Weighted stacking of seismic AVO data using hybrid AB semblance and local similarity

The hybrid framework of AB semblance and local similarity for optimized stacking of AVO data

Although work has been done on weighted stacking with local similarity (Sanchis and Hanssen, 2011), strong amplitude variations in the reflection events were usually ignored in both NMO velocity analysis and CMP stacking. In the tests using the synthetic and field data with class II AVO polarity-reversal anomalies, we choose a near-offset trace as the reference trace, not the equal-weight stacked trace in Liu et al. (2009) because the near-offset trace in this type of seismic data preserves the useful energy and usually has the highest SNR among all traces in the same CMP gather. The near-offset trace can effectively suppress the negative-stacking contribution from polarity-reversal traces. But the equal-weight stacked trace does not have correct or meaningful reflection signals due to the compromise of positive and negative polarities. It is worth mentioning that if the equal-weight stacked trace is chosen as the reference trace, positive and negative waveforms in the other traces would get the same weights, which makes weighted stacking deteriorate. Thus, the near-offset reference trace is capable of optimal weights regarding the amplitude variations and SNR. A workflow of weighted stacking using a hybrid framework of AB semblance and local similarity is shown in Figure 2.

workflow
Figure 2. Workflow of weighted stacking using hybrid AB semblance and local similarity.

Weighted stacking of seismic AVO data using hybrid AB semblance and local similarity