December 22, 2004
Journal Article

Geochemical Characterization Using Geophysical Data and Markov Chain Monte Carlo Methods: A Case Study at the South Oyster Bacterial Transport Site in Virginia

Abstract

The paper demonstrates the use of ground-penetrating radar (GPR) tomographic data for estimating extractable Fe(II) and Fe(III) concentrations using a Markov chain Monte Carlo (MCMC) approach, based on data collected at the DOE South Oyster Bacterial Transport Site in Virginia. Analysis of multidimensional data including physical, geophysical, geochemical, and hydrogeological measurements collected at the site shows that GPR attenuation and lithofacies are most informative for the estimation. A statistical model is developed for integrating the GPR attenuation and lithofacies data. In the model, lithofacies is considered as a spatially correlated random variable and petrophysical models for linking GPR attenuation to geochemical parameters were derived from data at and near oreholes. Extractable Fe(II) and Fe(III) concentrations at each pixel between boreholes are estimated by conditioning to the co-located GPR data and the lithofacies measurements along boreholes through spatial correlation. Cross-validation results show that geophysical data, constrained by lithofacies, provided information about extractable Fe(II) and Fe(III) concentration in a minimally invasive manner and with a resolution unparalleled by other geochemical characterization methods. The developed model is effective and flexible, and should be applicable for estimating other geochemical parameters at other sites.

Revised: April 5, 2005 | Published: December 22, 2004

Citation

Chen J., S. Hubbard, Y. Rubin, C.J. Murray, E.E. Roden, and E.L. Majer. 2004. Geochemical Characterization Using Geophysical Data and Markov Chain Monte Carlo Methods: A Case Study at the South Oyster Bacterial Transport Site in Virginia. Water Resources Research Vol. 40, no. No. 12:W12412. PNNL-SA-43212.