The Mg isotopic composition of marine pore fluids from ODP Site 807A (Ontong Java Plateau): Implications for the Cenozoic Mg chemistry of the ocean

The Mg isotopic compositions (δ²⁶Mg, relative to DSM3) of pore fluids from ODP Leg 130, Site 807A, as measured by MC-ICP-MS, are found to be between -0.25 and -0.79% (ave: -0.55%). The shallowest pore fluid measured was sampled at 13.4 mbsf while the deepest was sampled at 738 mbsf (a time span of ~35 Ma); the total section thickness is ~1350 meters. The sedimentary section is carbonate-rich, with a hole average CaCO3 content of ~92 wt%, and overlies basaltic basement. There is a systematic decrease in δ²⁶Mg from the top of the section (~ -0.8%) to deeper in the section (~ -0.3%) that coincides with a decrease in pore fluid Mg and an increase in pore fluid Ca concentrations. The δ²⁶Mg values at the upper and lower boundaries of the column are consistent with the isotopic composition of laboratory seawater (-0.84 ± 0.10%) and Kilbourne Hole olivine (-0.26 ± 0.05%) standards, respectively, the latter of which is isotopically identical to fresh basalt. The most basic interpretation of the data is that the pore fluids at 807A reflect modern seawater δ²⁶Mg at the top of the section and basement basaltic Mg at ~800 mbsf. Because the section contains ~550 meters of carbonate below the deepest pore fluid measured, this implies a strong basement influence on Mg isotopic composition well above the sediment-basement interface. That the bulk carbonate and pore fluid are out of equilibrium is an expected result, based on the large reactive length scale at 807A (~1500 m) [1-2]. Consequently, because the previously measured δ²⁶Mg values of bulk carbonates at 807A are ~ -4 to -5%, there is long-term leverage to change the isotopic composition of carbonates over time scales of tens of millions of years if we assume that the fractionation factor during diagenesis is different from ~0.9963 (the value assumed for biogenic carbonate formed in the surface ocean). Unlike Ca isotopes, the pore fluid is not well buffered with respect to Mg over geological time scales. Therefore, if carbonates exchange Mg with pore fluids over long time scales, as has been suggested previously [1-2], then the use of Mg isotopes as a paleo-proxy in marine carbonates is complicated. In addition, there is considerable structure in the pore fluid δ²⁶Mg values as a function of depth. Neither the pore fluid isotopic data nor the Mg concentration data fit a simple concave up diffusion profile with depth. Therefore, we hypothesize that there is a component of the pore fluid Mg isotope geochemistry that reflects variations in the Mg isotopic composition of seawater over time, similar to previous attempts at reconstructing seawater Mg concentrations [1]. We use numerical models to simulate deposition, recrystallization, and diffusion over million-year time scales in order to determine seawater δ²⁶Mg over time, constraining recrystallization rates based on previous work [1-2]. References [1] Fantle and DePaolo (2006) GCA, 70, 3883-3904 [2] Fantle and DePaolo (2007), GCA, 71, 2524-2546 [3] Higgins and Schrag (2010), GCA, 74, 5039-5053