Band: Seismic Frequency Band, Bandwidth, and Signal Processing in Geophysics
In geophysics and oilfield signal processing, a band is a contiguous range of frequencies within the acoustic or electromagnetic spectrum that is characterised by some shared property, such as being generated by the source, preserved by the recording system, or retained by a processing filter. The concept of a frequency band appears in virtually every quantitative measurement made in subsurface exploration: the seismic reflection band (typically 8 to 120 Hz for conventional land seismic and 5 to 200 Hz for marine seismic) determines what geological features can be resolved from reflection data; the sonic logging band (1 to 20 kHz for standard wireline sonic tools, 10 to 100 kHz for ultrasonic imaging tools) governs the wavelength in the formation and therefore the depth of investigation of acoustic measurements; the magnetotelluric band (0.001 to 10,000 Hz for crustal surveys, 1 to 100,000 Hz for shallow resistivity surveys) determines the depth range over which subsurface resistivity structure can be imaged. The boundaries of a band are defined by the frequencies at which the signal power falls to half its peak value (the -3 dB points) in the standard engineering convention, or by the corner frequencies of a filter applied to the data in a processing context.
The bandwidth of a measurement system or a dataset is the width of the frequency band, measured in hertz (Hz) or in octaves on a logarithmic scale. Wider bandwidth means more independent frequency components are available to construct the wavelet, which directly translates to sharper temporal resolution and the ability to distinguish between thinner geological layers. The Rayleigh resolution criterion defines the minimum separation of two reflectors that produces distinguishable reflection peaks as approximately one quarter of the dominant wavelength (lambda/4), so a seismic dataset with a dominant frequency of 40 Hz and a seismic velocity of 3,000 metres per second has a dominant wavelength of 75 m and a theoretical vertical resolution of approximately 19 m. Increasing the recorded bandwidth by recovering higher frequencies through careful acquisition design, processing, and deconvolution can reduce this resolution limit substantially, enabling detection of Montney sub-zones of 8 to 15 m thickness that would be below the tuning limit at lower bandwidth. Conversely, reducing bandwidth through aggressive filtering to suppress noise creates a coarser wavelet with lower resolution, which may prevent correlation of seismic events with specific lithostratigraphic units in the subsurface.
Key Takeaways
- Seismic frequency band and its geological implications: The seismic frequency band of a land acquisition programme in the WCSB typically spans 8 to 80 Hz in the raw field record after anti-alias filtering, with the usable signal band narrowing to 15 to 60 Hz after ground roll and noise suppression processing. The low-frequency end of the band is critical for rock physics inversion (low frequencies carry relative impedance information needed for absolute impedance and Vp/Vs ratio estimation), while the high-frequency end controls resolution of individual reservoir units. Loss of the low-frequency components below 10 Hz due to instrument coupling issues or low-cut filtering destroys the ability to distinguish between shale, sandstone, and carbonate lithologies from impedance inversion alone. Loss of high-frequency signal above 50 to 60 Hz due to earth attenuation in the Devonian section of the Alberta Basin limits resolution to approximately 15 to 20 m at 3,000 m depth, which may be acceptable for mapping major formation tops but is insufficient for characterising individual Duvernay benches of 3 to 8 m thickness.
- Acoustic logging bands and tool design: Wireline and LWD sonic tools generate acoustic energy in the 1 to 20 kHz band, with monopole tools using lower frequencies for refracted compressional (P-wave) and shear (S-wave) measurements at depths of investigation of 0.3 to 0.6 m into the formation, and higher-frequency dipole tools emphasising 2 to 6 kHz for flexural wave shear measurements in slow formations like unconsolidated Montney siltstones. Ultrasonic borehole imaging tools such as the UBI (Ultrasonic Borehole Imager) operate in the 200 to 500 kHz band, providing millimetre-scale circumferential images of borehole wall reflectivity and standoff, but with essentially no depth of investigation into the formation beyond the mud-filtrate invaded zone. The choice of frequency band in acoustic logging therefore reflects a direct trade-off between depth of investigation (lower frequency penetrates deeper) and spatial resolution of the measurement (higher frequency resolves smaller features).
- Bandwidth and vertical resolution: The tuning thickness of a thin bed, the minimum thickness at which the top and base reflections can be resolved as separate events, equals approximately lambda/4, where lambda is the dominant wavelength of the seismic wavelet. For a Montney B reservoir with P-wave velocity of 4,800 m/s and a dominant frequency of 40 Hz, the tuning thickness is approximately 30 m. If the acquisition and processing team can recover usable signal to 70 Hz using broadband recording and deghosting, the dominant frequency increases to approximately 50 Hz and the tuning thickness drops to 24 m. For Montney sub-zones of 8 to 12 m, this is still above the tuning limit, meaning that broadband processing alone cannot resolve individual Montney benches from standard surface seismic; only borehole seismic or seismic inversion constrained by well log data provides sub-tuning thickness characterisation.
- Band-pass filtering as the fundamental processing operation: Band-pass filtering, which retains frequencies within the signal band and attenuates frequencies outside it, is the most universally applied processing step in seismic data conditioning. It is applied to shot records to suppress noise before velocity analysis, to stacked sections to improve display quality, and to migrated volumes before attribute extraction. The choice of filter corners significantly affects interpretation results: a too-narrow passband removes genuine signal at the band margins and produces a smooth, lower-resolution volume; a too-wide passband retains noise at both the low and high ends, degrading signal-to-noise ratio and creating artefacts in amplitude analysis. WCSB seismic processing workflows typically specify band-pass filter panels at several corner-frequency combinations (for example, 5-10-70-90, 5-10-80-100, and 5-10-90-110 Hz) to allow the interpreter to visually assess signal quality at different bandwidths before selecting the final processing parameters for the deliverable volume.
- Broadband seismic acquisition and its impact on Duvernay characterisation: Conventional seismic recording uses an anti-alias filter with a corner frequency at 83% of the Nyquist frequency of the recording system, which for a 2 ms sample interval (500 Hz Nyquist) is approximately 415 Hz, far above the usable seismic band. Broadband acquisition techniques, which include deghosting of marine data, extended low-frequency recording using low-cut-free receivers, and simultaneous-source acquisition to improve the low-frequency signal-to-noise ratio, have expanded the recoverable seismic band from the conventional 10 to 60 Hz range to 5 to 90 Hz or beyond in premium surveys. In the WCSB, broadband Duvernay surveys acquired by several operators since 2018 have demonstrated improved definition of the Duvernay-Majeau Lake boundary, better discrimination of the Duvernay A, B, and C benches, and more reliable Vp/Vs inversion results for TOC and porosity prediction, directly reducing exploration risk on multi-well development programmes that depend on seismic-guided well targeting for CAD 12 to CAD 18 million Duvernay wells.
Seismic Frequency Bands and Earth Attenuation
Seismic waves lose amplitude as they travel through the earth due to geometric spreading, scattering, and anelastic attenuation. Anelastic attenuation, which converts wave energy to heat through viscous friction in fluid-filled pore spaces, is frequency-dependent: high-frequency components of the seismic wavelet are attenuated more strongly than low-frequency components, causing the frequency content of the recorded signal to shift toward lower frequencies as travel time (depth) increases. This preferential high-frequency attenuation is quantified by the quality factor Q, where a higher Q indicates less attenuation (Q = infinity is perfectly elastic). In the Alberta Basin, Q values for shale-dominated sections such as the Cretaceous Colorado Group range from 50 to 100 at seismic frequencies, while tight carbonate sections like the Devonian reef trends have Q of 100 to 300. The earth-filter effect of Q on the seismic band means that a source wavelet with 80 Hz high-cut content at the surface arrives at 3,000 m depth with a high-cut content of approximately 40 to 50 Hz, depending on the dominant Q of the travelled path, irrespective of the recording system's capability to record 80 Hz.
This depth-dependent bandwidth loss is one reason why borehole seismic data (VSP and walkaway VSP) provides higher-frequency signal than surface seismic for the same geological target: the seismic source wave travels only one-way from the surface to the receiver in the borehole, suffering half the attenuation of a surface-to-reflector round trip. A VSP receiver at 3,000 m depth records the downgoing wavelet with its full broadband content because it has not yet been attenuated by the return path, and the upgoing reflected wavelet from the target formation only 200 to 300 m below the receiver has lost correspondingly less of its high-frequency content than the same reflection observed at surface after a 6,000 m round trip. WCSB operators use VSP surveys to calibrate surface seismic bandwidth in each formation interval and to verify that the processing team's frequency estimates for the deliverable seismic volume are consistent with what the earth actually transmits at each depth horizon.
Acoustic Logging Bands and Shear Wave Measurement
The measurement frequency band of acoustic logging tools is matched to the wellbore and formation geometry to optimise the mode of wave propagation being measured. Standard monopole sonic tools operate in the 5 to 15 kHz band, generating compressional head waves that refract along the formation boundary and return to the tool's receiver array at the formation P-wave velocity. The monopole tool also generates a Stoneley wave, a guided mode that travels along the fluid-rock interface at the borehole wall, in the 1 to 3 kHz band; Stoneley wave velocity and amplitude are sensitive to formation permeability and are used in some Montney and Duvernay well evaluations to identify fracture zones and permeable intervals for completion targeting. Dipole tools fire asymmetrically to generate flexural waves in the 2 to 8 kHz band, which allow shear wave velocity measurement in formations too slow for refracted shear head waves (any formation where Vs less than 1,500 m/s, the acoustic velocity of the borehole fluid, cannot generate a refracted shear head wave and requires a dipole tool to measure Vs from the flexural mode).
The selection of optimal frequency for a dipole log in a given formation requires a balance between wavelength and borehole diameter. The dipole tool's optimal measurement frequency places the borehole diameter at approximately one wavelength of the flexural wave; for a 216 mm (8.5 inch) borehole and a formation shear velocity of 2,400 m/s, the optimal dipole frequency is approximately 11 kHz. In over-gauge boreholes, common in Buckinghorse shale above the Montney in the Dawson Creek area, the larger effective diameter requires a lower optimal frequency, and some high-frequency dipole tools perform poorly in these conditions, reporting artificially low shear velocities that skew mechanical property calculations if the frequency mismatch is not recognised and corrected. Field-variable frequency dipole tools that can be tuned to the formation and borehole conditions are available from major service companies and are specified for wells where borehole caliper variability is anticipated from pre-drill lithologic assessment.