LAUSR

Abstract Prediction of the neutralization of anthropogenic CO2 in the oceans and the interpretation of the calcite record preserved in deep-sea sediments requires the use of the correct kinetics for calcite dissolution. Dissolution rate information from suspended calcite-grain experiments consistently indicates a high-order nonlinear dependence on undersaturation, with a well-defined rate constant. Conversely, stirred-chamber and rotating-disc dissolution experiments consistently indicate linear kinetics of dissolution and a strong dependence on the fluid flow velocity. Here, we resolve these seeming incongruities and establish reliably the kinetic controls on deep-sea calcite dissolution. By equating the carbonate-ion flux from a dissolving calcite bed, governed by laboratory-based nonlinear kinetics, to the flux across typical diffusive boundary layers (DBL) at the seafloor, we show that the net flux is influenced both by boundary layer and bed processes, but that flux is strongly dominated by the rate of diffusion through the DBL. Furthermore, coupling that calculation to an equation for the calcite content of the seafloor, we show that a DBL-transport dominated model predicts lysoclines adeptly, i.e., CaCO3 vs ocean depth profiles, observed across the oceans. Conversely, a model with only sediment-side processes fails to predict lysoclines in all tested regions. Consequently, the past practice of arbitrarily altering the calcite-dissolution rate constant to allow sediment-side only models to fit calcite profiles constitutes confirmation bias. From these results, we hypothesize that the reason suspended-grain experiments and bed experiments yield different reaction orders is that dissolution rates of individual grains in a bed are fast enough to maintain porewaters at or close to saturation, so that the exact reaction order cannot be measured accurately and dissolution appears to be linear. Finally, we provide a further test of DBL-transport dominated calcite dissolution by successfully predicting, not fitting, the in-situ pH profiles observed at four stations reported in the literature.