System and method for prediction of reservoir parameters with uncertainty quantification

Abstract

A method is described for seismic inversion with uncertainty quantification including performing low frequency Markov Chain Monte Carlo (MCMC) processes on rock physics models to generate low frequency models (LFMs) of rock properties and training a deep neural network using the low frequency models and synthetic seismograms to generate a trained neural network. Given a seismic dataset, the trained neural network can generate a high frequency rock property model and then broad-band MCMC processes can be performed on the high frequency rock property model for uncertainty quantification. The method is executed by a computer system.