Application of adaptive parameterized S‐transform to delta sandstone reservoir identification

Abstract

Time-frequency analysis plays an important role in seismic interpretation and reservoir identification because it reveals the variation of seismic frequency contents over time. The S-transform is recognized as an efficient method for seismic time-frequency analysis, but its frequency distribution biases the actual Fourier spectrum due to the linear frequency-dependent term in the analytical window. In this paper, an improved S-transform, named adaptive parameterized S-transform, is proposed for identifying delta sandstone reservoirs. In the proposed method, three parameters are introduced into the S-transform to better adjust the time-frequency resolution in different frequency bands. Then these parameters are selected adaptively by using the concentration measure to provide a highly energy-concentrated time-frequency representation for seismic signals. Two synthetic examples validate the effectiveness of the proposed method. Analysis of field seismic data illustrates that it can provide a better reservoir interpretation with a high resolution. Comparisons between the wavelet transform, S-transform, and modified S-transform prove the superiority of the proposed method in describing reservoir distribution and decreasing the uncertainty of reservoir detection. This paper presents a complementary approach to current methods for reservoir detection and identification.