HIGH-RESOLUTION, TWO-DIMENSIONAL, NUMERICAL MODELS OF BIVALVE MOLLUSK SHELLS: BULK SAMPLING, DISEQUILIBRIUM FRACTIONATION, AND FALSE POSITIVES

Stable oxygen isotope profiles from bivalve shell carbonate (δ¹⁸O_carb) have emerged as important archives of past environmental conditions because they record water temperature and isotope variation (δ¹⁸O_water). Their utility is predicated on the assumption that bivalves precipitate oxygen isotopes in thermodynamic equilibrium with their environment. Several recent studies, however, have questioned the assumption of equilibrium fractionation in some species. Claims of disequilibrium fraction must be supported by tightly constrained environmental data (i.e., water temperature and δ¹⁸O_water), and carefully designed shell sampling strategies. The latter is particularly important when asserting disequilibrium fractionation in different regions of a single shell. In other words, when comparing intra-shell δ¹⁸O_carb values it is essential that samples be contemporaneous and time-averaged to the same degree. Here we employ a high-resolution, two-dimensional, numerical model of a generic venerid shell cross-section to evaluate δ¹⁸O_carb variability of bulk samples collected from different layers of the shell. The model assumes equilibrium fractionation across the entire shell. δ¹⁸O_carb values were calculated from modeled environmental conditions for hourly growth increments “deposited” over a ten-year lifespan. Hourly growth rates were calculated assuming optimal growth temperatures. Ontogenetically declining growth was modeled by incorporating reduced annual growth rates, progressively longer growth cessations, and seasonal changes in food availability. Finally, hourly growth increments, each with a known δ¹⁸O_carb value, were then fit to a logarithmic spiral assuming isometric growth. This procedure generates a 2D shell model with known δ¹⁸O_carb values across the entire cross-section. Bulk samples were then collected from different regions of the shell separated by the pallial line (i.e., inner and outer shell layers). Results suggest that contemporaneous samples, with similar degrees of time-averaging, yield nearly identical δ¹⁸O_carb values. However, non-contemporaneous samples can have δ¹⁸O_carb values that differ by >0.5‰, well outside IRMS error. These data suggest bulk samples from a single shell can yield oxygen isotope values that can easily be mistaken for disequilibrium fractionation.