|
<< Click to Display Table of Contents >> Navigation: Velocity > Interval Velocity Creation (3D data Dix 1D inversion) |
This module converts RMS (root-mean-square) velocity data into interval velocities for a 3D dataset, producing two output SEG-Y volumes simultaneously: one in the time domain and one in the depth domain. Interval velocity is the true velocity of each individual layer between two time or depth horizons, as opposed to RMS velocity which is an average influenced by all layers above a given time.
Two inversion methods are available. The Dix method applies the classic Dix formula sample-by-sample along each trace — it is fast and robust for smoothly varying velocity fields. The Inversion method performs a 1D iterative inversion at each bin location, working over a user-defined vertical calculation step; it can better handle velocity inversions and sharp layer contrasts but requires more computation time.
The module processes the full 3D inline/crossline grid but allows subsampling via inline and crossline skip parameters, which is useful for large surveys where a coarser interval velocity grid is sufficient for depth conversion or migration. Topographic corrections using a datum and replacement velocity are applied automatically to account for surface elevation variations. Both output files are written in SEG-Y format and can be directly used as input to depth conversion or migration velocity preparation workflows.
The collection of trace headers for the 3D dataset, carrying the inline and crossline bin identifiers and surface elevation (datum) values for each trace. The module uses this information to organize traces into the 3D inline/crossline grid, determine the extent of the survey, and apply per-trace topographic corrections. Binning must be applied to the dataset before running this module; an empty bin set will cause the module to stop with an error.
The link to the SEG-Y file containing the input RMS velocity field in the time domain. This file is read trace-by-trace during processing; the sample interval and trace length are determined automatically from the file header. Ensure that this input velocity volume is properly sorted and that its geometry matches the trace headers supplied above.
The full path and file name for the SEG-Y output volume that will contain the derived interval velocities in the time domain (ms). This file is created or overwritten when the module runs. Each output trace corresponds to one bin location in the 3D grid (taking into account any inline/crossline skip), with samples representing interval velocity values at successive two-way travel times. Accepted extensions are .sgy and .segy.
The full path and file name for the SEG-Y output volume that will contain the derived interval velocities in the depth domain (m). This file is generated alongside the time-domain output in a single processing run. Samples in each trace represent the interval velocity value at successive depth levels, which makes this volume directly usable as a velocity model for depth migration. Accepted extensions are .sgy and .segy.
The reference elevation (in metres) used as the flat datum plane for topographic corrections. When surface elevation varies across the survey area, traces from lower-elevation locations must be corrected so that all velocities are referenced to the same datum. If this value is left at its default of -99999.9, the module automatically determines the datum as the maximum receiver elevation found in the trace headers, which is the standard convention for a floating datum. Set an explicit positive value (for example, the mean sea level elevation of the survey area) to force a fixed reference plane.
The velocity (in m/s) assigned to the zone between the surface datum and the actual surface elevation of each trace location. Where the datum plane is above the real ground surface, the module fills the corresponding portion of the velocity trace with this constant replacement velocity instead of extrapolating from the data. The default value is 1500 m/s (approximate velocity of water or near-surface sediments). For land surveys with significant topography, use a value representative of the weathered layer or near-surface material, typically 1800–2500 m/s. The minimum allowed value is 150 m/s.
Controls the decimation of the output grid in the inline direction. A value of 1 outputs every inline; a value of 10 (the default) outputs every 10th inline. For large 3D surveys, using a skip value greater than 1 significantly reduces processing time and output file size while still providing a representative velocity grid for depth conversion. Set to 1 for a full-resolution output when bin-to-bin velocity accuracy is critical.
Controls the decimation of the output grid in the crossline direction. Works identically to Inline skip but applies across the crossline axis. The default value is 10. The combination of inline and crossline skip values determines the total density of the output velocity grid; for example, skip values of 10 in both directions produce a grid that is 100 times coarser than the full-resolution bin grid.
Selects the algorithm used to derive interval velocities from the input RMS velocity field. Two options are available:
Dix (default) — applies the Dix equation sample-by-sample. This is the standard, computationally efficient approach suitable for most surveys where the RMS velocity field is smooth and velocity generally increases with depth. The Calculation step parameter is disabled in this mode.
Inversion — performs a 1D iterative forward-modeling inversion at each bin location. The inversion works over vertical windows of the size defined by Calculation step and is better able to handle velocity inversions (decreasing velocity with depth) and sharp layer boundaries. Use this method when the Dix conversion produces physically unrealistic values such as very low or negative interval velocities.
Specifies the vertical window size (in ms) used by the Inversion method when computing interval velocities. This parameter is only active when Creation method is set to Inversion; it is grayed out and has no effect in Dix mode. The step is automatically rounded to the nearest multiple of the input data sample interval, and is forced to be at least three samples long. Smaller values produce finer vertical resolution in the interval velocity output but may amplify noise in the input RMS velocity field. Larger values smooth the interval velocity over a broader time window, which is more stable in noisy data. The default value is 1 ms; in practice, set this to a geologically meaningful layer thickness expressed in ms, such as 20–100 ms.
When enabled, this option prevents the computed interval velocity from decreasing too rapidly between consecutive depth/time samples. Sudden drops in interval velocity are common artefacts of noise or picking uncertainty in the input RMS velocity field, and they can lead to physically implausible models with very thin low-velocity zones. Enabling this option activates the Fraction of decrease parameter which controls the maximum allowed rate of velocity decrease. The default is disabled.
Controls the maximum fractional decrease of interval velocity permitted between successive samples when Limit decrease of VINT is enabled. The value is a dimensionless fraction between 0.01 and 1.0. A value closer to 1.0 allows very little decrease (stronger stabilisation), while a value closer to 0.0 permits large velocity decreases (weaker stabilisation). The default value is 0.99, which strongly suppresses velocity reversals. This parameter is grayed out and has no effect unless Limit decrease of VINT is checked.