ABSTRACT: Improved seismic imaging methods for more accurate interpretation and attributes
Whiting, Peter1, Carl Notfors2,
and Geoff Mansfield2
(1) Veritas DGC, Singapore, Singapore
(2) Veritas DGC
Conventional seismic data processing makes a number of significant approximations that inhibit the ability to accurately image and honour the recorded seismic data. Paying more attention to the actual physics of the problem can result in better and more accurate imaging, leading to more accurate interpretation, and more reliable seismic attributes.
The most common undesirable assumption in conventional processing is that the seismic energy travels along a straight raypath. Standard processing steps like spherical divergence, normal movoeut (NMO) and dip moveout (DMO) all are most commonly applied using a straight raypath approximation. The commonly quoted "hyperbolic" and "double-square root" approximations are simply straight raypath approximations. Processing seismic data with such approximate processing steps can lead to poorly imaged and badly focussed results that make interpretation more difficult. For the most accurate focussing and imaging of seismic reflections, it is important to avoid these approximations and use an imaging tool that accurately deals with the actual shape of the raypath.
Full Kirchhoff prestack time/depth migration can be used to correctly deal with raypaths that are not straight. The non-hyperbolic nature of diffraction surfaces can be accurately estimated and the seismic data can therefore be precisely focussed and imaged. The most obvious benefits are better steep dip imaging, better fault definition, more stable amplitudes, more accurate positioning and flatter gathers. All of these benefits have a direct effect on interpretation confidence and reliability of extracted attributes.
The output of gathers without the characteristic "curl-up" on far offsets due to the inadequacies of NMO, has a most obvious effect on the extraction of AVO attributes. Not only does full Kirchhoff prestack migration result in a well imaged stack, but the availability of generally flat gathers lead to more reliable attribute extraction. Comparisons of AVO attributes from conventional processing and Kirchhoff prestack imaging demostrate the superiority of the latter.
AAPG Search and Discovery Article #90913©2000 AAPG International Conference and Exhibition, Bali, Indonesia