Apparent Velocity
Apparent velocity is the speed at which a seismic wavefront appears to travel along a surface recording line or along a receiver array, as opposed to the true propagation velocity at which the wave travels through the subsurface medium. Because seismic energy typically arrives at the surface at an angle from the vertical rather than straight down, the wavefront sweeps laterally across the receiver spread at a rate that differs from the true velocity in the medium. The governing relationship is V_a = V / sin(theta), where V_a is the apparent velocity, V is the true propagation velocity in the medium, and theta is the angle of incidence measured from the receiver line surface (equivalently, the complement of the angle from the vertical). When a wave arrives exactly perpendicular to the receiver line (theta = 90 degrees), the apparent velocity equals the true velocity; as the angle of incidence decreases toward grazing incidence (theta approaching zero), the apparent velocity approaches infinity. For coherent noise such as ground roll (surface waves) arriving nearly horizontally along the receiver line, the apparent velocity equals approximately the Rayleigh wave propagation velocity (typically 150 to 600 m/s in shallow soil and rock), which is far lower than compressional wave velocities of 1,500 to 6,000 m/s at depth. This contrast in apparent velocity between signal (reflections arriving at near-vertical incidence) and noise (surface waves and refractions arriving sub-horizontally) is the physical basis for frequency-wavenumber (f-k) filtering, receiver group arrays, and all other coherent-noise-rejection methods used in seismic data processing worldwide.
Key Takeaways
- The apparent velocity formula V_a = V / sin(theta) links the observed lateral sweep rate to the true propagation velocity and the angle of approach: For a refracted wave travelling along a high-velocity refractor at true velocity V_r and returning to the surface at the critical angle theta_c = arcsin(V_1 / V_r) (where V_1 is the surface layer velocity), the apparent velocity along the receiver line equals V_r exactly. This is why the slope of the first-break travel-time curve (the plot of first-arrival time versus offset) directly gives the refractor velocity: the slope is the reciprocal of the apparent velocity of the head wave, which equals the reciprocal of the refractor's true velocity when the line is oriented in the dip direction. In Alberta Foothills exploration, where near-surface velocity inversions from low-velocity glacial till over high-velocity bedrock cause shot-to-shot variations in refraction apparent velocity, the first-break refraction velocities are used to build the near-surface velocity model for static corrections; the quality of the apparent-velocity picks in the first-break analysis directly controls the accuracy of the elevation statics that precede any meaningful reflector stack.
- Apparent velocity determines the spatial aliasing threshold for seismic receiver and source arrays: The Nyquist wavenumber criterion requires that a spatial wavefield be sampled at least twice per apparent wavelength to avoid aliasing. The apparent wavelength along a receiver line is lambda_a = V_a / f, where f is the temporal frequency of the wave. For a wave of true velocity V = 2,000 m/s arriving at theta = 30 degrees, the apparent velocity is V_a = 2,000 / sin(30) = 4,000 m/s; at a dominant frequency of 40 Hz, the apparent wavelength is 4,000 / 40 = 100 metres, requiring receivers at no more than 50-metre spacing to avoid aliasing this component. For ground roll with apparent velocity of 300 m/s at the same 40 Hz frequency, the apparent wavelength is 7.5 metres, requiring receiver spacing under 3.75 metres to avoid spatial aliasing of the noise. Since a receiver spacing that satisfies the signal Nyquist (50 m) will alias the ground roll (3.75 m threshold), the field acquisition strategy must use receiver group arrays of multiple geophones spread over a length of approximately one apparent ground-roll wavelength to attenuate the aliased noise by summing across the array before recording, rather than relying solely on dense individual receiver spacing to avoid aliasing.
- Frequency-wavenumber (f-k) filtering separates signal from noise by rejecting energy in the low-apparent-velocity zone of the f-k domain: The two-dimensional Fourier transform of a seismic shot record converts the time-distance (t-x) domain data into the frequency-wavenumber (f-k) domain, where each point represents a component wave with a specific temporal frequency f and spatial wavenumber k. The apparent velocity of any component is the ratio f/k (since V_a = lambda_a × f = (1/k) × f). On an f-k plot, signal energy (reflections with apparent velocities of 3,000 to 10,000 m/s) plots in a wedge-shaped region of low wavenumber and high frequency, while noise energy (ground roll at apparent velocities of 200 to 600 m/s) plots at high wavenumber and low frequency. The f-k filter applies a velocity-based fan mask that zeroes energy in the noise wedge (apparent velocities below 800 m/s, for example) while preserving energy in the signal cone. In the Montney play of northeast British Columbia, where loose glacial surface sediments generate strong dispersive ground roll with apparent velocities of 250 to 450 m/s depending on frequency, f-k filtering in the v_a = 200 to 600 m/s rejection band is routinely applied as the first step in the processing sequence to remove the noise before NMO correction and CDP stacking.
- Apparent velocity differences between compressional and shear reflections allow multi-component seismic processing to separate P and S wavefields: In three-component (3C) seismic acquisition, the vertical geophone records primarily compressional (P) wave energy while the two horizontal geophones record both P and converted shear (PS or C-wave) energy. At a given offset, the P reflection from a horizon arrives with a higher apparent velocity (NMO velocity corresponding to P-wave velocity) than the converted shear reflection (NMO velocity corresponding to approximately the average of P and S velocities). The moveout analysis that separates P and PS reflections in the CDP domain is a form of apparent-velocity discrimination: the P NMO velocity of 2,500 m/s contrasts with a PS NMO velocity of approximately 1,800 m/s, and the hyperbolic moveout difference can be exploited to separate the two modes for independent stacking. In Montney and Duvernay 3C seismic surveys, the PS shear reflection data provide Vp/Vs ratios that constrain fluid and lithology models independently of the P-wave amplitude variations, adding significant interpretive value for sweet-spot mapping at a modest incremental acquisition cost over a standard P-wave survey.
- Apparent velocity analysis of microseismic events locates hydraulic fracture activation in three dimensions during well completion: Microseismic monitoring of hydraulic fracture operations detects small-magnitude seismic events (M -3 to 0) generated when faults and natural fractures slip in response to pressure changes from the injected fluid. The monitoring array (perforated offset well or surface receivers) records the apparent velocities of P and S arrivals from each event across the receiver spread. By matching the observed apparent velocities on each receiver to those predicted by a velocity model for sources at various trial locations, the event origin time and location can be determined by least-squares inversion. The accuracy of this location depends critically on knowing the apparent velocity across the array for the specific P and S waves from each event, and errors in apparent velocity translate directly into location errors: a 5 percent error in P-wave apparent velocity produces a 5 percent error in the horizontal distance estimate, which at 400 metres from the monitor well corresponds to a 20-metre location error, enough to misidentify the fracture stage being activated or to incorrectly attribute activation to a natural fault rather than the intended perforation cluster. In the Duvernay Formation at Kaybob South, microseismic datasets covering multi-well pad completions record 2,000 to 8,000 events per stage across 10 to 14 stages per well, and the apparent-velocity-based location workflow is run in near-real time to provide fracture geometry feedback to the completion engineer.
Apparent Velocity in Seismic Data Processing, Acquisition Design, and Noise Rejection
The first practical application of apparent velocity in a seismic acquisition programme is the design of source and receiver arrays. A linear array of geophones summed (stacked) in the field acts as a spatial filter with a response that depends on the apparent velocity of each incoming wave. An array of n geophones spaced d metres apart has a maximum rejection at the wavenumber k_null = 1/(nd) (one null per array-length apparent wavelength). For a ground-roll component with apparent velocity of 350 m/s and dominant frequency of 15 Hz, the apparent wavelength is 23.3 metres; a 24-metre array of 6 geophones at 4-metre spacing produces nulls at approximately every half of this apparent wavelength, attenuating the ground roll by 15 to 20 dB relative to a single geophone while preserving deeper reflections whose apparent wavelengths at 40 Hz are typically 75 to 200 metres (well above the array length). In Alberta Foothills acquisition over foothills terrain, where strong linear ground roll from back-scattered waves complicates standard array design, source arrays (clusters of shot holes spaced 3 to 5 metres apart) are used in combination with receiver arrays to provide additional apparent-velocity discrimination in the 200 to 600 m/s noise band.
In 3D seismic acquisition, apparent velocity considerations drive the choice of receiver line orientation and the cross-line receiver spacing. A 3D receiver geometry in which the receiver lines are oriented obliquely to the dominant structural grain will record different apparent velocities on each receiver line direction: lines oriented parallel to the dip direction of a steeply dipping reflector will see low apparent velocities from cross-dip energy, while lines oriented perpendicular to strike will see the highest apparent velocities from up-dip reflections. The apparent velocity from the cross-dip direction must exceed the Nyquist threshold for the cross-line sampling interval, which imposes a maximum receiver line spacing for any given structural dip. In the Foothills belt west of Rocky Mountain House, where formation dips of 30 to 60 degrees in the thrust sheets produce apparent velocities as low as 2,500 m/s on cross-dip receiver lines, the 3D receiver line spacing must be limited to approximately 50 metres (compared to 100 metres acceptable in the flat Deep Basin to the east) to avoid spatial aliasing of the steeply dipping reflections in the cross-line direction.
The shot-interval velocity semblance analysis used in seismic processing is fundamentally a comparison of apparent velocities across a common midpoint (CMP) gather. The NMO velocity (V_NMO) used to align reflections before stacking is the apparent velocity of the reflection hyperbola as seen in the t-x domain: it is the velocity that correctly predicts the moveout of the reflector at the receiver line orientation. For a flat reflector, V_NMO equals the root-mean-square velocity of the overlying column; for a dipping reflector, V_NMO contains a component of the structural dip (apparent dip in the line direction), and the downdip V_NMO is lower than the updip V_NMO for the same reflector. This apparent-velocity anisotropy from structural dip (dip moveout, or DMO) must be corrected before the velocity analysis is valid for depth conversion; the DMO correction, standard in all modern processing sequences, removes the dip-induced apparent velocity component and gives a dip-independent NMO velocity that correctly represents the stacking velocity.
Refraction statics analysis, which corrects for near-surface velocity anomalies in land seismic data, relies entirely on measuring the apparent velocity of the first-arriving refracted wave along the receiver line. The field first-break arrival times are picked on each trace, and the inverse slope of the first-break travel-time plot gives the apparent velocity of the head wave from the near-surface refractor. In the Peace River area of northwest Alberta, the near-surface comprises variable thicknesses of Pleistocene glacial till (velocity 500 to 1,200 m/s) over Cretaceous shales and sands (velocity 1,800 to 2,400 m/s). The first-break apparent velocity transitions from the till velocity at short offsets (refraction from the base of till) to the bedrock velocity at longer offsets (head wave along the bedrock surface), and the crossover offset provides the till depth directly from the apparent-velocity break. This apparent-velocity segmentation of the first-break travel-time curves is the primary source of near-surface depth information used to compute the weathering statics corrections that align the deeper reflection events before stacking, and errors in apparent-velocity picks propagate directly into statics errors and residual reflection mis-ties between adjacent CDPs.