Exploring the controls on element ratios in middle Eocene samples of the benthic foraminifera Oridorsalis umbonatus
Culturing studies and empirically based core top calibrations have been used to infer that elemental ratios in benthic foraminifera can be used as proxies to reconstruct past variations in bottom water temperature and saturation state (Δ [CO 32−]). However the mechanisms linking elemental ratios to these parameters are poorly constrained. Here, we explore the environmental parameters influencing the incorporation of B, Li, Sr and Mg in Oridorsalis umbonatus in early Cenozoic sediments from Ocean Drilling Program Site 1209. We investigate the influence of middle Eocene variations in intermediate water Δ [CO 32−] using relationships developed from core top samples. The fidelity of bottom water Δ[CO 32−] reconstructions based on single element ratios is assessed by comparing the X/Ca-based reconstructions to each other and to carbon cycle proxy records (benthic foraminifera δ 13C, organic carbon content, foraminifera dissolution indices), and a seawater δ 18O reconstruction for Site 1209. Discrepancies in the reconstructed Δ[CO 32−] values based on these different metal ratios suggest that there are still gaps in our understanding of the parameters influencing X/Ca and demonstrate that caution is required when interpreting palaeo-reconstructions that are derived from a single elemental ratio. The downcore record of O. umbonatus Mg/Ca does not exhibit any similarities with the Li/Ca, B/Ca and Sr/Ca records, suggesting that the environmental parameters influencing Mg/Ca may be different for this species, consistent with temperature as the strongest control on this elemental ratio. This hypothesis is supported by the coefficients of multiple linear regression models on published Mg/Ca data. An incomplete understanding of the controls on elemental incorporation into benthic foraminifera hinders our ability to confidently quantify changes in saturation state using single X/Ca reconstructions over a range of timescales.