SYNT
Figure 1 Anisotropic synthetic data: zero-offset data section. The velocity is represented by a strong vertical and lateral linear gradient of 0.5 s^-1.

 

SUNTZEROisomig
Figure 2 Extended split-step Zero-offset isotropic migration of the anisotropic synthetic data (Figure ). In this case ${\eta=\epsilon=0.2}$ and ${\delta=0}$.

 

SUNTZEROANISOmig
Figure 3 Extended split-step zero-offset anisotropic migration of the synthetic data modeled with ${\eta=0.1}$ (Figure ). In this case ${\eta=\epsilon=0.2}$ and ${\delta=0}$.

 

SYNTPRE
Figure 4 Extended split-step anisotropic prestack migration of the anisotropic synthetic data (Figure ). This prestack imaged is obtained by using 5 reference velocities and constant ${\eta=0.2}$.

 

marmvel
Figure 5 Anisotropic synthetic data: Marmousi velocity model (focusing velocity) (, ).

 

MAR15TIPRE
Figure 6 Anisotropic prestack depth migration using 5 reference velocities and 5 reference $\eta$'s.

 

 

Azimuth moveout vs. dip moveout in inhomogeneous media

Biondo L. Biondi

biondo@sep.stanford.edu

ABSTRACT

Dip moveout (DMO) is often applied to prestack data to better preserve dipping events when performing partial stacks over ranges of offset. The tests presented in this paper, conducted on the SEG-EAGE salt data set, indicate that the application of azimuth moveout (AMO) in place of constant-velocity DMO yields better partial stacks in two important cases: first, when the velocity increases with depth. Second, when a salt body causes NMO-velocity conflicts between deeper flat reflectors and shallower dipping reflectors.

AMO is less sensitive to velocity variations than DMO because it is a residual operator, and thus it has less tendency to overcorrect reflections that have been moved out with too high NMO velocity. AMO has also the potential advantage over V(z) DMOs to be less sensitive to the given velocity function.