Assessing the temporal evolution of dissolved inorganic carbon in waters exposed to atmospheric CO2(g): A laboratory approach

• We investigated carbon cycling in artificial and natural water samples exposed to the atmosphere.
• Dissolved inorganic carbon (DIC) concentration decreased by C loss and increased from evaporation.
• DIC concentration increases and decreases were accompanied by carbon isotope (δ13C) enrichment.
• We generated conceptual models of the DIC and δ13C changes for carbon cycling in surface water.
• We show how kinetic and/or equilibrium isotopic fractionation can account for the δ13C increases.

SummaryDocumenting the transformation of dissolved inorganic carbon (DIC) during the interaction of surface waters (e.g., rivers, lakes) with atmospheric CO2(g) is vital for understanding carbon cycling. Investigations that mimic the continuum of changes in DIC concentrations and stable carbon isotope ratio of DIC (δ13CDIC) to equilibrium with atmospheric CO2(g) are difficult to conduct in natural settings because of multiple processes that occur in the water column, the interaction between water and sediments or rocks in stream channels and lake beds, as well as the variability in water residence times. Thus, laboratory simulations of the spectrum of DIC transformation provide insights which reduce the ambiguity in describing the mechanisms that control the behavior of DIC during surface water-atmospheric CO2(g) interaction. To test how surface water-atmospheric CO2(g) interaction affects DIC concentrations and δ13CDIC, we used three types of samples: (1) we prepared an artificial solution using NaHCO3 where the DIC concentration is near chemical equilibrium and the δ13CDIC is far from isotopic equilibrium with atmospheric CO2(g), (2) natural groundwater where the DIC concentration and the δ13CDIC are both sufficiently far from chemical and isotopic equilibrium with atmospheric CO2(g) and (3) lake water where the DIC concentration and the δ13CDIC are near chemical and isotopic equilibrium with atmospheric CO2(g). These samples allowed us to ascertain when only chemical or isotopic changes are occurring, or when both chemical and isotopic changes are occurring. The NaHCO3 solution was prepared by dissolving ∼6 g of laboratory grade NaHCO3 salt in 20 L of deionized water. Groundwater was collected from Stillwater, Oklahoma (36°08′22.20″N, 97°03′22.66″W) and lake water was collected from Lake McMurtry, Stillwater, Oklahoma (36°10′49.37″N, 97°10′52.9″W). The solution of NaHCO3, and groundwater (potential source of surface water) and lake water samples were exposed to the atmosphere in a laboratory setting for 850–1000 h until their DIC attained chemical and isotopic equilibrium with atmospheric CO2(g). All samples were prepared in duplicate and one set was agitated to simulate mixing in surface waters. The DIC concentrations of the NaHCO3 samples increased without C loss and the δ13CDIC was enriched to a steady state for the mixed sample. The increase in the DIC concentrations was modeled as evaporation and not as CO2(g) invasion since the pCO2 was higher than atmospheric throughout the experiment. The enrichment in the δ13CDIC was modeled as equilibrium carbon isotopic exchange with atmospheric CO2(g). The DIC concentrations in the mixed groundwater sample initially decreased due to CO2(g) outgassing and the accompanying enrichment in δ13CDIC was modeled as kinetic isotopic fractionation. After the initial decrease, the DIC concentrations increased continuously while the δ13CDIC was enriched to a steady state. Overall, the unmixed groundwater sample showed similar temporal δ13CDIC trends to the mixed groundwater sample, even though the unmixed sample did not achieve isotopic equilibrium with atmospheric CO2(g). Both the mixed and unmixed lake samples showed only small increases in temporal DIC concentrations and a slight initial decrease, followed by a small enrichment in the δ13CDIC during the experiment. The minor changes suggest that the lake samples were closer to chemical and carbon isotopic equilibrium with atmospheric CO2(g). The results of this study would apply in settings where the predominant process controlling carbon cycling is the interaction between the surface water DIC and atmospheric CO2(g).