Imaging subsurface migration of dissolved CO 2 in a shallow aquifer using 3-D time-lapse electrical resistivity tomography Esben Auken a , Joseph Doetsch a, ⁎, Gianluca Fiandaca a , Anders Vest Christiansen a , Aurélie Gazoty a , Aaron Graham Cahill b , Rasmus Jakobsen c a Department of Geosciences, Aarhus University, Aarhus, Denmark b Department of Environmental Engineering, Technical University of Denmark, Copenhagen, Denmark c GEUS—Geological Survey of Denmark and Greenland, Copenhagen, Denmark abstract article info Article history: Received 26 September 2013 Accepted 27 November 2013 Available online 4 December 2013 Keywords: ERT Monitoring Time-lapse inversion CO 2 Groundwater Electrical resistivity Contamination of groundwater by leaking CO 2 is a potential risk of carbon sequestration. With the help of a field experiment in western Denmark, we investigate to what extent surface electrical resistivity tomography (ERT) can detect and image dissolved CO 2 in a shallow aquifer. For this purpose, we injected CO 2 at a depth of 5 and 10 m and monitored its migration using 320 electrodes on a 126 m × 25 m surface grid. A fully automated acquisition system continuously collected data and uploaded it into an online database. The large amount of data allows for time-series analysis using geostatistical techniques for noise estimation and data interpolation to compensate for intermittent instrument failure. We estimate a time-dependent noise level for each ERT configuration, taking data variation and measurement frequency into account. A baseline inversion reveals the geology at the site consisting of aeolian and glacial sands near the surface and marine sands below 10 m depth. 3-D time-lapse ERT inversions clearly image the dissolved CO 2 plume with de- creased electrical resistivity values. We can image the geochemical changes induced by the dissolved CO 2 until the end of the acquisition, 120 days after the injection start. During these 120 days, the CO 2 migrates about 25 m in the expected groundwater flow direction. Water electrical conductivity (EC) sampling using small screens in 29 wells allows for very good verification of the ERT results. Water EC and ERT results generally agree very well, with the water sampling showing some fine-scale variations that cannot be resolved by the ERT. The ERT images have their strength in outlining the plume's shape in three dimensions and in being able to image the plume outside the well field. These results highlight the potential for imaging dissolved CO 2 using non-intrusive surface electrical resistivity tomography. © 2013 Elsevier B.V. All rights reserved. 1. Introduction Geological carbon sequestration is a promising technique for reducing CO 2 release into the atmosphere by capturing the CO 2 (e.g., at power plants) and injecting it into deep reservoirs for long-term storage. Potential storage formations are abandoned oil and gas fields, saline formations and coal beds (Benson et al., 2005). Irrespec- tive of the storage formation, the reservoirs must be sealed by several layers of fully intact cap rock that prevent leakage into shallow forma- tions. The risk of leakage from properly chosen reservoirs with adequate cap rock is very small. Nevertheless, it is important to monitor the migra- tion of the CO 2 for the safe and efficient operation of underground CO 2 storage (Benson et al., 2005). Efficient operation of the CO 2 injection requires monitoring and simulation of the migrating CO 2 in the reservoir. Monitoring outside the reservoir is mostly important for leakage detection and serves regulatory as well as public perception purposes. Although unlikely, the increased reservoir pressure can lead to leakage of reservoir brine (high salinity water) or CO 2 into shallow aquifers (Birkholzer et al., 2009). Therefore, before permitting any geological CO 2 storage, monitoring strategies for leaked brine and CO 2 need to be in place. Brine leakage is often considered more critical and its modeling (Nicot, 2008) and detection (Günther et al., 2013) is currently being investigated. Reservoir brine is mostly of high salinity and can therefore increase salinity levels in groundwater above acceptable drinking water limits. At the same time, leaked brine is characterized by a strong decrease in electrical resistivity, due to the dissolved salt. Here, we concentrate on monitoring the CO 2 itself and not the brine it is replacing. Under reservoir conditions with temperatures above 31 °C and pres- sures above 73.8 bars, CO 2 exists in a supercritical phase, which partly behaves like a fluid, partly like a gas. The potential of geophysical methods for monitoring of supercritical CO 2 in the reservoir has been demonstrated in several pilot storage studies (e.g., Giese et al., 2009; Journal of Applied Geophysics 101 (2014) 31–41 ⁎ Corresponding author at: Hydrogeophysics Group, Aarhus University, C. F. Møllers Allé 4, DK-8000 Aarhus, Denmark. Tel.: +45 87162373. E-mail address: joseph.doetsch@geo.au.dk (J. Doetsch). 0926-9851/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jappgeo.2013.11.011 Contents lists available at ScienceDirect Journal of Applied Geophysics journal homepage: www.elsevier.com/locate/jappgeo