Correlation window
![]() | |
Series | Investigations in Geophysics |
---|---|
Author | Öz Yilmaz |
DOI | http://dx.doi.org/10.1190/1.9781560801580 |
ISBN | ISBN 978-1-56080-094-1 |
Store | SEG Online Store |
From the same field data example (Figure 3.3-5), the results of residual statics corrections (for the right half of the stacked section) are examined by using different correlation windows while keeping all other parameters constant. The maximum allowable shift was 80 ms in these tests. From Figure 3.3-29, note that a correlation window confined to the mute zone (400 to 1200 ms) is not desirable. It does not provide sufficient statistics because of the low fold of coverage and the shortness of the data window available for crosscorrelation with the pilot traces.
With high-fold data, the mute zone problem is handled to a degree by limiting the correlation to small offsets. In this particular part of the profile, a large window including both the mute zone and deep data (800 to 2300 ms), a deep window (1700 to 2300 ms), a deep large window (1400 to 2800 ms), or a deep narrow window (1500 to 1700 ms) made no difference. This is probably because of the good signal-to-noise ratio in this part of the profile.
These observations are verified by the diagnostics based on the CSP and CRP stacks in Figures 3.3-30 and 3.3-31. Ungained stack responses are shown in Figure 3.3-32. In particular, note the relatively poor stack response using a window confined to the mute zone (400 to 1200 ms).
Figure 3.3-5 CMP stacks derived from gathers in Figures 3.3-2 and 3.3-3. Stack (a) before and (b) after residual statics corrections. NMO correction was applied using preliminary velocity picks derived from the spectra in Figure 3.3-4.
Figure 3.3-29 Test of correlation window: CMP gathers after residual statics corrections using five different correlation windows. Figure 3.3-32 shows the CMP stacks.
Figure 3.3-32 Test of correlation window: CMP stacks after residual statics corrections using five different correlation windows. CMP gathers are shown in Figure 3.3-29.
The choice of the correlation window is more critical in conditions of poor signal-to-noise ratio. Refer to the same diagnostics for the left half of the stacked section in Figure 3.3-5. These diagnostics are shown in Figures 3.3-33 through 3.3-36. Again, a correlation window confined to the mute zone not only provides an inadequate solution, as in the previous case (Figure 3.3-29), but also can be devastating, as shown in Figure 3.3-33. In this case, the CMP gathers with no corrections have better signal quality. It now is apparent that a narrow window, even if it is outside the mute zone (such as the 1500 to 1700 ms window), may not provide sufficient statistics. The CSP stacks (Figure 3.3-34) and the CRP stacks (Figure 3.3-35) show the undesirable aspects of choosing a window within the mute zone or choosing a window that is too narrow. The ungained stacked sections (Figure 3.3-36) clearly demonstrate the adverse effects of an improper choice of correlation window. Moreover, note the poor quality pilot traces for each CMP gather in Figure 3.3-37 (to the left of midpoint 377). The shot and receiver static solutions shown in the graphs above the pilot traces are totally unreliable.
Figure 3.3-33 Test of correlation window: CMP gathers after residual statics corrections using five different correlation windows. CMP stacks are shown in Figure 3.3-36.
Figure 3.3-36 Test of correlation window: CMP stacks after residual statics corrections using five different correlation windows. CMP gathers are shown in Figure 3.3-33.
These test results suggest choosing a correlation window that (a) contains as much signal as possible to improve correlation values, and (b) is large enough and is outside the mute zone whenever possible.
See also
- Residual statics estimation by traveltime decomposition
- Residual statics estimation by stack-power maximization
- Traveltime decomposition in practice
- Maximum allowable shift
- Other considerations
- Stack-power maximization in practice
- Exercises
- Topics in moveout and statics corrections