The normalized W transform for seismic interpretation

ABSTRACT The W transform is an effective time-frequency analysis method to reveal the low-frequency anomalies in seismic data. However, for broadband seismic signals, the W transform (WT) exists as an obvious spectral energy-splitting artifact at the dominant frequency with the peak frequency shifted to both ends. We develop a normalized WT to address this splitting artifact by multiplying the WT with the standard deviation normalized along the dominant frequency. Because the normalized standard deviation tends to decrease outward along the dominant frequency, the new transform concentrates its spectral energy around the dominant frequency with high time-frequency resolution, especially the temporal resolution that depends on the detailed time-varying dominant frequency. By analyzing numerical examples and 2D seismic data of a gas field in the western Sichuan Basin, China, we illustrate the effectiveness of the new transform and highlight its advantages over the Stockwell-, W-, and three-parameter W transforms, providing better gas-prone interpretation and reservoir delineation. In addition, due to the high temporal resolution, the new transform is suitable for other geophysical analysis, such as thin-bed identification.