Sorry, you need to enable JavaScript to visit this website.

Skip to content Skip to navigation

Computational Geosciences Seminar

Date and Time: 
February 16, 2018 -
12:30pm to 1:30pm
Admission: 

open to all

Location: 
Mitchell 350/372
Contact Email: 
shela.aboud@stanford.edu
Contact Phone: 
650-721-2276
Event Sponsor: 
Center for Computational Earth and Environmental Science

Title: Statistical methods for geophysical inversion and data assimilation problems with applications to CO2 sequestration and near surface geophysics

Speaker: Dario Grana (University of Wyoming)

Abstract: The estimation of subsurface rock and fluid properties from geophysical data is a mathematical inverse problem that requires rock physics and seismic modeling, inverse theory, and spatial statistics. The probabilistic approach to inverse problems provides the posterior distribution of rock and fluid properties given the measured geophysical data and allows quantifying the uncertainty of the predicted results. This modeling problem includes both discrete properties, such as facies or rock-type, and continuous properties, such as porosity, mineral volumes, and saturation. To jointly estimate the posterior distributions of the model variables conditioned by the measured data, we adopt a Bayesian inversion method under the assumption of Gaussian mixture distributions of petrophysical properties and a Markov chain model for the discrete property. This inversion method can be combined with geostatistical algorithms to generate multiple realizations from the resulting Gaussian mixture random field including a spatial continuity model. The proposed methods will be illustrated through applications to CO2 sequestration and near surface geophysics studies.