Abstract

The isotopic composition of organic carbon buried in marine sediments is an appealing proxy for palaeo CO₂ concentrations due to the well-documented effect of CO₂ concentrations on carbon fractionation by phytoplankton. However, a number of factors, in addition to CO₂ concentrations, influence this fractionation. Included among these factors are cell geometry, in particular surface/volume ratios, growth rate, and the presence of CO₂ concentrating mechanisms. Other potentially confounding factors are calcification, diagenesis, and the nature of the growth-rate-limiting factor, e.g. light vs nutrients. Because of these confounding factors, palaeoreconstructions based on the isotopic composition of organic carbon (δ¹³C) will almost certainly have to be based on the isotopic signatures of organic compounds that can be associated with a single species, or group of physiologically similar species. Long-chain alkenones produced by certain species of coccolithophores may provide a suitable diagnostic marker. By combining the δ¹³C of the alkenone carbon with the δ¹³C of coccolith carbon and the Sr/Ca ratio of the coccoliths, it is possible to calculate the extent of carbon fractionation (ε_p) and estimate growth rates. However, active transport of inorganic carbon tends to make ε_p insensitive to CO₂ concentrations when the ratio of growth rate to CO₂ concentration exceeds 0.285/rkg mol^–1d^–1, where r is the effective spherical radius of the cell in microns. Palaeo CO₂ concentrations calculated from alkenone and coccolith δ¹³C data capture the gross features of CO₂ concentrations in the Vostok ice core, but explain only 30–35% of the variance in the latter. The absence of a higher correlation may in part reflect the impact of active transport, particularly during glacial times. The impact of active transport may have been less severe prior to the Pleistocene, since CO₂ concentrations are believed to have been higher than present-day values during most of Phanerozoic time.