Bayesian lithology/fluid prediction constrained by spatial couplings and rock physics depth trends

摘要

A Bayesian lithology/fluid inversion approach based on prestack seismic data and well observations is presented. The model includes rock physics depth trends and a spatially coupled prior model for the lithology/fluid characteristics. Moreover, uncertainties in global model parameters are included. A McMC algorithm that converges reasonably fast is defined for assessing the posterior lithology/fluid model. The inversion approach is evaluated on a real case from the North Sea.