•  
  •  
 

Turkish Journal of Earth Sciences

Abstract

Seismic amplitude anomalies are the primary tool for direct hydrocarbon indicator (DHI) screening, but amplitude anomalies that can resemble gas-sand DHI responses in poststack data can arise from tuning interference at quarter-wavelength bed thickness. Evaluating whether a bright spot is a tuning effect requires accurate local interval velocity, which is in practice delivered by full-waveform inversion (FWI) or reflection tomography—both of which produce band-limited models. How velocity-model resolution controls the spatial accuracy of tuning identification was quantified using the Marmousi-II benchmark. Eight band-limited velocity models spanning 3–80 Hz were constructed by two-dimensional wavenumber-domain low-pass filtering, mimicking the resolution attainable at successive stages of a multiscale FWI workflow. Each band was evaluated against the true, unsmoothed velocity model. Identification accuracy (quantified by the F1 score) declined monotonically from 0.96 at 80 Hz to 0.77 at 3 Hz; missed tuning warnings increased from 4% to 22% of true tuning beds, and false alarms increased from 5% to 24% of band flags. A resolution threshold near 20 Hz separated acceptable (F1 ≥ 0.90) from degraded identification. Deep sections showed progressively greater sensitivity to velocity smoothing below 10 Hz, owing to path-integrated velocity averaging and the greater tuning-thickness sensitivity associated with Q-related frequency loss. A spatially resolved tuning-risk map showed that missed warnings and false alarms clustered at structurally complex zones and stratigraphic pinch-outs—precisely the settings of interest in mature exploration provinces. The results provide a quantitative basis for suggesting that an effective resolution near or above 20 Hz may be needed before velocity models are used to support DHI screening in Marmousi-II-like settings. The reported numbers should be read as an upper bound; real FWI velocity errors are not strictly bandlimited, and field models will generally perform worse at equivalent peak frequencies.

Author ORCID Identifier

ERGİN KARACA: 0000-0002-5114-8975

DOI

10.55730/1300-0985.2114

Keywords

Seismic tuning thickness, velocity resolution, full-waveform inversion, tuning-risk mapping, Marmousi-II

First Page

584

Last Page

595

Publisher

The Scientific and Technological Research Council of Türkiye (TÜBİTAK)

Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

Share

COinS