acquired from the Gulf of Mexico (GOM), where we update the subsalt velocities in a target-oriented fashion.

Subsalt velocity analysis by target-oriented wavefield tomography: A 3-D field-data example