We examined the sensitivity of the AVO response due to the presence
of a overburden with complex velocity anomalies using a synthetic
data set.
We observed that AVO attributes calculated after prestack depth
migration using the true velocity model are sensitive to the velocity
anomalies.
Introducing errors in the migration-velocity, we found that the AVO gradient attribute
is much more sensitive to velocity errors than AVO intercept attribute.
For velocity errors up to 5%, we can see a maximum of AVO intercept errors
of , whereas for velocity errors of only 1%, the inversion of
AVO gradient attribute has an error of 185%.
These results are specific for the synthetic data used; different
results could be obtained by modeling different velocity anomalies.
We observed some boundary artifacts in the modeled data and we noted that amplitude values after migration are more sensitive to these boundary artifacts than amplitude values before migration. These boundary artifacts become worse when we introduce velocity errors in the migration-velocity. We need to do further work to evaluate the influence of boundary artifacts on the amplitudes; we also would like to compare the results using other migration methods, such as Kirchhoff prestack migration.