time processing work flow
Signal Processing, is where the "heavy lifting" occurs to improve the signal-to-noise ratio (S/N) and enhance the resolution. This stage is highly iterative and requires careful testing to ensure the geology is preserved while the noise is removed. Here is the detailed breakdown of the Signal Processing steps (2.1 to 2.6):
1. Designature & De-bubbling (Marine) / Correlation (Land)
Goal: To remove the source signature and bubble effect, making the wavelet "zero-phase" and spikes.
- Detailed Parameters:
1. Filter Length: Typically 200–400 ms.
2. Prediction Distance (Lag): For de-bubbling.
Phase: Conversion from mixed-phase (raw) to zero-phase.
- Controlling Parameters/QC:
Signature Stability: Comparing near-field hydrophone records (marine) or sweep signatures (land).
Autocorrelation: Checking that the "side-lobes" and bubble pulses are removed.
- Issues to Check:
Phase instability; over-whitening that creates "ringing" artifacts; ghosting effects from the sea surface.
2. Noise Attenuation (Multi-Domain)
Goal: Removing specific types of noise like Ground Roll (land), Swell Noise (marine), or Random Noise.
- Detailed Parameters:
1. f-k Filter (Frequency-Wavenumber): Velocity cut-off for linear noise.
2. f-x Deconvolution: Prediction filter window size (e.g., 20 traces).
3. Tau-p Transform: Slowness range for coherent noise rejection.
- Controlling Parameters/QC:
1. Only Panels: The most critical QC—viewing exactly what was removed. If you see geological reflectors in the "noise-only" panel, you are losing signal. Spectral Analysis: Ensure low-frequency signal (5-10 Hz) isn't accidentally killed during ground-roll removal.
- Issues to Check: "Signal Leakage" (removing data with noise); "Smearing" of reflections; spatial aliasing.
3. Surface Consistent Deconvolution (SC-Decon)
Goal: To equalize the wavelet variations caused by different source and receiver locations.
Detailed Parameters:
1. Operator Length: (e.g., 120ms or 200ms).
2. Pre-whitening: (e.g., 0.1% to 1.0%) to stabilize the inversion.
3. White Noise Level: Prevents division by zero in frequency domain.
- Controlling Parameters/QC:
1. Four-Component Decomposition: Checking Source, Receiver, Offset, and CMP terms independently.
2. Autocorrelation Windows: Checking before/after deconvolution to ensure the wavelet is shortened (spiked).
3. Issues to Check: Boosting high-frequency noise; creating "notches" in the spectrum.
4. Refraction Statics (Land/OBN Specific)
Goal: Correcting for delays caused by the "weathering layer" (loose soil/sand near the surface). Detailed Parameters:
Weathering Velocity (Vw): Velocity of the top-most layer.
Replacement Velocity (Vrep): Velocity of the consolidated rock.
Datum Elevation: The reference level (e.g., Mean Sea Level).
Controlling Parameters/QC:
First Break Picks: The accuracy of the "picks" is the single most important control.
LMO Corrected Gathers: Check if the reflections are "wavy" (bad statics) or "smooth" (good statics).
Issues to Check: "Cycle-skipping" in picking; incorrect datum shifts leading to structural errors.
5. Velocity Analysis (NMO & RMS Velocity)
Goal: To determine the speed of seismic waves through layers to "flatten" reflections (NMO correction) and prepare for stacking.
Detailed Parameters:
Analysis Grid: Typically 500m x 500m or 1km x 1km (depending on structural complexity).
1. Semblance Window: Time gate (e.g., 40–60ms) and velocity range (e.g., 1400–6000 m/s).
2. Interpolation: Algorithms to create a continuous velocity volume from point picks.
Controlling Parameters/QC:
1. Semblance Sharpness: High-amplitude energy "bullseyes" on the semblance plot indicate high-confidence picks.
2. Gather Flatness: Visually checking the NMO-corrected gathers; if they "smile" (curved up) or "frown" (curved down), the velocity is wrong.
3. Interval Velocity ($V_{int}$): Calculated via Dix Equation; must be geologically plausible (no sudden jumps to 10,000 m/s).
Issues to Check:
1. Multiples: Picking multiple energy instead of primaries (usually lower velocity).
2. Dipping Events: Dipping reflectors appear with higher "apparent" velocity, which can smear the stack if not handled by DMO or Migration.
6. Anisotropy Analysis (Fourth-Order / ETA Correction)
Goal: To correct for the fact that seismic waves travel at different speeds vertically vs. horizontally (VTI), which causes "hockey-stick" reflections at far offsets.
Detailed Parameters:
ETA (h) Parameter: A dimensionless value representing the degree of anisotropy.
Offset Range: Usually requires "Long Offsets" (offset > depth) to accurately calculate.
Non-Hyperbolic Moveout Equation: Using higher-order terms beyond standard NMO.
Controlling Parameters/QC:
Far-Offset Flatness: The primary QC—ensuring the "tail" of the reflection (the hockey stick) is flattened and aligned with the near-offset.
Stack Resolution: Anisotropy correction significantly improves the frequency and sharpness of the final stack, especially for deep targets.
Issues to Check:
V-ETA Ambiguity: Velocity and Eta are coupled; an error in velocity can be "hidden" by an incorrect Eta pick, leading to structural depth errors later.
7. Multiple Removal (Demultiple)
Goal: Removing "ghost" reflections that bounce between layers (SRME/Radon).
Detailed Parameters:
1. Radon Moveout: Defining the velocity separation between primaries and multiples.
2. SRME Aperture: The spatial range for predicting surface multiples.
Controlling Parameters/QC:
1. Velocity Gathers: Checking if the "curved" events (multiples) are gone, leaving the "flat" events (primaries).
2. Autocorrelograms: Confirming the periodic energy is removed.
Issues to Check:
3. Overlap between primary and multiple (killing the signal); loss of near-offset data.
8. 5D Interpolation & Regularization
Goal: Filling in missing traces (dead shots/receivers) to create a uniform grid for imaging. Detailed Parameters:
1. Spatial Bandwidth: Maximum frequency to interpolate.
2. Number of Iterations: (e.g., 50-100 iterations).
Controlling Parameters/QC:
1. Missing Shot Maps: Comparing "Before" vs "After" fold maps.
2. Stack Sections: Checking if "holes" in the section are filled realistically without looking "smeared."
3. Issues to Check: Creating "synthetic-looking" data; artifacts in areas of high structural dip.