Refining ancient carbon dioxide estimates: Significance of coccolithophore cell size for alkenone‐based p CO 2 records

Long‐term alkenone‐based p CO 2 records are widely applied in paleoclimate evaluations. These p CO 2 estimates are based on records of the carbon isotope fractionation that occurs during marine haptophyte photosynthesis (ɛ p 37:2 ). In addition to the concentration of aqueous CO 2 (CO 2(aq) ) the magnitude of ɛ p 37:2 is also influenced by algal growth rates and cell geometry. To date, the influence of haptophyte cell geometry on the expression of ancient ɛ p 37:2 values has received little attention. This study evaluates changes in cell geometry of ancient alkenone‐producing algae at Deep See Drilling Project Site 516 in the southwest Atlantic Ocean by analyzing individual coccolith dimensions, which are proportional to algal cell volume and surface area. We show that during part of the early Miocene, mean cell sizes of alkenone‐producing algae were smaller relative to modern Emiliania huxleyi . Cell size variations coincide with significant changes in ɛ p 37:2 , with a distinct 6‰ decrease in ɛ p 37:2 at ∼20.3 Ma associated with a 27% increase in haptophyte cell sizes. These changes in cell size impact ɛ p 37:2 ‐based interpretations of growth rate variation and CO 2(aq) estimates for this southwest Atlantic site. After correcting for cell geometry, CO 2(aq) estimates at Site 516 are consistent with those reported from other oligotrophic sites during this time, resulting in overall low atmospheric p CO 2 estimates (<350 ppmv) for the early Miocene.