Application of instantaneous amplitude and spectral decomposition to determine the location and thickness of a hydrocarbon reservoir

Spectral decomposition is a robust attribute in the characterisation of post-stack seismic data. Reflected wavelets from specific geological lithologies show distinct patterns in the frequency spectrum, which is related to the wavelet frequency content and the interference pattern of a structure within a layer. Generally, the frequency spectrum of a wavelet has a smooth pattern through the entire bandwidth. However, the structure of a layer, such as its thickness and lateral distribution, will appear as a periodic function in the frequency domain. This research investigates the distinct patterns of reflective layers with different thicknesses using instantaneous amplitude and fast sparse S-transform (ST). To verify the efficiency of both methods, we have examined these approaches on channel-shaped and wedge-shaped synthetic models. The results showed that the fast sparse ST could predict features of both synthetic models better than the instantaneous amplitude. Next, the fast sparse ST is performed on experimental field data selected from a time interval containing a hydrocarbon channel reservoir. We used a sparse deconvolution algorithm to suppress the wavelet effect and balance all the frequencies. The results of this research showed the efficiency of the fast sparse ST in predicting a hydrocarbon channel reservoir's thickness.