Peg-Leg Multiple

Key Takeaways

• Peg-leg multiples arrive at intermediate times that do not align with simple multiples of primary reflection times — the defining characteristic that separates peg-legs from full water-bottom multiples is their timing; a full water-bottom multiple arrives at exactly twice the two-way travel time to the water bottom, and its harmonics (second, third order) at three and four times that time; a peg-leg involving a reflection from a reflector at two-thirds of the water-bottom depth arrives at (2/3 + 1 + 2/3) = 2.33 times the water-bottom primary time — a non-integer multiple that falls between the easily recognized harmonic multiples and masquerades as a legitimate deep reflector; this timing ambiguity makes peg-legs one of the most interpretation-hazardous types of seismic noise, because an interpreter who does not specifically check for peg-leg contamination may map what they believe is a deep reservoir target that is actually the ghost image of a shallow reflector bouncing around the water column.
• Normal moveout (NMO) velocity discrimination is the primary tool for separating peg-leg multiples from primaries — all multiple reflections (including peg-legs) have a lower NMO velocity than primary reflections at the same two-way time, because multiples travel longer ray paths in the shallower, slower part of the section even though their arrival time mimics a deeper (faster) primary; this velocity difference creates a characteristic that can be exploited in common midpoint (CMP) gather processing: applying NMO correction using primary velocities will over-correct multiples (pulling their arrival time too far up) while correctly flattening primaries, and the resulting difference in moveout behavior after NMO allows multiples to be attenuated by stacking (constructive interference for correctly flattened primaries, destructive interference for over-corrected multiples); sophisticated multiple suppression algorithms including parabolic Radon transforms (velocity filters that operate in the tau-p domain) and extended SRME (surface-related multiple elimination) explicitly exploit this velocity contrast to separate and remove peg-leg multiple energy from the data.
• SRME (surface-related multiple elimination) can predict and subtract peg-legs from the seismic record — the most powerful data-driven multiple suppression method, SRME, works by exploiting the fact that surface-related multiples (including water-bottom peg-legs) can be predicted by auto-convolution of the seismic data with itself; the mathematical basis is that any surface-related multiple is composed of the same primary reflections convolved together, and the seismic data itself contains all the primaries needed to reconstruct the multiples; SRME creates a "multiple model" by correlating the data with itself at appropriate time shifts corresponding to the predicted peg-leg geometry, then adaptively subtracts this model from the original data; the method works without any knowledge of the subsurface velocity structure, making it particularly valuable in structurally complex areas where velocity models are unreliable; extended SRME handles 3D acquisition geometries and near-offset extrapolation to predict the full range of peg-leg multiple types present in marine data.
• Inter-bed multiples between shallow reflectors are the most difficult peg-leg category to suppress — while water-bottom peg-legs are the dominant multiple type in marine data, inter-bed multiples generated entirely within the subsurface (where strong reflectors such as shallow evaporite horizons, basalt flows, or hard carbonate stringers act as internal reflectors for peg-leg paths) are harder to predict because they involve the seismic energy bouncing between reflectors that are not easily identified from the surface; these internal peg-legs can appear at any time in the seismic section and are not as systematically predictable as water-bottom multiples; demultiple methods that rely on knowing the reflector causing the multiple bounce (model-based methods) are needed for inter-bed peg-legs, requiring either detailed velocity models from full waveform inversion or integration with borehole data to identify the specific shallow reflectors generating the problematic peg-leg energy.
• Seismic interpreters use wave equation modeling to identify peg-leg contamination in prospect areas — before relying on a seismic reflection for a well location decision, interpreters routinely model the expected arrivals of all multiples (including peg-legs) from the major reflectors in the section to check whether the event of interest could be a multiple rather than a primary; this is done by computing synthetic shot records that include all orders of multiple reflections using the known velocity model and reflectivity sequence from nearby wells; if a modeled peg-leg from a known shallow reflector arrives at the same time and moveout as the prospective reflection in the real data, the interpreter must treat the event as potentially contaminated and apply additional multiple suppression or acquisition design changes before committing to a well decision; in frontier basins without well control, peg-leg modeling from shallow seismic data alone provides the best available guide to multiple contamination risk.