What do oyster shells tell us about paleoclimatology?

In the context of the current global warming, it is crucial to acquire new paleoclimatic data around past analogues. For decades, bivalve shells have constituted one of the most common supports for paleoclimatic archives based on stable isotope approaches to constrain the seasonal gradient of temperature. One of the key points to conduct this kind of project reliably is to constrain precisely what the analyzed organisms do record and what it is possible to restore from their analysis? Thus, the prerequisite for any (paleo)climatic study is to determine the rhythms that govern the biomineralization of the analyzed organisms. Additionally, it is essential to verify the significance of their geochemical signal.

Thus, in this work, we present results combining a sclerochronological and geochemical approach applied to oyster shells. The shells were cultivated in sites from the French Atlantic coast where the environmental conditions (temperature, salinity, ...) were measured continuously during two years. The shells were chemically labeled (Mn²⁺) in order to obtain a precise time frame within each shell that can be revealed under cathodoluiminescence.

The results of the sclerochronological study indicate that oysters mineralize their shells with a rhythmicity following tidal cycles (2 calcite increments / day). A lunar and seasonal cyclicity is also highlighted. High resolution geochemical measurements (d¹⁸O) confirm that oysters can reliably reproduce seasonal temperature variations, even if geochemical drifts are observed during the first year of growth due to a very high growth rate. We also tested the new ∆₄₇thermometer for these mollusks. Our results are consistent with a strong correlation between seawater temperature and ∆₄₇ for all samples except for the juvenile samples that also exhibited an isotopic disequilibrium for the d¹⁸O. This correlation is consistent with earlier calibrations reprocessed in the recent I-CDES metrological scale, adding to the evidence that many different types of carbonates conform to statistically indistinguishable relationships between ∆₄₇ and crystallization temperature.

Finally, these results are applied to reconstruct the paleoclimatic evolution around a hyperthermal event that occurred 40 Ma ago, the Middle Eocene Climatic optimum.