TQWT and WDGA: innovative methods for ground roll attenuation

The existence of noise in seismic data has an undesirable effect on seismic interpretations. An important type of coherent noise is ground roll, in which time and frequency overlap with useful seismic signals. The tunable Q-factor wavelet transform is a new discrete wavelet transform method and is considered as a new tool for ground roll attenuation. The simple tunable Q-factor wavelet transform by using a single wavelet with explicit oscillatory behavior is utilized for each trace. The enhanced tunable Q-factor wavelet transform (for seismic signals), which benefits from several mother wavelets with distinct Q-factors (for each part of a signal), is presented here and it is shown that it can adjust reliably to the natural non-stationary behavior of a seismic signal in time and space. Random noise has an undesirable effect on Q-factor ranges, with extreme consequences. Wavelet domain ground roll analysis is used as a substitute approach to thresholding for ground roll attenuation. The simple tunable Q-factor wavelet transform and the enhanced tunable Q-factor wavelet transform are compared with each other and with the f-k method by using synthetic and real data samples. (paper)