Perspectives on geochemical proxies: The impact of model and parameter selection on the quantification of carbonate recrystallization rates

Diagenetic reactions in marine sediments, such as the recrystallization of carbonates, can impact the accuracy of paleo-environmental and paleo-climatic reconstructions by geochemical proxies. The extent to which the recrystallization of carbonates affects the chemistry of sedimentary archives depends on the reaction rate, extent of isotopic disequilibrium, and duration of reaction. The reaction rate, which is obviously critical, can be constrained by the elemental and isotopic compositions of pore fluids. Such constraints are affected by assumptions regarding the temperature in the sedimentary column relative to the temperature of formation, the burial rate, pore fluid advection, the composition of the sediments (carbonate-rich versus siliciclastic), and the porosity of the sediment column. In this study, we use a steady-state analytical solution to the diagenetic equations to constrain depth-dependent reaction rates (and extents of recrystallization) based on the Ca isotopic compositions of pore fluids in sedimentary columns at multiple ocean drilling sites (Sites 807, 984, 1170, and 1171), which encompass a diverse range of sedimentary compositions and conditions. We find that carbonates in siliciclastic sediments are generally less altered by diagenesis than their carbonate-rich counterparts. The discrepancy in recrystallization rates between siliciclastic and carbonate-rich sedimentary sections is, however, significantly smaller than previously estimated, suggesting that siliciclastic archives are not immune to diagenetic effects. While we find that diagenesis can decouple contemporaneous proxies of sea surface temperature (Mg/Ca and δ18O), our calculations also reveal that δ18O-based temperature estimates are more robust in siliciclastic sections relative to carbonate-rich sections. Sensitivity tests of the calculated extent of recrystallization suggest that uncertainties in porosity and burial rate are generally the greatest sources of error to proxy reconstruction from diagenetically altered sediments. The conclusions drawn using the analytical solution are benchmarked against a depth-dependent, forward numerical model using the CrunchFlow software (Steefel et al., 2015); ultimately, this comparison demonstrates that the assumptions necessary in deriving the analytical solutions have a relatively minor impact on the resulting conclusions.