Strong lateral velocity anomalies
are challenging for one-parameter residual-moveout analysis.
The migrated ADCIGs may display complex moveouts
that are not accurately described by a a one-parameter family of curves.
Furthermore, when the residual velocity error is large enough,
the moveout at far angles may be larger than the dominant wavelength,
and thus stack-power spectra may display local maxima.
The negative effects of these phenomena can be avoided by imaging
only the low-frequency components of the data,
when they are available.
A more general solution is to smooth the stack-power function along
the residual-moveout parameter axis.