Seismic Time-frequency Spectral Decomposition by Matching Pursuit to Detect Channels

Seismic data, being non-stationary in nature, have varying frequency content in time. Time-frequency decomposition of a seismic signal aims to characterize the time-dependent frequency response of subsurface rocks and reservoirs. There are a variety of spectral decomposition methods; these include the Short Time Fourier Transform (STFT), Continuous Wavelet Transform (CWT) and Matching Pursuit Decomposition (MPD). In STFT, time-frequency resolution is fixed over the entire time-frequency space by preselecting a window length. Therefore, resolution in seismic data analysis becomes dependent on a user specified window length. The CWT dilates and compresses wavelets to provide a time-scale spectrum instead of a time-frequency spectrum. Converting a scalogram into a time-frequency spectrum using the center frequency of a scale gives an erroneous attenuation in the spectrum. Decomposition (MPD) do not involve windowing of the seismic data and thus have the best combination of temporal and spectral resolution as compared to STFT and the continuous wavelet transform (CWT). Castagna et al, 2003, used matching-pursuit decomposition for instantaneous spectral analysis to detect low-frequency shadows beneath hydrocarbon reservoirs. In this paper we analyze performance of matching pursuit method in mapping channel sediments at one of the south-west oil- field in Iran.