Amplitude Variation with Angle Inversion Based on Propagator Matrix Modelling

Most existing amplitude variation with angle (AVA) inversions are based on the exact Zoeppritz equation and its linearized approximations. However, such modellings consider the primary (P-wave) waves only and ignore transmission loss, multiples and many other converted wave modes. It means that strict pre-processing is required to satisfy the requirement of the input data to inversions, which is difficult to accomplish. To overcome this problem, a new AVA inversion based on the propagator matrix method of Carcione (PMC) is proposed in this abstract. PMC is derived from the 1D wave equation and can accurately simulate full wave responses within horizontally multi-layered models. For computational accuracy, we use a Bayesian framework to define the augmented function and choose the L-BFGS optimization method to solve it. Through a model test and real data application, it is found that the proposed AVA inversion based on PMC can effectively eliminate transmission loss and internal multiples and helps to provide better S-wave velocity and density estimations.