Seismic synthetic data sets modeling and processing

Two seismic surveys were computed over a $2 \times 3$ km central block of the modeled reservoir. The first survey was recorded at the beginning of production (t₀), and the second one was recorded after 3 years of production, and two years and a half of water injection (t₁). The data sets covered an area $6 \times 6$ km, with 30 m spacing between common-midpoints (CMP's) in both directions. Each data set contains two different offsets (source-receiver distance): zero offset and 3 km offset. Multi-offset surveys allow the analysis of amplitude variation with offset (AVO), which can be useful for identifying variations in the fluid type. The background velocity function was linearly increasing with depth, starting from a velocity of $2\;{\rm km/s}$ and a constant gradient of $.28\;{\rm s^{-1}}$.

The wavelet used in the modeling had a maximum frequency of about 45 Hz, giving a wavelength of about 50 m in the target. This is reasonable, but a lower limit on the resolution expected for a target of this depth. Further modeling will be done with higher frequencies.

The modeled data was imaged using a Kirchhoff 3-D prestack migration program. The kinematics and amplitudes used by the migration were obtained by numerically tracing rays through the given background velocity function. The migration code had the appropriate anti-aliasing corrections and amplitude tapering to preserve the correct amplitudes. The two different offsets ( and 3 km) present in the data sets were migrated and analyzed separately.