NMO and Deconvolution

 

nmo-ann
Figure 12 The upper panel shows a CMP gather after NMO correction. The near offset is to the left, the far offset to the right. The lower panel shows the CMP gather after deconvolution, bandpass filtering, and NMO correction.

Following the divergence correction, I perform a normal moveout correction (NMO) using the estimated RMS velocities. A resulting CMP gather can be seen in the upper panel of Figure . The near offsets are at the left side and increase to the right side. A significant change in shape and amplitude of the source wavelet is visible at the transition zone between the 100 m and 50 m group spacings. The loss in frequency and amplitude is caused by the strong hydrophone array attenuation. In the direction of the far offsets, where the hydrophone groups are shorter, the wavelet sharpens and regains its large amplitudes. These waveform shapes and amplitude changes can seriously degrade any subsequent AVO analysis. In order to regularize the wavelet as much as possible, I perform a single trace source wavelet deconvolution before NMO correction which estimates a different filter for each trace. Deep water (more than 3 km) and shallow dips in the area cause the wavefront propagation angles to change only moderately as a function of depth. For example, in Figure at offset 2.65 km ( transition between 100 m and 50 m groups spacing), the seafloor reflection propagates at 22$^{^{\circ}}$while the BSR reflection propagates at 20$^{^{\circ}}$. This small change in propagation angles makes trace-by-trace deconvolution a valid tool in this case. Although the application of single trace, as opposed to surface-consistent, deconvolution can result in undesired changes in amplitude and waveform, the Blake Outer Ridge seismic data are of high enough quality that single trace deconvolution with careful quality control gives very consistent results. After deconvolution, the data are bandpass filtered to a range of frequencies where signal is found at all offsets. This removes spurious deconvolutional high-frequency noise. Then the data are again NMO corrected.

A resulting CMP gather after deconvolution, bandpass filtering and normal moveout correction can be seen in the lower panel of Figure . The waveform has been well calibrated along all offsets. The remaining amplitude loss, especially that induced by the hydrophone attenuation, will be addressed in the amplitude calibration section.