|
<< Click to Display Table of Contents >> Navigation: Statics > Refraction statics solver - model-free |
Refraction statics solver - model-free computes surface-consistent static corrections from first-break pick times without constructing an explicit weathering layer velocity model. Instead of building a near-surface model (as a layer-based solver does), this module directly decomposes the observed first-break time anomalies into source and receiver components using an iterative least-squares approach. The result is a set of static shift values for each source and receiver location that, when applied to the seismic data, flatten the refraction arrivals and improve the quality of subsequent stacking and imaging.
The solver uses first-break picks provided by a preceding automated or manual pick module. It iterates to minimise the residual between the measured pick times and the time predicted by the current source-plus-receiver static model, applying an optional outlier rejection pass at each cycle. The results are saved as a static correction item and displayed on two quality-control graphics: a map of average solver error per source/receiver location and a graph of convergence error versus iteration count.
Note: This module is deprecated and is retained for legacy workflow compatibility only. For new projects, use the current Refraction statics workflow.
The first-break picking result item produced by a preceding refraction picking module. This item stores the first-arrival times for every source-receiver pair together with the associated trace geometry. The solver reads these pick times and the corresponding source and receiver coordinates to set up the linear system for static decomposition. The module will report an error if this item is invalid or if no picks have been made yet.
The maximum number of solver iterations used to converge the least-squares decomposition of source and receiver statics. Default: 50. Minimum: 1. In each iteration the solver updates source and receiver statics based on the residual misfit and then re-estimates the reference delay time function. The process converges when the residual stops decreasing. Most datasets converge within 20–50 iterations. Inspect the Convergence graphic to confirm that the error has plateaued before accepting the result; if the error is still decreasing at 50 iterations, increase this value.
The number of outlier rejection passes performed inside each main iteration. Default: 2. Minimum: 1. After updating the static estimates in each iteration, the solver flags pick pairs whose residual time anomaly exceeds a threshold as outliers and excludes them from the subsequent update. Running multiple outlier rejection passes per iteration improves robustness against erroneous picks (for example, picks on multiples or direct waves rather than the refraction). Increasing this value slows the computation but makes the solver more tolerant of noisy picks.
The offset bin size (in meters) used to group traces when computing the reciprocal delay time function for the solver. Default: 50 m. Minimum: 1 m. Traces with source-receiver offsets falling within the same bin are averaged together to compute a representative time function for that offset range. A smaller increment gives a denser offset sampling and a more detailed refraction velocity analysis, but requires more picks per bin to be statistically stable. Use a value comparable to the nominal receiver group interval of your survey.
The minimum source-receiver offset (in meters) included in the solver calculation. Default: -1 m (use all offsets). When set to -1, the module automatically uses the minimum offset present in the picks. Set a positive value to exclude very short-offset picks that may be contaminated by direct wave interference or where the refraction is not yet the first arrival. This value must be less than MaxOffset if both are specified.
The maximum source-receiver offset (in meters) included in the solver calculation. Default: -1 m (use all offsets). When set to -1, the module automatically uses the maximum offset present in the picks. Set a positive value to exclude long-offset picks where the refraction path is too deep to be relevant for weathering static corrections, or where pick quality degrades. For most land surveys a practical upper limit is 2–5 times the expected depth to the refractor.
The replacement velocity (in m/s) used to convert the solved delay times into static shift values referenced to the processing datum. Default: 1500 m/s. Minimum: 1 m/s. After the solver decomposes the pick times into source and receiver components, these components are combined with the surface elevation and the datum elevation to compute the total static shift for each location. The static shift is the elevation correction divided by this velocity. Set V0 to the replacement velocity used in your processing datum scheme — typically the same value as used in the datum static correction module that follows this solver.
When enabled, the solver accepts multiple first-break picks per trace (for example, picks on several distinct refraction branches corresponding to different refractor depths). Default: off. When disabled, only the primary (earliest) pick for each trace is used. Enable this option when the picking dataset contains multi-branch picks and you want the solver to account for all branches simultaneously, which can improve accuracy in areas with laterally varying refractor geometry.