next up previous [pdf]

Next: Waveform inversion workflow Up: Shen et al.: Refraction Previous: Shen et al.: Refraction

Introduction

Typically, conventional ray-based methods (White, 1989; Hampson and Russell, 1984; Olson, 1984) are used to derive near-surface velocity. Such smooth solutions may be adequate for areas with simple near-surface velocity distributions, but in geologically complex areas, smooth velocities are not accurate enough for imaging deeper reflectors (Marsden, 1993; Bevc, 1995; Hindriks and Verschuur, 2001). In such cases, waveform inversion (Tarantola, 1984; Pratt et al., 1998; Mora, 1987) tends to give more accurate results (Sheng et al., 2006; Ravaut et al., 2004; Sirgue et al., 2009) by simulating seismic wave propagation.

With more and more long offset data being acquired, particularly on land, it is not uncommon to see shingling in data (Figure 1).

realshot
Figure 1.
Shot gather of 2D land dataset showing shingling phenomenon, red lines denote approximate first-break picks, notice how they become discontinuous as offset increases.[NR]
realshot
[pdf] [png]

Shingling is defined here as where first-break traveltimes are only piecewise continuous when we look at data in shot gathers. This phenomenon is likely to be caused by a near-surface low velocity layer (Figure 2 ). Shingling poses a serious problem for ray-based inversion methods. Note that at the shot gather in Figure 1, the first-break amplitudes decay so much that that at a certain offset, the only reliable pick on a trace becomes the later refraction arrival in time. The discontinuity is caused by tracking different events as we move further away from source location. Since ray-based methods assume the same event is being picked, they tend to fail in these types of situations.

shingling
shingling
Figure 2.
A synthetic data case showing a) a 1D velocity model with low velocity layers in the near-surface and b) the resulting shot gather. Green arrows denote decaying refractions on the right and where they occur in near-surface model on the left. [NR]
[pdf] [png]

Waveform inversion relies on the wave equation and is able to solve the velocity structure quite well even in the presence of low velocity layers. Results from waveform inversion are highly dependent on the accuracy of the starting model and on the frequency content of the data. To relax the dependence of the waveform inversion on the initial model, we propose the work flow described in the next section.


next up previous [pdf]

Next: Waveform inversion workflow Up: Shen et al.: Refraction Previous: Shen et al.: Refraction

2011-05-24