A box model was used to estimate oceanic carbonate system parameters (total dissolved inorganic carbon, alkalinity, CO2(aq), CO3 2−, HCO3 −, and pH) in the surface and deep ocean over the past 20 Ma. Geological data were used to constrain boundary conditions, including the carbonate compensation depth (CCD) and seawater temperature, salinity, and ionic composition. In addition, ∼14,000 δ 13C measurements of planktic and benthic foraminifera were compiled. Sensitivity tests were performed to evaluate potential errors arising from uncertainties in model boundary conditions. Values of surface water pH were modeled assuming four different histories of atmospheric CO2, and compared with proxy-based estimates of pH. The system is underconstrained, and therefore multiple CO2 histories can be invoked to explain the same pH trend, depending on the assumptions made regarding ocean overturning and biological productivity. Despite this uncertainty, we conclude that if CO2 was strongly coupled to temperature and ice volume over the last 20 Ma, predicted values of surface ocean pH fit poorly with current proxy-based pH reconstructions.
Roberts, C. D. & A. K. Tripati, 2009. Modeled reconstructions of the oceanic carbonate system for different histories of atmospheric carbon dioxide during the last 20 Ma. Global Biogeochemical Cycles 23: GB1011. Article (subscription required).
