|
地球物理学进展 2009
Uncertainty analysis for AVO inversion
|
Abstract:
More abundant and effective lithologic information can be obtained from inversion of pre-stack data in comparison with from ordinary post-stack impedance inversion. The noise and other factors of pre-stack data can heavily influence the stability of the inversion problem. So it is very important to predict the reliability of AVO estimates. Based on Bayesian theorem, it is initially assumed that the prior probability distributions are uniform and the noise is independent and Gaussian. In this case, the uncertainty analysis equation can be obtained. The variance in parameter covariance matrix can be used to investigate the uncertainty in the problem. The modeling studys show that the uncertainty of AVO inversion is a function of the prestack noise, the ratio of the P- to S-wave velocity, the fold and the range of angle available for inversion. The predicted errors by the method give interpreters tools to help quantify the reliability of their AVO interpretation.