Geologic Carbon Storage in the Lower Ordovician Arbuckle Group Saline Aquifer in Kansas

A small scale field test at Wellington oil field operated by Berexco LLC in Sumner County, Kansas will inject up to 40,000 tonnes (770,000 MCF) of CO₂ over approximately nine months into the lower 46 m thick porous and permeable interval of the Gasconade Dolomite in the 300 m thick Lower Ordovician Arbuckle Group saline aquifer. The objective is to validate the predicted CO₂ plume based on a 3D geocellular geomodel obtained from a 3D multi-component seismic volume and two new wells (one cored and both tested) drilled to basement. The plume behavior obtained from a compositional simulation will be tested and validated through visualization, sampling, and indirect detection using an array of monitoring technologies including in situ fluid sampling via U-tube, downhole continuous active seismic monitoring (CASSM), and vertical seismic profiling using acoustic recording from a continuous fiber optic cable installation. Also, CGPS and InSAR for potential surface deformation and a seismometer array will be installed. The CO₂ injection, funded by DOE-NETL and cost sharing partners, is pending approval of a Class VI geosequestration permit.

Besides mapping the CO₂ plume dispersal, the test will estimate the tonnage of CO₂ sequestered and refine estimates of the regional carbon storage in the Arbuckle in southern Kansas. The current estimates range between 9-75 billion tonnes in a 9650 km² area that has been extensively characterized by geological and geophysical methods.

Wells in the overlying Mississippian oil reservoir at 460 m above the injection zone and a nest of wells in the USDW will be monitored to detect any leakage. Encompassing static and dynamic models that incorporate Arbuckle injection, caprocks, and freshwater aquifers will evaluate flow, storage, seals, and risk.  Best practice methodologies for modeling, monitoring, verifying, and accounting of the CO₂ should provide a framework for carbon management in the midcontinent.