A 20 million year record of planktic foraminiferal B/Ca ratios: Systematics and uncertainties in pCO2 reconstructions

We use new and published data representing a 20 million long record to discuss the systematics of interpreting planktic foraminiferal B/Ca ratios. B/Ca-based reconstructions of seawater carbonate chemistry and atmospheric pCO₂ assume that the incorporation of boron into foraminiferal tests can be empirically described by an apparent partition coefficient, (Hemming and Hanson, 1992). It has also been proposed that there is a species-specific relationship between K_D and temperature (Yu et al., 2007). As we discuss, although these relationships may be robust, there remain significant uncertainties over the controls on boron incorporation into foraminifera. It is difficult to be certain that the empirically defined correlation between temperature and K_D is not simply a result of covariance of temperature and other hydrographic variables in the ocean, including carbonate system parameters. There is also some evidence that K_D may be affected by solution ratios (i.e., pH), or by . In addition, the theoretical basis for the definition of K_D and for a temperature control on K_D is of debate. We also discuss the sensitivity of pCO₂ reconstructions to different K_D-temperature calibrations and seawater B/Ca. If a K_D-temperature calibration is estimated using ice core pCO₂ values between 0 and 200 ka, B/Ca ratios can be used to reasonably approximate atmospheric pCO₂ between 200 and 800 ka; however, the absolute values of pCO₂ calculated are sensitive to the choice of K_D-temperature relationship. For older time periods, the absolute values of pCO₂ are also dependent on the evolution of seawater B concentrations. However, we find that over the last 20 Ma, reconstructed changes in declining pCO₂ across the Mid-Pleistocene Transition, Pliocene glacial intensification, and the Middle Miocene Climate Transition are supported by the B/Ca record even if a constant coretop K_D is used, or different K_D-temperature calibrations and models of seawater B evolution are applied to the data. The inferred influence of temperature on K_D from coretop data therefore cannot itself explain the structure of a published pCO₂ reconstruction (Tripati et al., 2009). We conclude the raw B/Ca data supports a coupling between pCO₂ and climate over the past 20 Ma. Finally, we explore possible implications of B/Ca-based pCO₂ estimates for the interpretation of other marine pCO₂ proxies.