Spectral Decomposition of Seismic Data With Variational Mode Decomposition-Based Wigner–Ville Distribution

Abstract

Wigner–Ville distributions (WVDs), which provide superior time-frequency resolution and energy distribution concentration relative to other traditional Fourier-based or wavelet-based seismic time-frequency methods, are an important tool for spectral decomposition and have the potential to yield better seismic interpretations for highlighting geophysical responses in particular frequency bands. However, the existence of cross-term interference in WVDs limits their application. To effectively suppress the cross-term interference in a WVD without reducing the time-frequency resolution and energy aggregation, we propose a variational mode decomposition (VMD) based WVD approach. VMD is first used to decompose the multicomponent seismic data into a series of narrow band limited intrinsic mode functions (IMFs). Next, we calculate the WVDs of these IMFs. Finally, the maximum amplitude volume above the average amplitude and the peak frequency volume are extracted from these WVDs for seismic interpretation. A synthetic data example demonstrates the effectiveness of the VMD-based WVD approach and its superiority comparison to the WVD, smoothed pseudoWVD and empirical mode decomposition based WVD approaches. Real seismic data applications show that spectral decomposition with VMD-based WVD has a strong ability to highlight hydrocarbon-related information.