High-Resolution Spectral Reflectivity Inversion: A Case Study on Thin Bed Reservoirs

Carlos Cabarcas¹, John Castagna¹, and Oleg Portniaguine^3
1 The University of Oklahoma, Norman, OK
² Fusion Geophysical,

Conventional wisdom generally considers a thin layer (thickness smaller than 1/4 seismic wavelength) to be below seismic resolution. Therefore, the top and base reflections from a thin reservoir cannot be separated. However, these well known resolution limitations (Widess, 1973, Kallweit 1982) are based on the study of very special cases (Chung and Lawton, 1995) which may be inappropriate for “real world” situations. For example, resolution below 1/4 wavelength is possible for even signals. Also, the incorporation of instantaneous spectral decomposition into a seismic reflectivity inversion can be tailored to overcome conventional resolution limitations and effectively remove the seismic wavelet.

Our spectral inversion algorithm is applied to the Boonsville public domain dataset. This dataset was acquired over a well-known giant gas field located in the Forth Worth Basin, Texas. The aim of our algorithm is to improve the characterization of thin bed reservoirs. The resolution improvement and detail obtained over the Atokan Age thin bed reservoirs of this field are very remarkable. The Caddo Formation reservoirs are better delineated and extensively correlated with production information. Elusive compartments recognized in several wells but impossible to identify on the seismic volume are easily mapped after the inversion. Spectral reflectivity inversion describes with sufficient confidence isolated thin geologic bodies.

It should be noted that this methodology does not provide a unique solution and it is also very dependent on a good first approximation to the source wavelet. Nevertheless, the results after reflectivity inversion are more interpretable and useful than the original seismic data.

AAPG Search and Discovery Article #90039©2005 AAPG Calgary, Alberta, June 16-19, 2005