Oxygen and carbon isotope fractionation in calcitic deep-sea corals: Implications for paleotemperature reconstruction

Inhabiting areas of the ocean where paleoenvironmental records are sparse, deep-sea corals represent valuable yet largely untapped Holocene records of intermediate and deep ocean variability. $δ$18O and $δ$13C were analyzed in nine live-collected deep-sea gorgonian corals (Isididae and Coralliidae) in order to further develop the “lines” paleotemperature method. Least squares linear regression analysis for full lifespan $δ$18O vs. $δ$13C (corrected for $δ$18Owater and $δ$13CDIC) was utilized to yield equations of the form y = mx + b. $δ$18O and intercept values were found to be a function of temperature, and to approximate calcite $δ$18O equilibrium. The corals in this study extend the previously reported calibration (Hill et al., 2011) over a broader range of temperatures from 5 °C to 11.2 °C. When combined with the data from Hill et al. (2011), a new expression for the relationship between the $δ$18Ointercept value and temperature is proposed: T (°C) = -4.12 ± 0.38 ($δ$18Ointercept) + 12.32 ± 0.75(R2 = 0.90, p value {\textless} 0.0001) Error estimates are ± 0.7 °C for corals living at cold temperatures (2 °C), ± 1.4 °C in warmer waters (11 °C), and ± 0.5 °C at the mean water temperature of the data set (4.6 °C). The first multi-specimen verification of the “lines” method was performed on three co-located bamboo (Isididae) corals and found to give nearly coincident $δ$18O intercepts. Detailed intraspecimen sampling reveals $δ$18O and $δ$13C isotopic variability within coeval portions of the skeleton. In one specimen, “lines” method analysis was utilized on multiple samples taken from the same temporal increment of the skeleton, yielding multiple $δ$18O intercepts. Calculated temperatures using the calibration proposed here describe a temperature range of 7.9 to 10.3 °C, which approaches the temperature range of 11.1 ± 0.7 °C at the coral collection location.