Experimental calibration of manganese incorporation in foraminiferal calcite

Abstract

In the context of recent climate change and increased anthropogenic activities in coastal areas, which both may have a negative impact on dissolved oxygen concentration, there is an increased interest to better understand the mechanisms and evolution leading to hypoxia in marine environments. The development of well calibrated proxies is crucial to obtain reliable environmental reconstructions of past oxygen content and of historical development of hypoxia. Manganese is a redox element of interest for this purpose because manganese oxides are reduced to soluble Mn2+ in oxygen-depleted conditions, which can be incorporated in biogenic calcite. The Mn/Ca ratio in benthic foraminiferal calcite is therefore a promising proxy to reconstruct past oxygen variations. In this study, we calibrate this proxy by measuring (with Laser Ablation ICP-MS) the Mn/Ca ratio of benthic foraminifera calcified under controlled conditions in laboratory experiments. Two benthic foraminiferal species (Ammonia tepida and Bulimina marginata) calcified in 4 different dissolved manganese concentrations (from 2.4 to 595 µmol L−1) corresponding to in situ Mn concentrations encountered in bottom and/or pore waters in low oxygen marine environments. There is a statistically significant positive linear correlation (R2 > 0.9) between Mn/Cacalcite and Mn/Caseawater. However, the two species show different partitioning coefficients (DMn of 0.086 and 0.621, for A. tepida and B. marginata, respectively), although they calcified in exactly the same stable conditions. These results highlight a strong species specific effect on Mn incorporation, which is probably due to different biological controls during biomineralisation processes. There is also ontogenetic variability (determined through a comparison of successive chambers) that is different between the two species and also varies as a function of the dissolved Mn concentration. A conceptual model is proposed to explain these data.