High Resolution Time-Reassigned SynchroSqueezing Transform with Iterative Group Delay Estimation

Abstract

Abstract High-resolution post-processing methods for spectral decomposition play an important role in seismic signal analysis, enabling the detection of low-frequency shadows (LFS) associated with hydrocarbon accumulations. However, conventional timeߝfrequency transforms often suffer from energy smearing and limited temporal resolution when analyzing nonstationary seismic data. To address these limitations, the High-Resolution Time-Reassigned Synchrosqueezing Transform (HRTSST) is proposed, incorporating iterative group-delay refinement, magnitude-based stabilization, and selective energy compaction. These mechanisms improve the localization of timeߝfrequency coefficients while maintaining signal interpretability under interference and noise. Synthetic experiments demonstrate improved spectral sparsity and reduced mode mixing compared with established methods, including the Time-Reassigned Synchrosqueezing Transform (TSST), Time-Reassigned Multi-Synchrosqueezing Transform (TMSST), and Second-Order Transient-Extracting Transform (STET). Application to a 3D seismic dataset from the F3 block in the North Sea further confirms the capability of HRTSST to delineate LFS anomalies consistent with gas-bearing intervals identified from well logs and seismic amplitude attributes. The resulting timeߝfrequency representation suppresses spurious dispersion and improves spectral contrast, supporting more reliable hydrocarbon detection and reservoir delineation in complex structural settings. These results demonstrate the potential of HRTSST as an effective tool for seismic spectral analysis and subsurface characterization in exploration geophysics.