``Under-migrated'' effect

The second problem which can arise involves the completeness of our migration with respect to frequency. The range of $\omega$ needed to cycle through is a function the data spectrum. Figure 3 shows a CMP gather in ($\omega$,k_h) domain. To be correct, our migration must cycle through the top and bottom lines in the plot. However, we must be careful not to cycle too high, as temporal aliasing is still a factor (maximum frequency can be no higher than the Nyquist, $\frac{1}{2 \Delta t}$).

 

spect Figure 3 A CMP in ($\omega$,kh) domain. We must cycle equal to or past $\omega$at the line, but not past temporal Nyquist.