Multi-component seismic-resolution analysis using finite-difference acquisition modelling

Abstract

Various rules-of-thumb (e.g. Fresnel Radius, Rayleigh Limit) are commonly used to predict seismic resolution, based on image wavelength. However, seismic resolution ultimately depends on more fundamental parameters including survey design, source bandwidth, geology and data processing. A more instructive analysis is possible via numerical modelling of the acquisition process. Here we demonstrate the improved insight available with this approach, with examples taken from the coal and petroleum sectors. We use viscoelastic finite-difference modelling to simulate 2D multi-component acquisition sequences. The ability to allow for anelastic attenuation is important as it permits a more realistic comparison of the resolution achievable on P-wave and PS-wave imagery. Analysis of a typical coal target suggests that barren-zones of width 5-10 metres can be resolved. The interplay of wavelength and attenuation is such that the PS-wave image is likely to exhibit comparable, or slightly reduced, lateral resolution where statics are not a problem. Resolution can be downgraded significantly if statics are more severe, and in practice this is likely to have greater impact on the PS image. A second example examines detection of lens-like features at petroleum depth. The resolving ability on the P-wave imagery is broadly consistent with analytical predictions (100m laterally and 40 m vertically). In terms of resolution, the PS-images are perhaps less competitive than at the coal scale, because the longer path lengths yield greater relative attenuation. Realistic numerical modelling, simulating the full acquisition and processing sequence, leads to a more pragmatic understanding of seismic resolution issues. It is a valuable tool for survey planning and image interpretation.