Low temperature stable mineral recrystallization of foraminiferal tests and implications for the fidelity of geochemical proxies

摘要

Fossilized foraminiferal tests are widely used as proxies to constrain paleoenvironments, including past sea surface and bottom water temperatures and the chemical composition of ancient seawater. However, relatively little is known about the extent to which calcareous foraminiferal tests exchange with seawater prior to burial. In this study, we used a Ca-45 radiotracer approach to quantify the extent and rate of foraminiferal calcite-fluid exchange over a three-month period. Modern foraminiferal tests exhibited faster exchange with a Ca-45-spiked solution (2.6.10(-4) to 2.0.10(-5) mol/m(2)/d) than fossil foraminiferal tests (7.6.10(-5) to 5.3.10(-6) mol/m(2)/d) from ODP Site 807. The impact of dissolved Si on the exchange rate of modern foraminifera was minimal (2.6.10(-4) to 9.2.10(-6) mol/m(2)/d). Unlike foraminiferal calcite, inorganic calcite exchanged at remarkably slow rates (6.6.10(-6) to 5.9.10(-7) mol/m(2)/d) suggesting that biogenic calcite is more reactive than inorganic calcite. Increased Ca2+, Mg2+, and Sr2+ concentrations in fluids from parallel tracer-free reactors indicated preferential dissolution of tests, though there was no overt change in test morphology during experiments (as imaged by scanning electron microscopy and micro-CT scanning). Time-dependent box model simulations of Ca-45, as well as aqueous elemental chemistry, support the hypothesis that calcite-fluid exchange occurred via dissolution-reprecipitation of foraminiferal tests that was induced by intra-test chemical heterogeneity. The short-term exchange reactions investigated in this study are comparable to the reactions that can occur in the seawater column and at the sediment-seawater interface. This work suggests that dissolution-reprecipitation in such settings has the potential to overprint the chemistry of foraminiferal tests without overt physical alteration, which can lead to a potential bias in paleoenvironmental reconstructions. Model-based estimates suggest a maximum of 2-