Groundwater‐surface water interaction, dissolved organic carbon oxidation and dissolution in carbonate aquifers

Abstract The high primary porosity and permeability of eogenetic karst aquifers permit water recharged through secondary dissolution features to be temporarily stored in aquifer matrix porosity. The recharged water contains elevated dissolved organic carbon (DOC) concentrations that, when oxidized, enhance limestone dissolution and impact carbon cycling. We evaluate the relationship between DOC oxidation and limestone dissolution using observations at a stream sink‐rise system and reversing spring in the Floridan aquifer, north‐central Florida, USA, where subsurface residence times of recharged water are days and months, respectively. We estimate water chemical compositions during surface water‐groundwater interactions at these two systems with mixing models of surface water and groundwater compositions and compare them with measured DOC, dissolved inorganic carbon (DIC), Ca 2+ and dissolved organic nitrogen (DON) concentrations. Differences between measured and modelled concentrations represent net changes that can be attributed to calcite dissolution and redox reactions, including DOC oxidation. DOC losses and Ca 2+ gains exhibit significant (p < 0.01) inverse linear correlations at both the reversing spring (slope = −0.9, r 2 = 0.99) and the sink‐rise system (slope = −0.4, r 2 = 0.72). DOC oxidation in both systems was associated with decreases in the molar C:N ratio (DOC:DON). Significant (p < 0.01) positive linear correlations between increases in Ca 2+ and DIC concentrations after correcting for DIC derived from calcite dissolution occurred at both the reversing spring (slope = 1.3, r 2 = 0.99) and the sink‐rise system (slope = 1.61, r 2 = 0.75). Greater deviations from the expected slope of −1 or +1 at the sink‐rise system than at the reversing spring indicate DOC oxidation contributes less dissolution at the sink‐rise system than at the reversing spring, likely from shorter storage in the subsurface. A portion of the deviation from expected slope values can be explained by the dissolution of Mg‐rich carbonate or dolomite rather than pure calcite dissolution. Despite this, slope values reflect kinetic effects controlling incomplete consumption of carbonic acid during dissolution reactions.