Application of the local maximum synchrosqueezing transform for seismic data

Abstract Seismic signal analysis is the main step in data processing through the petroleum exploration via costly seismic investigations. Precision of target delineation by seismic data for exploratory drilling strongly depends on resolution of the seismic image. However, resolution of the seismic image is restricted by the band-limited nature of the seismic signal and inherent deficiencies in signal enhancement and random noise attenuation methods. Generally, it is required to increase resolution of the seismic data both in time and frequency domains. General solution to achieve this goal is employing time – frequency transformation (TFT) methods. The most applicable and conventional TFT method is the short time Fourier transform (STFT). Nevertheless, the STFT method does not provide sufficient resolution for further seismological interpretation on data. Thus, reassignment and simultaneous synchrosqueezing transform methods were introduced that increase resolution in time by reassignment of coefficients to their true position. However, they are still not appropriate for obtaining required resolution in seismic signal. The local maximum synchrosqueezing transform (LMSST) was introduced as an efficient method for signal analysis. The LMSST method yet was not used for analysis of seismic data and the procedure of its parameters optimization for applying on seismic data is unclear. Therefore, in the presented study, we propose a strategy for defining optimum parameters of the LMSST method and the procedure of its application on seismic data, where high resolution seismic image is required for hydrocarbon exploration. The proposed strategy was applied on two synthetic data examples, considering the quality factor expression and contamination by random noise and a field data example from a natural gas reservoir. Result of applying the proposed strategy on synthetic and field data examples and comparison of results with the leading-edge methods, revealed that this strategy could be considered as an alternative to the common signal enhancement methods. Additionally, the proposed method consists of the advantage of easy implementation and full reconstruction of the original signal. It was also proved that the proposed strategy is a robust method against added random noise compared to other competitive methods when applied on nonstationary seismic signal.