Continuous time-varying Q-factor estimation method in the time-frequency domain

Field data

For the field data test, we select a 2D poststack seismic section (Figure 11a) that contains 201 seismic traces, each of which has a time length of 3 s and a time sampling interval of 4 ms. For comparison, we perform the automatic gain control (AGC) processing on the original seismic section. The processed result after AGC is shown in Figure 11b. AGC can enhance the weak amplitude of the deep part, but it cannot improve the resolution of the deep data. Next, we estimate the time-varying equivalent Q-factors for each seismic trace using the LCFS method and select the time corresponding to the maximum value of the local centroid frequency as the reference time. The estimated equivalent Q profile is smoothed to some extent in the spatial direction to preserve the lateral continuity. Figure 11c shows that the Q-factors calculated using the LCFS method have a reasonable distribution range and the characteristics of continuous variation in the direction of time and space. Figure 11d shows the seismic section after inverse Q filtering based on the time-varying equivalent Q-factors. Comparing Figures 11a,  11b, and  11d, we can see that the resolution of the deep part of the seismic data after inverse Q filtering has improved, the absorption and attenuation of the formation are reasonably compensated, and the original structural characteristics are well maintained.

powdata,agcpowdata,smqts,cmsig
Figure 11. Processing result of field data. Poststack field data (a), the seismic data after AGC (b), the estimated time-varying equivalent Q-factors (c), the seismic data after inverse Q filtering (d).

Continuous time-varying Q-factor estimation method in the time-frequency domain