RECONSTRUCTING SEASONALITY USING OXYGEN ISOTOPES FROM BIVALVE MOLLUSK SHELLS: GEOARCHAEOLOGICAL IMPLICATIONS FOR ENVIRONMENTAL INTERPRETATION

Past human patterns of site occupation and migration were linked, in part, to annual climate variability. Annual temperature variability (herein referred to a seasonality), is recorded by stable oxygen isotope variability in bivalve mollusk shell carbonate. Because clams were commonly exploited for food, their discarded shells are common in many coastal archeological sites, providing useful clues for understanding the link between climate and past human movements.

Interpretation of sclerochronological records, however, is complicated by variable growth cessations and/or variable growth rates. To robustly quantify the impact of growth rate variation on the resolution and fidelity of bivalve mollusk shell archives, we used a Monte Carlo simulation to model stable oxygen isotope profiles from synthetic mollusks. Modeled shell oxygen isotope (δ¹⁸O) values were constructed using known patterns of water temperature and water-δ¹⁸O variation and hourly growth rates were calculated using simulated optimal growth temperatures (OGT). Together, these variables were used to construct complete ontogenetic δ¹⁸O profiles, which were subsequently virtually subsampled to reflect actual sampled profiles. This procedure was repeated using stochastically generated OGTs to simulate actual population variability. In addition, we conducted a spatial sampling analysis on a synthetic mollusk to identify sampling biases when comparing samples collected from different regions of the shell. This simulation allowed us to illustrate complications that may arise when comparing samples collected from regions of a shell with different spatiotemporal resolution. Models were compared with observed biogeochemical samples collected from an ontogenetic series grown under known environmental conditions.

Our models highlight the importance of understanding how organismal biology impacts reconstructed environmental histories. Finally, we make specific suggestions for sampling strategies given constraints on the resolution of environmental reconstructions.