next up previous print clean
Next: Conclusions Up: Shragge and Artman: Imaging Previous: Methodology

Experiments

In this section, we present an example of the application of the shot-profile migration algorithm to a synthetic teleseismic earthquake data set. (For additional information regarding the synthetic data set refer to Shragge (2003)). The first step in imaging process is identifying the probing energy, and separating from all other phases emanating directly from the source. In teleseismic imaging, this is simplified by the fact that P- and S-waves are well separated in time at large epicentral distances due to differences in P- and S-wave velocity magnitude. In this case, we use P-waves because the provide a better S/N ratio than latter arriving phase. To characterize the teleseismic source, we exploit the fact that teleseismic arrivals are planar after traversing large distances Shragge (2003). Knowing the location of the earthquake, precise dip and azimuthal orientation of these plane waves can be calculated using standard 1-D reference earth models. Having thus characterized the source, we can construct a wavefield for use in the migration to come. An example of a modeled teleseismic source wavefield is presented in Figure [*]a.

 
sourcedata
sourcedata
Figure 2
a) Constructed source wavefield; b) SV component of the (rotated) receiver wavefields. The data are immersed in a large field of zeros to allow for the correct migration of back-scattered energy that does not reflect from the free-surface within the length of the receiver spread.
view burn build edit restore

The first stage in preprocessing the data is to identify zero time. Although this generally involves time-windowing about the estimated arrival time, with synthetic data it is a straightforward task. The dip and azimuth parameters from source characterization are used to define the appropriate receiver wavefield rotation from [Z,N,E] to [P,SV,SH]. Figure [*]b presents one component of the rotated data.

We present the results of two different migrations in Figure [*]. Figure [*]a and [*]c present the forward-scattered P-S converted mode and the backscattered P-P mode images, respectively.

 
fsbs
fsbs
Figure 3
a) Forward-scattered P-S mode image; b) Synthetic Model; c) Backscattered P-P mode image.
view burn build edit restore

The model used to generate the finite-difference data is presented for reference in Figure [*]b. The appropriate exponent and velocity models for are given in Table 1. The incorporation of v(x,z) velocity models in the wave-equation migration leads to better positioning of model structure relative to other telseismic imaging techniques Shragge et al. (2001); Shragge (2003). This method also affords all the common advantages of wave-equation depth imaging in complex media.


next up previous print clean
Next: Conclusions Up: Shragge and Artman: Imaging Previous: Methodology
Stanford Exploration Project
7/8/2003