Most bivalves and gastropods calcify indistinguishably from dual clumped isotope equilibrium

Molluscan shell-carbonates are extensively used to reconstruct paleo-temperatures at sub-annual resolution. The accurate application of two widely used temperature proxies, the shell carbonate oxygen isotope (ẟ18O) and carbonate clumped isotope (Δ47) composition, is based on the assumption that kinetics in the DIC-H2O-CaCO3 system were either absent or invariant during shell formation and/or can be corrected for via empirical calibration. Here, we analysed the dual clumped isotope composition, i.e., ∆47 and ∆48, of a wide range of modern and Eocene molluscs (bivalves and gastropods) to investigate the potential importance of kinetics during molluscan biomineralisation. We show that Δ47 and Δ48 of most of our modern samples are indistinguishable from equilibrium. For these samples, ∆47-derived temperatures conform to corresponding growth temperatures within their fully propagated 95% uncertainties of ≤ +/-2.3°C. Significant departures from equilibrium values are only obtained for two samples characterised by growth temperatures <10°C. Together, these results strongly imply that bivalve and gastropod shell carbonates represent key archives for accurate and highly precise reconstructions of sea surface temperatures by means of ∆47 clumped isotope thermometry. Kinetic limitations of this thermometer may only become important at relatively low temperatures. ∆ -derived temperatures for our Eocene samples (~39 Ma) from the Hampshire Basin (paleo-latitude ~40°N)) show a range of 17.3-23.2°C. These paleo-temperatures are in good agreement with previous sea surface temperatures for the mid-Eocene mid latitude regions based on foraminifera clumped isotopes, adding confidence to both datasets. In addition, in order to aid the accurate reconstruction of seawater ẟ18O values, we compiled published oxygen isotope fractionation data for molluscs and established relationships that describe the temperature dependence of oxygen isotope fractionation between water and molluscan calcite and aragonite, respectively. Applying the equation for aragonite to the Eocene samples, we obtain reconstructed seawater δ18O for the Hampshire Basin between -2.3‰ and - 3.5‰ VSMOW.