Multi-sparsity-based spectral attributes for discontinuity detection

Abstract

Structural and stratigraphic discontinuities, such as faults and channels, generally contribute to the construction of traps and reservoirs. Spectral decomposition can utilize the sensitivities of different frequency components to different geological conditions to identify these geological anomalies. The sparse inverse spectral decomposition (SISD) involves a sparse constraint of time–frequency spectra, and one critical parameter is the sparsity which determines the time–frequency resolution. A small sparsity gives a low temporal resolution result that cannot be used for thin-bed detection. Conversely, a large sparsity provides a high-resolution result, but it may lose weak reflection signals. The complex geological conditions in the subsurface will lead to some difficulties in detecting the discontinuities by using the SISD method with a fixed sparsity. To address this issue, we propose multi-sparsity-based spectral attributes by fusing the amplitude spectra results of three different sparsities to detect subsurface discontinuities. Compared with the fixed sparsity, the multi-sparsity-based spectral attributes can detect more geological details and highlight geological edges more clearly. The application on a 3D real data with an area of 230 km2 from deep formation in Northwest China exhibits its effectiveness in discontinuity detection. The proposed method can detect the weak or small hidden geological details more and better than the fixed sparsity method, suggesting that it may serve as a future tool for detecting the distribution of geological abnormalities in subsurface.