|
<< Click to Display Table of Contents >> Navigation: Modeling > Reflection Tomo 2D/3D |
The Reflection Tomo 2D/3D module builds a near-surface depth velocity model by performing reflection tomographic inversion. Starting from an RMS time-migrated velocity field, the module iteratively updates a depth velocity model so that modelled reflection travel times match the observed moveout across offsets. The inversion uses Fresnel-zone weighting to constrain the velocity updates to the resolution limit of the seismic data.
Use this module when refraction-based statics methods are insufficient and you need a depth velocity model consistent with reflection moveout. The module supports both 2D line and full 3D survey geometries by toggling the 3D case flag.
Boolean flag. When enabled the module treats the dataset as a 3D survey with both inline and crossline axes. Disable for 2D processing along a single line.
Link to the RMS velocity cube derived from time-domain migration velocity analysis. This field drives the conversion to depth and is used as the reference moveout model for the tomographic inversion.
Container defining the spatial extent and sampling of the output depth velocity grid. Set the minimum, maximum, and step values in both X and Y to cover the survey area with the desired resolution.
Minimum inline coordinate of the output velocity grid.
Maximum inline coordinate of the output velocity grid.
Inline grid spacing of the output velocity cube. Default: 25 m.
Minimum crossline coordinate of the output velocity grid.
Maximum crossline coordinate of the output velocity grid.
Crossline grid spacing of the output velocity cube. Default: 25 m.
Container group controlling the geometry and parameterization of the output depth velocity model.
Maximum velocity (m/s) expected for the deepest refractor in the near-surface model. Default: 6000 m/s. This value caps the velocity of the lowest model layer and prevents unrealistic high-velocity solutions during the iterative update.
Reference datum elevation (m) to which statics corrections will be computed. Default: 1000 m. Set this to the floating datum or final datum elevation used in your processing flow.
Maximum depth (m) of the output velocity model below the datum. Default: 1500 m. The model will be built from the surface down to this depth. Set this to the base of the near-surface zone you wish to characterize.
Vertical sample interval (m) of the output depth velocity model. Default: 10 m. Finer intervals increase output resolution but also increase memory requirements.
Inline spatial sampling interval (m) of the output velocity cube. Default: 200 m. This controls the lateral resolution of the tomographic velocity model along the inline direction.
Crossline spatial sampling interval (m) of the output velocity cube. Default: 200 m.
Inline grid spacing (m) of the internal travel-time table used for ray tracing. Default: 25 m. Finer spacing improves travel-time accuracy but increases memory and computation time.
Crossline grid spacing (m) of the internal travel-time table. Default: 25 m.
Radius (m) of the lateral smoothing applied to the velocity update at each iteration. Default: 200 m. Increasing this value produces a smoother model that is more geologically plausible but less sensitive to local velocity variations. Use a value comparable to the expected spatial wavelength of near-surface velocity anomalies.
Half-width (m) of the Fresnel zone used to weight ray contributions to the velocity update. Default: 10 m. A wider Fresnel zone increases stability of the inversion by averaging more ray paths but reduces lateral resolution.
Dominant frequency (Hz) of the seismic data used to estimate the Fresnel zone width during inversion. Default: 5 Hz. Set this to the dominant frequency of the reflections used for picking. This controls the theoretical resolution limit of the velocity update.
Lateral window length (m) over which a low-frequency trend is estimated and removed from the picks before inversion. Default: 2000 m. A larger window captures only very long-wavelength trends; a smaller window removes shorter-scale background variations.
Vertical (time or depth) window length (m) over which the low-frequency trend is computed. Default: 100 m.
Number of outer inversion iterations in which the full velocity model is rebuilt and the complete ray set is re-traced from scratch. Default: 5. More iterations improve convergence but increase computation time. Monitor the residual misfit to decide when to stop.
Number of inner iterations applied to a local region around each gather before updating the global model. Default: 5. Increasing local iterations improves fitting of fine-scale velocity variations.
When enabled, allows the inversion to produce velocity inversions (velocity decreasing with depth) in the model. Default: disabled. Enable only when geological evidence supports the presence of low-velocity zones.
Lower bound on the velocity values permitted in the inverted model (m/s). Default: 1000 m/s. Prevents the inversion from generating physically unrealistic low-velocity solutions.
Upper bound on the velocity values permitted in the inverted model (m/s). Default: 7000 m/s. Prevents the inversion from producing geologically unrealistic high-velocity layers.
Minimum source-receiver offset (m) of traces used in the inversion. Default: 0 m. Exclude near-offset traces if they are dominated by direct-wave arrivals or surface waves that would corrupt the reflection moveout analysis.
Maximum source-receiver offset (m) of traces used in the inversion. Default: 5000 m. Traces beyond this offset are excluded to avoid stretch artefacts and NMO cycle skipping at large offsets.
Maximum allowable difference (m) between the observed pick position and the position predicted by the current velocity model. Picks that deviate more than this threshold are treated as outliers and excluded from the inversion update. Default: 100 m. Reduce this value to reject noisy or erroneous picks more aggressively.
Container group with parameters controlling the residual statics solver applied after the tomographic velocity update.
When enabled, computes residual static corrections after each global iteration to account for pick errors and small near-surface anomalies not resolved by the tomographic velocity model.
Number of global solver iterations used during the residual statics estimation step. Default: 2.
Number of local solver iterations for the residual statics estimation. Default: 2.
Number of azimuthal directions used to stratify the offset-azimuth distribution during the inversion. Default: 1 (azimuth-independent). Increase for 3D surveys with strong azimuthal anisotropy to improve the accuracy of the velocity update in different offset-azimuth sectors.
Lateral search radius (m) used to estimate the initial near-surface velocity (V0) from nearby picks before starting the full tomographic iterations. Default: 100 m. Increasing this value stabilizes the initial V0 estimate in areas with sparse pick coverage.