sulfate reduction, respectively. Importantly, comparison of the ERT and characterization datasets
suggests that biogeochemical transformations preferentially occurred in a thin, high-hydraulic
conductivity zone. SP datasets (i.e. open circuit potential) were also collected in conjunction with
the 2006 experiments conducted in the '2005' flow cell. With this technique, the magnitude of
the cell potential can be estimated using the difference between the cathodic and anodic half-cell
potentials, with a given anodic reaction (e.g. sulfide- vs. iron-oxidation) yielding a value
characteristic of the dominant metabolic process. The measured potentials are distinct for each of
the two wellbore locations, with sulfate-reduction dominating the response downgradient from
the point of acetate injection. The SP phenomena will be presented at the Fall 2006 AGU
meeting (Williams et al., 2006). Although more work is necessary to confirm our complex
resistivity and SP observations, these exciting results further suggest that geophysical methods
should be useful for monitoring system transformations in high resolution and in a minimally
invasive manner.
2c). Petrophysics. In this component of the project, we are investigating the controls on the
geophysical responses through development, augmentation, and testing of petrophysical models.
During this year, we have explored the relationships between complex resistivity and seismic
responses to the development of precipitates. For the complex resistivity, we have decided to
work with a modified Cole-Cole model to link between electrical parameters with the volume of
precipitates that are developed. For the seismic waveform data, are working with a double-
porosity, or patchy saturation model, to relate the seismic attributes to precipitate formation.
2d) Estimation Framework . We are developing a stochastic framework for estimation of time-
lapse geochemical parameters using extensive geophysical data, sparse geochemical data, and
the petrophysical models, which were described in Section 2c. We are using a Markov Chain
approach for this complex, multi-parameter estimation problem that builds on our previous work
and that includes time and zonation associated with heterogeneity. The estimation framework
will be presented at the Fall 2006 meeting (Chen et al., 2006). Development of this estimation
framework will enable a more quantitative interpretation of the geophysical signatures than has
been heretofore possible.
2e) Reactive Transport Modeling. This component focuses on validating and calibrating reactive
transport models against the controlled laboratory experiments. We are using the pore-water
chemistry determined over the course of lab experiments and the solid-phase mineralogy
determined via post mortem analysis to develop a defensible description of the reaction network
(pathways and rates). The reactive transport modeling is carried out using CRUNCH and
TOUGHREACT codes. The reactive transport model is being calibrated to the geophysical data,
but only by using the independent constraints provided by the microbiological, chemical, and
physical data. This is a key step, since the geophysical data will be crucial in developing a high-
resolution data set at the field scale, where complete microbiological, chemical, and physical
characterization of the subsurface material will not be feasible.
To date, modeling of precipitation rates using solute spatial profiles from previously-
conducted column experiments indicates FeS and ZnS accumulation rates that agree closely with
those determined by 0.5 N HCl extraction. Based on these results, the direct reaction of dissolved
H2S with magnetite present in the column appears to be negligible. However, careful analysis of
the reaction stoichiometry of the columns (microbially-mediated sulfate reduction by