Thin-Bed
Reflectivity Inversion and Seismic Interpretation
Chopra, Satinder1, John P.
Castagna2 (1) Arcis Corporation,
Enhancing the bandwidth of surface
seismic data has always been a desirable goal for geoscientists. Conventional
wisdom dictates that in the presence of noise and consequent broadening of the
seismic wavelet during its subsurface journey, the resolution limit is a
quarter of the dominant wavelength of the data. This limit follows from the
Widess model, which is essentially a special case of a realistic model. Based
on an analytical analysis of a realistic model, it is found that the seismic
amplitude and frequency vary continuously far below the conventional view of
the limit of seismic resolution and it is possible to infer thickness below the
seismic sample rate. This implies that frequency beyond the seismic data
bandwidth can be recovered.
Thin-bed spectral inversion method is a
novel way of removing wavelet from the seismic data and extracting
reflectivity. This inversion process does not require stringent assumptions for
its performance. It does not require any a priori model, any reflectivity
assumptions or horizon constraints and neither is a well constraint mandatory,
though having at one well control point is helpful. For data with high
signal-to-noise ratio, thicknesses far below tuning can be resolved.
Appreciable noise in the data deteriorates the performance of the inversion
outside the frequency band of the original seismic data, but the method still
enhances high frequencies within the band without blowing up noise as
conventional deconvolution would do. Nevertheless, the highly resolved seismic
data retrieved in the form of reflectivity data is very useful for making
accurate interpretations and proves to be advantageous in many ways.
AAPG Search and Discover Article #90063©2007 AAPG Annual Convention, Long Beach, California