Automated spectral recomposition with application in stratigraphic interpretation

Conclusion

Automated spectral recomposition using separable nonlinear least squares represents the seismic spectrum as a sum of Ricker components and efficiently estimates their peak frequencies and amplitudes. With the seismic spectrum reconstructed from component frequencies, spectral recomposition can be used in seismic interpretation. We adopted Ricker wavelet in the analysis because of its popularity. Other wavelets may also be used with corresponding estimation numerical strategy.

Applying spectral recomposition, we have been able to better visualize seismic images in both cross sections and stratal slices. This technique has improved the interpreter's ability to image the various elements of the depositional system in extracted stratal slices. Spectral recomposition can also be used in forward modeling, in studying how different frequency components attenuate in the subsurface, newand in estimating thin-bed thicknesses using tuning frequencies. It provides a robust and phase-independent approach to seismic thickness estimation. Compared with conventional methods involving adjacent peaks and troughs picking, spectral recomposition requires only peak-frequency and amplitude estimation.

Automated spectral recomposition with application in stratigraphic interpretation