Multicomponent prestack joint AVO inversion based on exact Zoeppritz equation

The Zoeppritz equation describes the relationship between reflection and transmission coefficients, incidence angles and elastic properties. Currently, most of the prestack joint AVO or AVA (amplitude variation with offset or angle) inversion methods are based on various approximations of the exact Zoeppritz equation. However, because these approximations have certain assumptions and limitations, such as limited incident angles or weak contrast of elastic parameters on either side of the interface, the conventional linear prestack joint AVO inversion based on different approximations isn't suitable, which reduces the inversion accuracy theoretically. In this paper, we develop a multi-objective prestack joint AVO inversion method using both PP and PS seismic data based on the exact Zoeppritz equation. To solve the nonlinear multi-objective inversion problems, we use a developed fast nondominated sorting genetic algorithm (NSGA II) to estimate three elastic parameters, including P-wave velocity, S-wave velocity and density. Compared with the conventional weighted stacking optimization means of objective functions, this approach is robust because it can simultaneously cope with more than one objective function without introducing weight coefficients. Model tests illustrate that the nonlinear multi-objective prestack joint AVO inversion method shows great potential to estimate three elastic parameters with the exact Zoeppritz equation. We also analyze the influence of the initial model and the search window on the inversion results. Additionally, the seismic data with random noise examples reveal the stability and anti-noise interference ability of our joint AVO inversion method.