Instantaneous spectral attributes to detect channels

Abstract

Channels filled with porous rock and encased in a nonporous matrix constitute one of the more important stratigraphic exploration plays. Although attributes such as coherence can be used to map channel width, they are relatively insensitive to channel thickness. In contrast, spectral decomposition can be used to map subtle changes in channel thickness. The peak spectral frequency derived by using the short-window, discrete Fourier transform (SWDFT) is an excellent tool for mapping such changes along an interpreted horizon. We show that by use of instantaneous spectral attributes, we can generate equivalent maps for complete seismic volumes. Because we are often interested in mapping high-reflectivity channels encased in a lower-reflectivity matrix, we find that a composite plot of the peak frequency and the above-average peak amplitude accentuates highly tuned channels. Finally, by generating a composite volume using peak frequency, peak amplitude, and coherence, we can establish not only the channel thickness, but also its width. We demonstrate the value of such 3D volumetric estimates through application to (1) a marine survey acquired over Tertiary channels from the Gulf of Mexico and (2) a land data survey acquired over Paleozoic channels from the Central Basin Platform, west Texas, United States. The channels in both marine and land surveys can be highlighted through composite-volume analysis.