Nonlinear three-term AVO inversion based on exact Zoeppritz equation

Traditional three-term AVO inversion is usually inaccurate for density inversion, while density is vital in fluid identification, fluid saturation calculation, porosity calculation of conventional reservoirs, TOC content calculation, and cracks prediction of unconventional reservoirs. Therefore, utilizing amplitude information of large offset data and PS wave data, which include abundant density information, is important to improve the stability and accuracy of density inversion results. Based on framework of Bayesian inversion, we introduce Trivariate Cauchy distribution as prior distribution, use exact Zoeppritz equation to construct the objective function of the three-term AVO joint inversion, simplify the objective function by second-order Taylor, and obtain the iterative updating formula of model parameters. Then, we use large offset data to improve inversion results, and realize PP and PS data joint inversion. Synthetic data and field data test results show that this new method not only can directly invert P-wave velocity, S-wave velocity and density, but also has high accuracy, especially for the density inversion results. By comparing PP-wave or PS-wave model data inversion results with PP and PS-waves joint inversion results, it is obvious that joint inversion has better stability, higher precision and stronger antinoise ability. This demonstrates feasibility and effectiveness of the method. Compared with Aki-Richards approximate formula inversion results, it is noted that this method has higher precision and better antinoise ability.