Integration of Bayesian Linearized Inversion into Geostatistical Seismic Inversion

Modeling uncertainty in seismic inversion problems is a topic of interest for both the oil and gas industry and the academia. Although recent advances in methodologies for sampling the posterior space of the petro-elastic properties of interest, integrating the a priori knowledge, they still have high computational cost. Global Stochastic Inversion, an iterative geostatistical seismic inversion methodology, stands out due to its spatial constraining capacity and a priori knowledge integration. However, it is very computationally expensive in searching the model parameter space. On the other hand, Bayesian Linearized Inversion procedures are fast if done trace-by-trace but it inefficient at spatial modeling, specifically when sampling the posterior distribution. This paper proposes a hybrid methodology to tackle the disadvantages of both inversion procedures. Experimental results using a real dataset suggests faster convergence and a better uncertainty modelling when applying the proposed methodology contrary to conventional Global Stochastic Inversion.