NMO CORRECTION AND SPECTRA OF MISSING TRACES

One of the most important considerations in interpolation is to avoid interpolating aliased events. It is very difficult to distinguish steeply dipping events from aliased data. NMO processing provides a good means of moving the steeply dipping events to the nonaliased dip range of the frequency-wavenumber domain. Figure 2 shows the spectrum of a CMP gather before and after NMO correction. Notice that the NMO correction has significantly reduced the steep dips.

 

fig2
Figure 2 (a) Synthetic CMP gather, (b) amplitude spectrum of the CMP gather, and (c) amplitude spectrum of the NMO-corrected CMP gather.

If there existed any missing trace, its spectrum would spread over a wide wavenumber range. Figure 3 shows three amplitude spectra, which correspond to the interlaced missing, truncated, and randomly missing CMP gathers, respectively. Among the three spectra, only the amplitude spectrum of the interlaced missing CMP gather clearly indicates that a conventional frequency-wavenumber domain interpolation method can be applied ( i.e., Fourier transforming the data, applying a high-cut filter, and inverse Fourier transforming). The others, the amplitude spectra of truncated and randomly missing CMP gathers, are blurred by the missing traces, and it is difficult to find a clear-cut passband.

 

fig3
Figure 3 The amplitude spectrums of NMO corrected CMP gathers of (a) the interlaced missing, (b) the truncated missing and (c) the randomly missing traces.