Introducing the modular approach for high-resolution time-frequency analysis of seismic signals: application for low-frequency shadow detection

This paper explores the utilization of time-frequency representations in seismic signal analysis to extract valuable insights from seismic signals. The study highlights the importance of high-resolution time-frequency representations in extracting crucial information from seismic data. Specifically, the research focuses on the low-frequency shadow as a seismic attribute, aiding in the identification of gas reservoirs. The low-frequency shadow is characterized by a decrease in frequency and an increase in amplitude near the base of a gas layer in a seismic section. Acknowledging the limitations of conventional methods such as short-time Fourier transform and high-resolution techniques such as synchrosqueezing transform in accurately representing non-stationary seismic signals in time-frequency representation, this paper introduces a novel high-resolution method named as modular high-resolution time-frequency representation. This innovative approach excels in analyzing non-stationary signals, accurately detecting the low-frequency shadow, and signal reconstruction. To demonstrate the effectiveness of the proposed method, the research outlines its algorithm and applies it for comparison with other methods using synthetic seismic data. The results showcase its ability to generate sparse time-frequency maps compared to conventional methods. Furthermore, the study evaluates the effectiveness of the proposed method for low-frequency shadow detection in real seismic data by applying it to 2D seismic data from a gas field in southwestern Iran. The findings emphasize its significant potential for accurately detecting low-frequency shadow locations.