Reservoir Characterization Based on Seismic Spectral Variations

It is seismic frequency, rather than the amplitude, providing the essential information for fluid discrimination in reservoir geophysics. In this study, I propose to exploit the spectral variation in the frequency domain for reservoir characterization. For a seismic profile in the time-space domain, one can generate a vector of the instantaneous frequency spectrum at every point. These spectral vectors at different time-space locations have different variational characteristics. It is the spectral variation pattern, rather than the amplitude in the frequency spectrum, playing a key role in reservoir characterization. I demonstrate this innovative technology in the characterization of coalbed methane (CBM) reservoirs. The prediction process takes account implicitly the factors of CBM reservoirs (such as viscosity, elasticity, cleat system, wave interference within a coal seam etc.) affecting the frequency spectrum, and based on its variation pattern, one is able to quantitatively predict the spatial distribution of potential reservoirs.