Interval Q factor estimation based on an improved frequency shift method

Abstract

To characterize seismic attenuation information quantitatively, the qualify factor Q is routinely used in seismic exploration. To obtain a more reliable Q value than the peak frequency shift (PFS) method, the improved frequency shift (IFS) method computes the peak frequency by the centroid frequency (CF) to improve the noise immunity. However, the IFS method assumes all wavelets are the standard Ricker shape and ignores the effect of attenuation on the wavelet, which greatly affects the accuracy of estimated Q values, especially in interlayer computations. Some analytical methods try to replace the original hypothesis with attenuated Ricker wavelets, but there are too many mathematical approximations, resulting in inaccurate analytical results. Here, a new formula is derived by matching the spectrums with the attenuated Ricker wavelets. The peak frequencies (PFs) of the reference and attenuated wavelet are calculated by the Gauss-Newton iteration method. Combined with fitted PFs, we derive another new formula to calculate the interval-Q value directly. With these modifications, the PFS improved (PFSI) method not only improves the accuracy of the Q factor but also has noise robustness. The synthetic data tests and field data application demonstrate the feasibility and effectiveness of the proposed method.