The foundations for wavelet analysis highlight how frequency and phase control seismic response [2] . When thin‑bed effect was first articulated [3] , it showed that beds thinner than one‑quarter of the dominant wavelength (λ) produce composite reflections. The wedge model became the standard framework for analyzing tuning effects. This was further formalized by examining various resolution criteria [4] , concluding that the ultimate limit of resolution is governed by the bandwidth and dominant frequency of the seismic wavelet. In essence, higher frequency (broader bandwidth) yields finer resolution in direct proportion.

Subsequent research has focused on improving resolution. Spectral decomposition was introduced [5] . The applications of spectral attributes were expanded to stratigraphic interpretations [6] . AVA and inversion frameworks sought to extract fluid and lithology information [7] - [9] , though these remain subject to tuning limits. More recent work emphasizes bandwidth expansion via broadband seismic, Q‑compensation, and Full Waveform Inversion (FWI) to sharpen imaging. Still, tuning remains a persistent source of volumetric uncertainty.

Seismic Tuning, Thickness Resolution and Wedge Models

When a bed’s thickness is less than approximately λ/8 (one-eighth of the dominant wavelength), the separate reflections from the top and bottom cannot be distinguished [3] . They form a single composite reflection whose amplitude is roughly proportional to the bed’s thickness. The bed is effectively “tuned” – below this limit, thinner beds yield progressively weaker composite amplitudes, approaching zero amplitude as thickness tends toward zero. At exactly zero thickness, the reflections from the interface would perfectly overlap in opposite polarity and cancel out. For a zero-phase wavelet that in the ideal case of equal but opposite reflection coefficients, the maximum amplitude occurs when bed thickness is λ/4 (quarter wavelength). λ/4 is therefore often referred to as the tuning thickness, corresponding to the peak constructive interference of a thin bed’s reflections. Beyond λ/4, further increases in thickness cause the top and bottom reflections to separate in time such that their interference diminishes.

Quantitatively, the relationship between amplitude and thickness for a thin-bed reflection is often visualized by a wedge model and tuning curve. Figure 6 shows a tuning curve derived from the wedge model ( Figure 6(A) ) (for a simplified case of a gas sand layer encased in shales) [10] . Figure 6(B) shows the synthetic seismic section when the wedge is illuminated by a zero-phase Ormsby wavelet. The thickness estimates and amplitude tuning curves measured from the seismic response are shown as Figure 6(C) and Figure 6(D) , respectively. Labels are added to these curves to indicate the thickness of the wedge in units of the dominant wavelength (λ).

The vertical axis ( Figure 6(D) ) is the peak amplitude of the composite reflection (normalized to the amplitude of a single interface reflection), and the horizontal axis is the wedge thickness expressed in meters. At zero thickness, the amplitude is zero (since the top and base reflections cancel exactly). As thickness increases, amplitude rises almost linearly at first - in this “thin” regime (<~λ/10), amplitude is roughly proportional to thickness. The onset of tuning can be marked around λ/5, beyond which the simple linear relationship breaks down as interference becomes significant. The curve reaches a maximum at the tuning thickness (λ/5 or 18 m in this case), where amplitude is at its peak constructive interference. Notably, this maximum amplitude is higher than the amplitude would be for a very thick bed with the same reflectivity - a thin bed at tuning can produce a “super-bright” reflection. For instance, a bed at tuning thickness can have considerably more reflective power than the isolated reflectors would suggest, purely due to wavelet interference. Beyond this point, as thickness grows further, the two reflections start to separate, and the composite amplitude begins to drop. By ~λ/2 thickness, the amplitude often falls below the tuning peak. In the example curve, after λ/4 the amplitude declines and eventually flattens out to the level of a single-interface response (100% of true reflectivity) once the top and base are fully resolved (no interference). This means a thick bed (>>λ/4) will show the “correct” amplitude for the top reflection (assuming standard processing and no attenuation), but a thinner bed (<λ/4) will either exaggerate or mute that amplitude. Specifically, in the zone just below λ/4, amplitude is boosted (tuning boost), whereas much thinner beds (λ/4) see amplitude diminished (tuning roll-off).

These theoretical insights have direct implications for interpreting seismic DHIs. A “bright spot” on seismic could mean a thick gas sand with strong reflectivity, but it could equally be a thinner sand tuned to λ/4 that produces an anomalously high amplitude. Conversely, the absence or dimming of an amplitude could mean there is simply no pay, but it might also mean the pay is present but below tuning thickness or tuned out by destructive interference. Cases have been documented where thin-bed tuning led to underestimation of net pay when using seismic amplitude alone [10] [11] . As shown in Figure 6 , a wedge model mimicking gas sand found that beds thinner than ~λ/4 yielded measured thicknesses and amplitudes that deviated substantially from true values, introducing amplitude responses at “tuned’” thicknesses that could mislead interpreters.

Bonn-Ario field is in the Coastal Swamp depo belt of the Niger Delta basin, 40 km South-East of Port Harcourt. The study interval is part of the Agbada formation, a paralic sequence of sands and shales up to 4 Km thick in the central parts of the basin.

The Tertiary Niger Delta lies in the Gulf of Guinea, equatorial West Africa, between latitudes 3˚ - 6˚N and longitudes 5˚ - 8˚E ( Figure 7 ) [12] . It occupies the site of a triple-junction created by South Atlantic opening during the Late Jurassic breakup of Africa and South America [13] [14] . Deltaic progradation began in the Early Tertiary (Eocene) and built one of the world’s largest regressive deltas, covering roughly 300,000 km² [15] , containing an estimated 500,000 km³ of sediment [16] and reaching sediment thicknesses in excess of 10 km at the basin depocenter [17] .

Stratigraphically the delta is divided into three lithostratigraphic units that record an overall progradational stratigraphic succession: the marine (pro-delta) Akata Shale Formation, the paralic (delta-front) Agbada Formation, and the continental (delta-plain) Benin Formation ( Figure 8(A) ). The Akata Formation is the basal Tertiary unit and consists predominantly of marine shales that act as the Niger Delta’s principal source rock; thickness estimates for this unit reach up to about 7 km [18] . Above it, the Agbada Formation comprises alternating sand and shale sequences and represents the “true” deltaic portion of the basin; the proportion of shale decreases progressively toward the upper part of the Agbada Formation. The Benin Formation, the youngest unit, is characterized by continental sands and gravels with only minor coastal-plain shale interbeds and attains thicknesses of up to ~2 km in the central delta [14] .

Structurally the basin is commonly described as a linked extension-contraction system: extensional regimes dominate on the onshore delta top while contractional features are more apparent toward the deep-water toe of thrust. The Niger Delta evolved as a prograding extensional complex riding above a ductile, overpressured substrate - the marine Akata shales. Characteristic structural elements include roll-over anticlines (often with collapsed crests), back-to-back structures, and growth (syn-sedimentary) faults ( Figure 8(B) ). Many of these faults cut through parts of the Agbada Formation, the largest of which commonly sole out onto detachment planes near the top of the Akata Formation.

Amplitude extractions are used for fluid contact delineations when changes in pore fluid are likely to produce reliable, measurable changes in seismic amplitudes and when those signals can be calibrated against well data. The changes are expressed as strong impedance contrast with the surrounding rocks and visible amplitude anomalies on the seismic data (bright/dim spots). In reservoirs with limited down-dip well control where the fluid contacts (Gas-Water Contact, Oil-Water Contact or Gas-Oil Contact) have not been penetrated, geoscientists commonly turn to seismic amplitude extractions.

Our case study, the Bonna-Ario M1000 reservoir, has demonstrated that using seismic amplitude shut-off as a proxy for the GWC would lead to an underestimation of the gas column height by ~35 ft and invariably lead to an underestimation of the GIIP. In this case, GIIP estimates using the GWC “seen” by the wells was 1.89 trillion cubic feet (TCF), and seismic amplitude-derived GWC was 1.60 TCF ( Figure 12 ). In other words, seismic‑only volumes underestimated the GIIP by ~15%. This magnitude of error could materially affect reserves booking, project sanctioning, and investment decisions. For reservoirs globally, tuning can evidently bias volumetric assessments toward conservatism in well data-sparse reservoirs where no fluid contacts have been logged by the wells.

In the SPE PRMS [1] oil and gas resources estimation rules, seismic indicators fall under reliable technology - a grouping of one or more technologies (including computational methods) that have been field tested and have been demonstrated to provide reasonably certain results with consistency and repeatability in the formation being evaluated or in an analogous formation - that may be used to extend hydrocarbon presence below the lowest known hydrocarbon in reservoirs where no hydrocarbon (fluid) contacts have been drilled by the wells for Probable (2P) Volumes estimation [22] . This study has, however, shown that seismic amplitude extractions alone will not be sufficient to accurately define fluid contacts. Equating seismic amplitude cutoffs with fluid contacts in sub‑tuning areas of reservoirs introduces systematic pay thickness and volumetric underestimation risks. Rather, amplitude extractions should be treated as the λ/4 resolution limit of the seismic data. For more accurate estimation of the in-place volumes in such reservoirs, this resolution limit must be determined through wedge modelling and AVA analysis, and factored into the seismic amplitude-derived fluid contact estimate.

The broadband PSDM seismic volume covering the Bonn-Ario field has been shown to have a bandwidth of ~75 Hz typical of conventional seismic acquisitions. This pales in comparison to modern broadband seismic data with useable frequencies up to 100 Hz, with extensions above 150 Hz for high resolution applications [23] [24] . Advances in broadband seismic acquisition and processing will mitigate, but not eliminate tuning-induced pay and volume mis-estimations. Seismic amplitude extractions paired with AVA analysis and wedge modelling have been demonstrated to provide a reliable, cheap, quick and consistent estimation of the pay thickness and fluid contacts.

This study demonstrates how tuning thickness controls seismic amplitudes and can lead to the underestimation of GIIP when amplitude shut offs are misinterpreted as fluid contacts. For the Bonn-Ario M1000 reservoir, wells outside the seismic amplitude outline encountered > 20 ft gas pay. Forward modeling and wedge analysis confirm that the observed seismic amplitude Shut off before the logged GWC is resolution‑driven, and for the M1000 reservoir, could potentially lead to a 35 ft underestimation of gas pay thickness and a ~15% underestimation of in-place volumes if the seismic amplitude derived GWC is used for volumetric estimations. This finding is of particular importance to reservoirs with no logged fluid contacts, where the estimated fluid contacts are seismic amplitude-derived even if the reservoir stratigraphic thickness may be above the tuning thickness. Analogues from the Niger Delta and multiple basins confirm the global relevance of this issue. Adoption of integrated workflows and advanced seismic processing can mitigate the risk, ensuring more accurate reserves booking and investment decisions. The following recommendations can be drawn from the M1000 reservoir case.

1) Calibrate seismic to wells and don’t ignore discrepancies. If wells find hydrocarbon outside the seismic anomaly, do not dismiss it as an outlier - it likely means the seismic picture is incomplete. Investigate the cause (tuning, noise, processing issues) and update the interpretation. An interpreter should actively seek to reconcile well and seismic data to avoid tuning pitfalls.

2) Tuning will mask or distort hydrocarbon volume. In our case study, seismic amplitude anomalies only registered on seismic where it thickened towards 35+ ft; thinner gas pay beyond the “bright” area was essentially invisible on the data. This resulted in a significant underestimation of gas initially in place by the seismically derived fluid contact estimates. Always consider that amplitude anomalies may not delineate true reservoir boundaries - they often delineate the λ/4 thickness contour. Seismic amplitude shut offs do not necessarily delineate the downdip limit of hydrocarbon accumulations.

3) Use AVA analysis, wedge models and synthetics as routine tools. Building a quick wedge model for your reservoir, using expected wavelet and rock properties, is extremely useful. It will tell you what the tuning curve looks like and at what thickness you can trust amplitudes. This can guide whether an amplitude map needs scaling. In the case study, the wedge model predicted a 35 ft tuning thickness, which aligned with observations on seismic data. We recommend doing this modeling prior to finalizing volumetrics in reservoirs where the fluid contacts are determined from seismic amplitudes.