Temperature dependence of oxygen isotope fractionation in coccolith calcite: A culture and core top calibration of the genus Calcidiscus

Reconstructions of seawater temperature based on measurement of oxygen isotopes in carbonates mostly derive from analyses of bulk sediment samples or manually picked foraminifera. The temperature dependence of O-18 fractionation in biogenic calcite was first established in the 1950s and the objective of the present study is to re-evaluate this temperature dependence in coccolith calcite with a view to developing a robust proxy for reconstructing "vital effect"-free delta O-18 values. Coccoliths, the micron-sized calcite scales produced by haptophyte algae that inhabit surface mixed-layer waters, are a dominant component of pelagic sediments. Despite their small size, recent methodological developments allow species-specific separation (and thus isotopic analysis) of coccoliths from bulk sediments. This is especially the case for Calcidiscus spp. coccoliths that are relatively easy to separate out from other sedimentary carbonate grains including other coccolith taxa. Three strains of coccolithophores belonging to the genus Calcidiscus and characterised by distinct cell and coccolith diameters were grown in the laboratory under controlled temperature conditions over a range from 15 to 26 degrees C. The linear relationship that relates 18O fractionation to the temperature of calcification is here calibrated by the equation: T [degrees C] = -5.83 x (delta O-18(Calcidiscus) - delta O-18(medium)) + 4.83 (r = 0.98). The slope of the regression is offset of similar to-1.1 parts per thousand from that of equilibrium calcite. This offset corresponds to the physiologically induced isotopic effect or "vital effect". The direction of fractionation towards light isotopic values is coherent with previous reports, but the intensity of fractionation in our dilute batch cultures was significantly closer to equilibrium compared to previously reported offset values. No significant isotopic difference was found between the three Calcidiscus coccolithophores, ruling out a control of the cell geometry on oxygen isotope fractionation within species of this genus. This also indicates that our culture calibration may be applied to all Calcidiscus coccoliths found in the sediment. We compared the culture calibration to delta O-18 measured from near-monogeneric Calcidiscus fractions separated out from core top sediments. We found concordant O-18 fractionation factors for the core top calibration with a good linear coefficient (r = 0.94). The near-monogeneric Calcidiscus assemblages seem, however, to record slightly heaviest delta O-18 values compared to the data of culture study. This discrepancy may be due to a possible seasonality effect on the production of Calcidiscus coccoliths. The uncertainty of the calibration is of similar magnitude to those of other proxies used for SST reconstruction, such as foraminiferal delta O-18 or the alkenone under-saturation index. This confirms that coccoliths can be used as a complementary or alternative substrate to foraminiferal shells for isotopic analyses. Comparing delta O-18 of coccoliths to these other SST proxies, or developing an interspecific comparison of coccolith geochemistry may give insights into the carbonate chemistry of seawater through key periods of the geological record. (C) 2012 Elsevier Ltd. All rights reserved.