Reconstructing Eocene atmospheric CO2 using the boron isotopic composition of larger benthic foraminifera

The Cenozoic Era, commencing ∼66 million years ago and extending to the present day, has witnessed substantial variations in Earth’s climate. Initially, the climate was significantly warmer than contemporary conditions, exhibiting a more uniform global temperature with minimal latitudinal gradients. During the early Cenozoic, tropical conditions were extending considerably beyond their current geographical limits, with luxuriant forests at the poles and an absence of permanent ice, and sea levels reaching more than 50 m higher than today. This warm interval, often designated as a “hothouse” world, was distinguished by elevated atmospheric carbon dioxide concentrations, potentially reaching ∼ 1,500 ppm. Understanding the link between this climate state and the concentration of key greenhouse gases over geological time scales is a crucial endeavour to better predict (and prepare for) anthropogenic climate change. To do so, the robust and proven proxy of boron isotopes in marine carbonates will be used in a new proxy archive and measured using a novel sample introduction technique. Traditionally, the marine carbonate boron isotope composition is measured by analysing the extracted boron in solution from the carbonate matrix. Although more precise, this methodology requires extensive cleaning and dissolving procedures which are prone to contamination. More recently, the development of boron isotope measurements using highly spatially-resolved analytical sampling techniques, such as laser ablation, has gained traction to assess sample heterogeneity and by-pass the need the of matrix separation. However LA-MC-ICPMS suffer from an unusual matrix interference, biasing the boron isotopic measurement in carbonate samples significantly. This thesis refines correction approaches to produce precise and accurate boron isotope measurements in carbonate samples, as well as exploring the sources of the interference. These methodological advances in LA-MC-ICPMS allowed for unprecedented sampling resolution and measurement precision, which is why this methodology was chosen to analyse a limited foraminifera sample set of modern Operculina ammonoides from empirical culture experiments which were isotopically labelled, and in which pH and DIC were decoupled from each other. Boron isotope measurements of these specimens demonstrated that the cultured sample set appears to record a δ11 B significantly higher than δ11 BB(OH)− , especially at lower pH. In contrast, wild specimens had a sensitivity to pH close to that of borate ion. We suggest that the low sensitivity of the laboratory culture sample set is the result of the difference between their micro-environment carbonate chemistry and that of the surrounding seawater, itself driven by a pH dependence of their relative rates of calcification and photosynthesis. As such, we demonstrate the utility of these organisms as a proxy archive while also highlighting how laboratory experimental design has the potential to drive important changes in the micro-environment and resulting shell chemistry of organisms of this size. Given that the genus Operculina originated in the late Palaeocene, the findings of this chapter paves the way towards deep-time palaeo-pH/CO2 reconstructions using foraminifer species which have a very closely related modern representative. Building on the analytical developments and the modern empirical calibration of O. ammonoides, a set of Eocene fossil LBF (Nummulites sp.) from the Paris and Hampshire basins was analysed to reconstruct the early-Eocene palaeoclimate. The trace element, δ11 B, and ∆47 compositions of LBF tests were characterised to assess the preservation state of the fossils collected from several expeditions and museum collections, finding that most fossils were well-preserved. The use of LA-MC-ICPMS to quantify the δ11B of fossil Nummulites allowed, for the first time, the characterisation of heterogeneity within a single specimen, finding that a single LBF can have significant δ11B variability. The measured δ11 B of several individual specimens per sample, served to reconstruct early-mid Eocene CO2, with an average value of approximately 1500 ppm, resulting in an estimated Earth climate sensitivity of 5.56.3 4.8. These results are consistent with previous reconstructions of temperature, CO2 , and Earth’s climate sensitivity, demonstrating the utility of LBF as a robust archive for reconstructing past seawater pH and CO2 levels, complementing other marine carbonate records and archives beyond ice cores. It also demonstrates the suitability of LA-MC-ICPMS as a precise and reliable methodology for isotopic palaeoreconstructions.