GEOLOGIC CARBON SEQUESTRATION RATES IN HYPERALKALINE LAKES: EOCENE GREEN RIVER FORMATION, NORTH AMERICA, AND HOLOCENE LAKE MAGADI, KENYA

With atmospheric CO₂ rising above 400ppm, global demand for carbon sequestration is growing. Promising avenues include land use and agricultural changes, supercritical CO₂ injection into subsurface reservoirs, and induced mineralization from solution. Mineral precipitation depends on abundant alkalinity; efforts to date have generally focused on precipitation of Ca or Mg carbonates, for example through fluid interaction with mafic rocks. The high solubility of Na salts has limited the usefulness of Na in possible mineralization reactions, but hyperalkaline lakes are environments in which sodium carbonate can precipitate.

Na carbonate deposits of the Eocene Green River Formation, western North America, indicate episodes of rapid deposition of C during some time intervals. Using published Ar/Ar chronologies (Smith et al., 2003), we estimate the Eocene CO₂ sequestration rate averaged ~40 kT CO₂/yr based on trona reserve volumes (Wig et al., 1995). Based instead on an inferred Na flux model of Smith et al. (2008), we estimate a somewhat higher rate of ~100 kT CO₂/yr.

The Late Pleistocene Magadi Basin in southern Kenya is another basin with substantial Na carbonate deposits and an available geochronology. Based on our estimates of basin morphology and trona thicknesses reported by Baker (1958), and published ¹⁴C dates (Hay, 1968), we estimate the average Holocene carbon dioxide burial rate was ~100 kT CO₂/yr in the lake.

In both of these paleolakes it is likely that magmatic CO₂ contributed to the dissolved inorganic carbon load, so not all the sequestered carbon was atmospheric. Nevertheless, these lakes demonstrate that abiotic evaporatively-driven processes can concentrate and precipitate substantial amounts of inorganic carbon.

These rough estimates of paleo-carbon sequestration rates are not far from the US Department of Energy target sequestration rates of 1 MT CO₂/year for viable geologic injection wells such as NETL’s injection test in Michigan, or the Sleipner (North Sea) Injection Field which has injected about 900 kT CO₂/yr since 1996. With the proper inflow chemistry, volume, and evaporation rates, therefore, natural or engineered hyperalkaline lakes can potentially sequester a significant mass of atmospheric carbon as Na carbonate minerals, potentially with low energy costs.