Carbon and oxygen isotope variability among foraminifera and ostracod carbonated shells

Abstract

This study investigates the effect of biological and environmental inter-individual variability on the meaning of d 18 O and d 13 C values acquired on small carbonated shells. First we present data obtained with a MultiPrep automated carbonate system on small sample sizes of a homogeneous carbonate material: Carrara marble. This demonstrates the capacities of the analytical system to reliably run small amounts of carbonates even down to 10 mg. Then we present two data sets obtained on real fossil samples of various size ( sensu number of individual organisms) calibrated against the NBS19 carbonate standard. Both datasets evidence a clear trend of between-biological sample standard deviation increase for both d 18 O and d 13 C measurements when the number of pooled specimens per sample decreases. According to the results obtained from a systematic study of a geologically homogeneous sample of coeval fossil Elphidium foraminifera, we estimate that there is 95% of chances to reach between-biological sample standard deviation values higher than 1.02‰ (d 18 O) and 1.45‰ d 13 C) based on single-cell measurements. Such values are one order of magnitude higher than the instrumental standard deviations associated with these stable isotope ratios. Conversely, a minimum of 35 (d 18 O) and 44 (d 13 C) pooled specimens of Elphidium appears necessary to reach a between-sample standard deviation £ 0.25‰ with a probability of 95%. Such biological intrinsic and irreducible variability between coeval individuals, and thus samples, clearly questions the interest for single-cell analyses, more precisely, for coastal marine species, such as Elphidium, subject to many environmental changes during their life-time. Indeed, strong variations in salinity or temperature, as well as biogenic fractionation, could influence the isotopic composition of an individual specimen. Results might be less problematic for an average community including several tests. This paper underlines uncertainties linked to specific environments in which selected organisms live, especially for paleoceanographic or paleoclimatic reconstruction purposes where secular oxygen and carbon isotope variations typically range from 0.5 to 1.5‰.