Modelling the electrical resistivity response to CO2 plumes generated in a laboratory, cylindrical sandbox

CO2 capture and storage into deep geological formations is one of the main solutions proposed to reduce the concentration of anthropic CO2 in the atmosphere. The monitoring of injection sites is a crucial issue to assess for the long term viability of CO2 storage. With the intention of detecting potential leakages, we are investigating the possibility of using electrical resistivity tomography (ERT) techniques to detect CO2 transfers in the shallow sub-surface. ERT measurements were performed during a CO2 injection in a cylindrical tank filled with Fontainebleau sand and saturated with water. Several measurements protocols were tested. The inversion of the resistances measured with the software R3T (Binley and Kemna (2005)) clearly showed that the CO2 injection induces significant changes in the resistivity distribution of the medium, and that ERT has a promising potential for the detection and survey of CO2 transfers through unconsolidated saturated media. We modeled this experiment using Matlab by building a 3D cellular automaton that describes the CO2 spreading, following the geometric and stochastic approach described by Selker et al. (2007). The CO2 circulation is described as independents, circular and continuous gas channels whose horizontal spread depends on a Gaussian probability law. From the channel distribution we define the corresponding gas concentration distribution and calculate the resistivity of the medium by applying Archie's law for unsaturated conditions. The forward modelling was performed with the software R3T to convert the resistivity distribution into resistances values, each corresponding to one of the electrode arrays used in the experimental measurements. Modelled and measured resistances show a good correlation, except for the electrode arrays located at the top or the bottom of the tank. We improved the precision of the model by considering the effects due to CO2 dissolution in the water which increases the conductivity of the saturated water in the vicinity of the plume. Finally, a special interest is given to the electrode configurations and measurement protocols. A numerical study was realized using R3T to assess the information that each electrode configuration can provide. We built a map of their sensitivity to resistivity changes. From these results, we studied the resolution of several electrode combinations by inverting the associated resistance in order to reduce the number of data really necessary to get a good imaging of the plume and thus accelerate the acquisition time.