Depth Imaging & Velocity Model Building

---

Where structures are complex, lateral velocity changes are strong or targets are small, time processing cannot place reflectors correctly and PSTM cannot fully correct mispositioning. Pre-stack depth migration corrects the image distortions caused by steep strata and velocity changes.

Kirchhoff pre-stack depth migration, reverse time migration (RTM), Gaussian beam migration, wave-equation (eGWM) and generalised Radon (GRT) migration, anisotropic VTI/TTI depth imaging, and velocity model building with Constrained Velocity Inversion (CVI) and grid tomography.

Horizon interpretation is the basis of the model. Working with the geologists, horizons are tracked from shallow to deep on continuous events, preferably velocity interfaces. We build a structural model with TTI anisotropy where needed, use velocities that belong to individual rock units, and test model scenarios against seismic diagnostics and geological input.

Image fidelity is measured objectively as a function of velocity, not judged by eye: residual moveout and local stack power on the migrated gathers, and event coherence and image-simplicity measures such as Varimax on the image cube. Structural velocity-consistency constraints are applied throughout.

A full onshore 3D project from initial model to final tomographic update. The initial model was built from PSTM velocities with CVI rather than a Dix conversion. Two iterations of grid tomography followed, with residual-moveout picking on a 1000 m x 500 m grid and travel-time errors measured up to 35 degrees, steered by dip, azimuth and continuity attributes so the model followed the geology.

Moving from PSTM to PSDM to RTM, geometric accuracy and sensitivity to the velocity model both increase. Where the model is uncertain, PSTM and PSDM give complementary views, while RTM suffers from model error. Anisotropic depth migration gives the most accurate positioning, but only as accurate as the velocity model, which is why geological input matters.