Core-top constraints on the ecology and paleothermometry of planktic foraminifera in the Indian Ocean

Abstract

The magnesium-to‑calcium ratio (Mg/Ca) of fossil foraminifera is a widely used geochemical proxy for reconstructing past ocean temperatures. Culturing experiments have shown that, for some species, nonthermal influences like salinity and carbonate chemistry play a non-trivial role in the incorporation of Mg into foraminiferal calcite. Inclusion of these variables into Mg/Ca calibrations have provided a better fit to core top and sediment trap data in some recent studies, but the widespread covariation of potential control variables in the open ocean and the lack of multivariate calibrations for some species warrants additional work. Here, we assess Mg/Ca calibrations for four common species of planktic foraminifera from the Indian Ocean: Globigerinoides ruber (white; sensu lato), Globigerina bulloides, Globorotalia inflata, and Globorotalia truncatulinoides (sinistral). Using a newly assembled database of paired Mg/Ca-δ¹⁸O observations from 115 Indian Ocean core tops, including 25 new observations from a 2018 cruise (CROCCA-2S), we build a multi-parameter environmental dataset informed by local constraints on each species' apparent calcification depth and growth season. We combine multivariate regression techniques with Bayesian factor analysis to develop several multivariate models and assess whether they improve model fit and performance over temperature-only equations. For G. ruber, the temperature-only model is the best empirical fit to our data (Mg/Ca = exp.(0.086*(T – 25)) + 1.49). Although we find a statistically significant relationship to seawater pH, its impact on estimated temperatures (± 0.65 °C per 0.1 pH unit) is smaller than the total uncertainty on the temperature-only model (±1.2 °C; 95 % CI). The Mg/Ca of G. bulloides is most reasonably described by a temperature-only model (Mg/Ca = exp.(0.068*(T – 25)) + 1.51). An apparent negative correlation between G. bulloides Mg/Ca and salinity is not consistent with prior culture and field studies, which have yielded only positive correlations (if any). This anomalous result is likely a byproduct of including core tops from the monsoon-sensitive Java/Sumatran coast. Finally, the Mg/Ca of subsurface-dwelling G. inflata and G. truncatulinoides fell within predicted ranges of some published calibrations, suggesting they might be effectively applied in this region. However, data for these two species were only available for cores from the southeast Indian Ocean where deep ocean mixing limits the dynamic range temperature, salinity, and other parameters across their habitat depth, preventing the creation of new, regionally robust calibrations. Overall, our work offers new calibrations for surface-dwelling species (G. ruber and G. bulloides) in the Indian Ocean and provides a framework for further refining multivariate controls on Mg/Ca of subsurface-dwelling species in the region.