摘要

Due to the viscous-elastic properties of subsurface media, seismic waves experience a loss of amplitude and distortion of phase during propagation. Various methods have been designed to compensate for this attenuation (Q-values) to achieve a high seismic resolution. Traveltime attenuation tomography is an efficient method for estimation of the Q-distribution using inversion. The method updates an initial Q-guess to the true Q-distribution. The Q-update amount is associated with the difference between the simulated traveltime attenuation (SAT) response of the model data and the recorded traveltime attenuation (RAT) of field data. The inversion system is a non-linear process that considers the case in which the travel time of the seismic wave depends on the Q-value to be estimated. Therefore, it is reasonable to calculate the Q-value using the non-linear inversion technique. The accuracy of RAT is essential for Q-tomography. Minute errors in the attenuation of the traveltime can cause unreasonable Q-inversion results. A stabilized method is designed to improve the accuracy of RAT by statistically selecting the optimum frequency band for regression and excluding unreasonable results. Synthetic and field data examples demonstrate the feasibility and effectiveness of this method for Q-estimation. Using a more accurate Q-distribution, seismic resolution can be improved after Q-compensation migration.