3-D Migrations

I ran standard and anti-aliased Kirchhoff 3-D poststack time migrations on the Halliburton data subset. The standard migration corresponds to the left impulse response of Figure , i.e., a 45 degree aperture and $\cos^2\theta$ obliquity dip-filtering. The anti-aliased migration corresponds to the right impulse response of Figure , which has the additional anti-aliasing local triangle filtering applied as described previously. The migrations take approximately 20 CPU hours per run for the standard migration, and 30 CPU hours for the anti-aliased migration, on the SEP 32-node CM5. In this section, I will briefly compare the migrated images obtained by both methods.

Figure shows inline sections sliced from the standard (upper) and anti-aliased (lower) migration volumes, at a constant y=0.225 km. Aliasing of the type shown in the left panel of Figure has severely contaminated the upper migration image. Even the horizontal structure has been degraded by steep dip spatial aliasing of the migration operator. The anti-aliased migration (lower) is much more coherent in terms of horizontal structure, and possibly images two salt-sediment interfaces. Also, the anti-aliased image clearly shows shallow faulting above the interpreted salt location, including a slightly clockwise-rotated fault block at 9.5 km and 1.5 seconds.

Figure shows inline sections sliced from the standard (upper) and anti-aliased (lower) migration volumes, at a constant y=0.875 km. Again, the anti-aliased image is much more coherent in all aspects, although visually a little more low-frequency in content. The third inline section of Figure is at a constant y=2.300 km. Here, the anti-aliased image clearly shows the salt peak, as well as the intersection between the salt and anticlinal sediments where the salt appears to have pierced through. Also, some near vertical faulting is apparent directly above the salt peak at the 8 and 9 km locations. These features are barely visible in the standard migration, and would be much more challenging to interpret without the aid of the anti-aliased image.

Figure shows a crossline section sliced from each migrated volume at a constant x=7.1625 km location. Note the coherent resolution of the anticlinal sediment structure in the anti-aliased migration (right panel). With a little imagination, we might be seeing the top of salt which underlays the sediment package at 2.5 seconds, and the fuzzy bottom of the salt intrusion at 3.0 seconds. Below the interpreted top of salt, a depth migration with salt and sediment velocity specification is probably required to focus the deeper reflection events.

Figure shows a time slice from each of the standard (upper) and anti-aliased (lower) migration volumes, at a constant $\tau=0.128$ seconds. The anti-aliased time slice shows a 3 km long continuous radial fault associated with the salt intrusion. Since these time slices are essentially at the surface, the salt approximately lies between and underneath the two pseudo-parallel radial faults, and the salt peak penetrates the sediment anticline beneath the lower right corner of the plot. The anti-aliased migration image clearly shows this fault continuity and the fact that it penetrates through to the surface, in comparison to the more ambiguous result from the standard migration time slice in the upper panel.

Figure shows a time slice from each of the standard (upper) and anti-aliased (lower) migration volumes, at a constant $\tau=1.752$ seconds. In the anti-aliased image, the salt peak has clearly penetrated the anticlinal sediment structure in the lower right corner of the plot. A radial fault (salt-filled?) is barely discernible, but definitely seems to curve into the major discontinuity in folded sediments at $\{x,y\}=\{8.5,1.7\}$ km. The second radial fault is not well imaged at this pseudodepth. Note the contrast between the coherency of the upper and lower panels.

Figure shows time slices at a constant $\tau=1.880$seconds. In the anti-aliased migration image, both radial faults seem to be clearly imaged, but almost absent from the standard migration panel. The fault reflections are fairly strong amplitude, perhaps indicating they are salt-filled. The two faults bound a fault block which contains reflectors that dip steeply and acutely into the plane of the page, especially at x=9.5 km in the y=1.5 km range. In contrast, the anticlinal sediment structure to the left dips more within the horizontal plane of the page.