High-resolution time frequency analysis using Wigner-Ville Distribution and the Maximum Entropy Method: Application for gas and hydrates identification

Abstract

Time-frequency (TF) analysis can reveal important details of seismic data and provide valuable information for reservoir characterization. Resolution in the TF plane is clearly critical for interpretation. Many methods applied to timefrequency representation introduce spurious or cross-terms, essentially when using bilinear functions such as the Wigner Ville Distribution (WVD). Most techniques proposed to overcome this shortcoming use a smoothed kernel which in turn can adversely affect the component’s concentration in the TF plane. We propose to apply the Maximum Entropy Method (MEM) to WVD to obtain a robust and high resolution time-frequency representation of seismic traces, and we also introduce a formula to estimate instantaneous frequency (IF) in time domain. We apply our approach to the spectral decomposition and IF analysis of a seismic dataset from the Gulf of Mexico. The analysis allows us to identify hydrates and a gas pocket.