Joint Inversion of Seismic and Electromagnetic Data Using Statistical Rock-Physics Models and Markov Random Fields

• Activity Number: 655
• Type: Contributed
• Date/Time: Thursday, August 2, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistical Computing
• Author(s): Jinsong Chen*+ and Michael G. Hoversten
• Companies: Lawrence Berkeley National Laboratory and Chevron Energy Technology Company
• Keywords: Inverse problems ; Bayesian ; MCMC ; Monte Carlo ; Geophysical ; Uncertainty

Abstract

Joint inversion of seismic and electromagnetic (EM) data requires rock-physics relationships to link seismic attributes to electrical properties. Ideally, we can connect them through reservoir parameters by developing physical-based models using nearby borehole logs. This is difficult in the exploration stage because information available is insufficient for the development. Conversely, it is easy to derive statistical relationships among geophysical attributes and reservoir parameters from the borehole logs. We develop a Bayesian model to invert seismic and EM data using statistical rock-physics models and Markov random fields. We apply the developed model to a synthetic case that simulates a CO2 monitoring application. We derive statistical rock-physics relations from borehole logs at one location and estimate seismic velocity ratio, acoustic impedance, density, electrical resistivity, lithotypes, porosity, and water saturation at three different locations. Comparison of the inversion results with the true values shows that the correlation-based statistical rock-physics models provide significant information for improving the joint inversion results.