next up previous print clean
Next: Conclusions Up: Shragge: Teleseismic PSM Previous: Synthetic model and data

Experiments

The results of applying the hybrid survey-sinking algorithm to the synthetic data set are presented in this section. The v(z) velocity profiles used in these migrations consist of a 40-km-thick crust overlying an upper mantle half-space. The values of the P- and S-wave velocities employed are given in the first two rows of Table 2.

Figure [*]a presents the forward-scattered P-S migration result for the SV component shown in Figure [*]c where the source is incident from the right. Both the crust-mantle (white-to-gray in model) and relict crust-mantle (black-to-gray in model) discontinuities are well imaged, but the crust-relict crust (white-to-black in model) boundary is notably absent due to weak amplitudes of incidence-angle dependent P-S conversions. Figure [*]b presents the backscattered P-P migration result for the P-wave component presented in Figure [*]d. The structure imaged in this case is predominantly the crust-mantle discontinuity. This is expected since the magnitude of the P-P reflection coefficient for the mantle-relict crust discontinuity is very weak as the change in P-wave velocity is less than two percent. Also note the vertical mis-positioning of structure in Figures [*]a and [*]b arising due to use of a v(z) reference model.

 
Figure5
Figure5
Figure 6
Results for applying migration algorithm to data shown in Figures [*]c and [*]d. (a) Forward-scattered P-S image; (b) Backscattered P-P image.
view burn build edit restore

Figure [*] presents the results of stacking the migrations of all 6 generated events for 4 of the 6 primary elastic scattering modes. Results show good reconstruction of the crust-mantle discontinuity. Panels (a), (c) and (d) also exhibit recovery of the relict crust, though with differing relative amplitudes. Forward-scattered P-S and backscattered P-P (Panels a) and b)) do not suffer from cross-mode contamination to the same degree as modes in panels c) and d). This is attributed to these two modes being the first scattered arrivals in their respective wave sections. These modes are themselves separated by the orthogonality of waves ending in P- and S-phases.

Forward-scattered P-P and backscattered S-P images are not shown since they are severely degraded. Forward-scattered P-P energy has little-to-no sensitivity to travel-time variations since the scattered wavefield moves at the same velocity and in the same direction as the source wavefield. This yields small-to-zero time delays insufficient for acceptable resolution of model structure. Backscattered S-P energy is twice converted and, over the range of angles considered here, has a much lower amplitude relative to the earlier arriving P-P energy.

 
Figure6
Figure6
Figure 7
Stack of all events for given scattering modes. a) P-S forward-scattering; b) P-P backscattering; c) P-S backscattering; and d) S-S backscattering.
view burn build edit restore


next up previous print clean
Next: Conclusions Up: Shragge: Teleseismic PSM Previous: Synthetic model and data
Stanford Exploration Project
7/8/2003