Multiple Reflection | Oil Authority

A multiple reflection in seismic exploration is an unwanted reflected wave that has undergone more than one reflection in the subsurface before being recorded at the surface receiver — the seismic energy bounces between two or more reflectors (or between a reflector and the surface) multiple times, arriving at the receiver after the primary reflection from the same depth and adding a spurious signal to the seismic record that, if misinterpreted as a primary reflection from a deeper interface, would indicate a false subsurface layer; multiple reflections are one of the most significant sources of seismic noise in petroleum exploration, particularly in marine environments where the sea floor and water surface create a strong horizontal reflector pair that generates intense short-period multiples, and in sub-salt environments where the high-impedance salt base generates long-period peg-leg multiples that can obscure primary reflections from sub-salt reservoirs.

Key Takeaways

Multiple classification by reflection path distinguishes two main categories: free-surface multiples, which involve at least one reflection from the Earth's surface (or the sea surface in marine acquisition), and interbed multiples, which bounce entirely within the subsurface between two reflectors without returning to the surface; free-surface multiples include water-bottom multiples (sea floor reflection followed by water surface reflection, creating a mirror-image replica of the seafloor reflection at approximately twice the water column two-way travel time), peg-leg multiples (one leg of the ray path involves a subsurface reflector while the other leg bounces between two different interfaces), and long-period multiples from large impedance contrasts like the sea floor-to-sea surface round trip; interbed multiples occur between any two subsurface reflectors with sufficiently strong impedance contrast to sustain multiple round trips, with the most geologically significant being sub-salt interbed multiples generated between the salt base and the underlying high-impedance carbonate or anhydrite reflectors.
Multiple attenuation methods exploit the physical differences between primaries and multiples to separate them in the seismic data processing workflow — the most important difference is moveout velocity: multiples travel longer path lengths and have lower root-mean-square (RMS) velocities than primaries at equivalent reflection times (because the multiple has bounced from shallower reflectors while the primary has propagated to a deeper interface), allowing velocity discrimination using normal moveout (NMO) velocity filtering or parabolic Radon transform demultiple; however, when the multiple's velocity is close to the primary velocity at the same two-way time (common in shallow water or in thick, low-velocity sedimentary sections), velocity discrimination fails and alternative methods like predictive deconvolution, surface-related multiple elimination (SRME), or extended SRME must be applied.
Surface-Related Multiple Elimination (SRME) is the industry standard data-driven multiple attenuation method for marine seismic data — SRME predicts free-surface multiples by convolving the recorded seismic wavefield with itself (an operation that generates the predicted multiple contribution), then subtracts the predicted multiples from the recorded data to leave only primary reflections; SRME requires dense, complete shot-receiver offset sampling to produce accurate multiple predictions, making it best suited to 2D and 3D marine surveys with regular receiver spacing; the Delft University research group that developed SRME in the early 1990s and the subsequent industry development of 3D SRME, iterative SRME, and extended SRME for interbed multiples have collectively produced the most significant advances in marine seismic multiple attenuation of the past 30 years.
Multiple contamination in sub-salt seismic imaging is particularly severe because the high-impedance salt-sediment interfaces generate strong reflections, and the multiple energy from shallow salt flanks and over-thrusts can arrive at the same travel times as primary reflections from sub-salt reservoir targets; identifying whether a reflection in a sub-salt seismic section is a primary from a sub-salt reservoir or a multiple from a shallower interface requires modeling the expected multiple arrivals from the interpreted salt geometry and comparing the modeled multiple arrival times and amplitudes with the observed data; misidentification of multiples as primary sub-salt reflectors has led to expensive exploration wells drilled on false closures, making multiple attenuation quality control a critical step in GoM and other sub-salt exploration workflows.
Radon transform (parabolic and hyperbolic) demultiple uses the different moveout curvature of multiples versus primaries in common midpoint (CMP) gathers — multiples have a more hyperbolic moveout (lower velocity = more curvature) than primaries at the same two-way time — to separate them in the Radon (tau-p) domain where primaries and multiples occupy different regions; the high-resolution parabolic Radon transform maps each event in the data to a point in the Radon domain defined by its residual moveout after NMO correction, allowing multiples (which have more residual moveout due to their lower velocity) to be muted in the Radon domain before transforming back to the time domain with only primaries remaining; the Radon demultiple method is computationally intensive and requires careful velocity model building to define the moveout difference between primaries and multiples, but it can attenuate multiples in data where SRME is less effective (land seismic, shallow water with poor near-offset sampling).

Fast Facts

Multiple reflections have been recognized as a seismic processing problem since the earliest days of reflection seismology in the 1920s, when geophysicists first noticed that some reflections appeared at twice the two-way travel time of known shallow reflectors and therefore represented reverberations rather than primary arrivals from deeper interfaces. The development of predictive deconvolution by Enders Robinson at MIT in the 1950s provided the first systematic mathematical approach to multiple suppression, treating the seismic trace as a convolution of the source wavelet with the reflectivity series and using inverse filtering to remove the periodic component attributable to multiples. Today the suppression of multiples is one of the largest components of seismic data processing costs in marine exploration, with specialized demultiple algorithms running on clusters of thousands of processing cores for months on large 3D datasets in the deepwater GoM, offshore Brazil, and North Sea.

What Is a Multiple Reflection?

A perfect seismic reflection experiment would record only energy that traveled straight down from the source, reflected once from a subsurface interface, and traveled straight back up to the receiver. The real subsurface is not so cooperative. Strong acoustic impedance contrasts at shallow interfaces — the sea floor, the water surface in marine acquisition, major unconformities — act as mirrors that trap seismic energy in reverberation cycles, sending energy up and down between reflectors many times before it finally arrives at the surface receiver with a delay corresponding to a deeper reflector that may not exist.

This is the multiple reflection: energy that has taken a more complicated path than the simple down-up primary reflection, arriving at the receiver later and masquerading as a signal from a deeper, fictitious reflector. In a seismic section contaminated by strong multiples, false reflectors appear at depth, apparent dip structures are created where none exist, and the amplitude of true primary reflections is obscured by the superimposed multiple energy.

In shallow water marine environments, where the strong sea floor reflection bounces repeatedly between the sea floor and the sea surface, multiples can dominate the seismic record from the sea floor reflection time downward, effectively masking all primary reflections deeper than the first water-bottom reverberation. Eliminating or suppressing these multiples is not a refinement of the seismic processing workflow — it is a prerequisite for any meaningful geological interpretation of the subsurface.

Multiple Attenuation in Seismic Processing

Predictive deconvolution exploits the periodicity of simple water-bottom multiples — because the multiple travel time is a fixed addition (one water-column two-way time) to the primary arrival time, the multiple appears as a periodic repetition of the primary wavelet at integer multiples of the water column travel time; autocorrelation of the seismic trace detects this periodicity, and designing an inverse filter that removes the periodic component (while preserving the non-periodic primary arrivals) attenuates the water-bottom multiple series; deconvolution is computationally efficient and works well for simple, short-period multiples in relatively shallow water, but fails for long-period peg-leg multiples where the periodicity is not a simple multiple of a single time interval.

Velocity discrimination residual moveout analysis in common offset or CMP gathers provides multiple attenuation complementary to SRME by targeting the velocity difference between primaries and multiples — after NMO correction with the primary velocity field, residual moveout remains on multiple events because their lower velocities result in over-correction (the events are too flat relative to the NMO velocity), while primary events are flat; applying a high-resolution Radon or F-K filter in the residual moveout domain selectively attenuates the multiple events while preserving the flat primaries, providing a velocity-based separation that works best when the primary-multiple velocity difference is at least 3 to 5% at the relevant two-way times.

Multiple Reflection Across International Jurisdictions

Canada (AER / WCSB): WCSB onshore seismic surveys in Alberta and British Columbia encounter multiple reflection problems primarily from strong shallow reflectors including the base of glacial till (up to 100 meters thick), the McMurray oil sands (strong impedance contrast at the top of the oil sands), and the Devonian carbonate unconformity (a major acoustic impedance boundary that generates peg-leg multiples interfering with deeper Devonian carbonate exploration targets); AER does not specifically regulate seismic data quality standards, but industry practice through the Canadian Society of Exploration Geophysicists (CSEG) includes multiple attenuation as a standard component of seismic processing workflows for WCSB exploration data; the WCSB's shallow glacial cover and the presence of the high-impedance McMurray oil sands at moderate depths (50 to 500 meters) make multiple attenuation more challenging for sub-Cretaceous targets than equivalent marine seismic surveys where SRME is highly effective.

United States (API / BSEE): GoM deepwater seismic acquisition and processing for sub-salt exploration requires among the most sophisticated multiple attenuation workflows in global petroleum exploration — the combination of strong salt interfaces, variable water depth (0 to 3,000 meters), and deep targets (greater than 10,000 meters below sea floor) creates a complex multiple environment where water-bottom multiples, salt flank multiples, and interbed multiples all superimpose on the primary reflections from sub-salt reservoir targets; major GoM operators including Shell, ExxonMobil, Chevron, and BP invest heavily in proprietary multiple attenuation algorithms and high-performance computing infrastructure for processing their deepwater 3D seismic datasets, and BOEM uses subsalt seismic image quality as a component of its assessment of exploration risk and resource uncertainty in OCS lease evaluation.

Norway (Sodir / NORSOK): NCS marine seismic acquisition in the North Sea and Norwegian Sea encounters significant multiple reflection problems from the high-impedance Paleocene chalk reflectors (notably the Tor and Ekofisk Formations) that generate strong interbed multiples interfering with deeper targets, and from the shallow hard seabed in the northern North Sea and mid-Norwegian shelf where the rocky glaciomarine seabed creates complex water-bottom multiple patterns; Sodir's NCS seismic data acquisition and processing standards include requirements for multiple attenuation quality control in the final processed stack submitted for the NCS seismic database, and Norwegian geophysical research companies including CGG Norway, PGS (Petroleum Geo-Services), and TGS (TGS-NOPEC) have contributed significantly to global multiple attenuation methodology development for deep-water and hard-bottom marine environments.