Developing estimation frameworks that integrate indirect and direct measurements during remediation

Sources:  Susan Hubbard & Dan Hawkes

To use the correlations between geophysical signatures and subsurface biogeochemical transformations for quantitatively estimating biogeochemical states and transformations, other ESD research has focused on developing estimation frameworks that can integrate the spatially extensive (but indirect) geophysical data with the direct (but sparse) wellbore measurements, both collected over time during a remediation treatment. To meet this objective, ESD’s Jinsong Chen and coworkers (Chen et al., 2009) developed a state-space Bayesian estimation approach that uses Markov Chain sampling methods to find solutions. The state-space approach was applied to time-lapse SIP data collected during laboratory column biostimulation experiments that precipitated iron and zinc sulfides during sulfate reduction. Chen and his team developed a dynamic petrophysical model to link the time-lapse SIP signatures with precipitate development and associated permeability reduction. Consistent with laboratory findings, Chen’s petrophysical model assumed that once dispersed FeS and ZnS precipitates were initially formed, they were followed by a coating of dispersed microbial cells with FeS and ZnS precipitates, and then by the formation of precipitate-coated dispersed-cell clusters that clogged pore spaces. Figure 4 illustrates the estimation results, including the geophysically obtained evolution of precipitate volume fraction and permeability at a single-column sampling point. The figures show that the geophysically obtained estimates compare favorably with measurements obtained using geochemical or hydrological datasets collected during the column experiment. This study suggests that the developed state-space approach permits the use of geophysical datasets for quantifying the evolution of both remediation-induced end products and hydrological feedbacks. More recently, Chen et al. (WRR, submitted 2011) has extended this methodology to use with time-lapse surface SIP datasets, and has successfully applied it to the Rifle, CO field dataset, as shown in Figure 4.

Figure 4. Comparison of geophysically obtained estimates of  the volume fraction of evolved, dispersed precipitates (left) and the associated reduction in permeability (right) over the duration of a biostimulation experiment that led to the production of sulfides. The geophysically obtained estimates are compared to direct geochemical (left) and hydrological (right) measurements made during the column experiments. Dashed lines indicate uncertainty associated with the geophysically obtained estimates. This study (Chen et al., 2009) is the first to illustrate the use of time-lapse geophysical data for quantifying complex and coupled biogeochemical-hydrological properties associated with in situ remediation treatments.