An effective Bayesian model for lithofacies estimation using geophysical data

A Bayesian model coupled with a fuzzy neural network (BFNN) is developed to enhance the use of geophysical data in lithofacies estimation. Prior estimates are inferred from borehole lithofacies measurements using indicator kriging, and posterior estimates are obtained by updating the prior using geophysical data. The novelty of this study lies in the use of the fuzzy neural network for the inference of the likelihood function. This allows spatial correlation of lithofacies as well as nonlinear cross correlation between lithofacies and geophysical attributes to be incorporated into lithofacies estimation. The effectiveness of BFNN is demonstrated using synthetic data emulating measurements at the Lawrence Livermore National Laboratory (LLNL) Site.