|
Activity Number:
|
349
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, August 6, 2008 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Physical and Engineering Sciences
|
| Abstract - #301639 |
|
Title:
|
Inverse Modeling of Full-Waveform, Single-Well Geophysical Data Using a Bayesian Model and Markov Chain Monte Carlo Methods
|
|
Author(s):
|
Jinsong Chen*+ and Thomas M. Daley and Carlos Torres-Verdin
|
|
Companies:
|
Lawrence Berkeley National Laboratory and Lawrence Berkeley National Laboratory and The University of Texas at Austin
|
|
Address:
|
1 Cyclotron Road, MS90-1116, Berkeley, CA, 94720,
|
|
Keywords:
|
geophysical inversion ; elastic wave equations ; Bayesian ; MCMC
|
|
Abstract:
|
Single-well seismic methods have been used for environmental investigations and natural resources exploration as cost-effective approaches for subsurface imaging. Current modeling studies primarily focus on numerical simulation of the governing elastic wave equations. In this study, we develop a Bayesian inversion method to estimate elastic parameters and their spatial distribution in the surrounding medium from full-waveform seismic data. We use a staggered-grid finite difference method to forward model full seismic waveforms in a 2D cylindrical coordinate system. We use Metropolis-Hastings and slice sampling methods to explore the joint posterior probability distribution function. Both synthetic and field case studies show that the developed Bayesian model is effective for estimating elastic properties as well as their associated uncertainty.
|
