Amino acid racemization kinetic modeling and the late Quaternary paleoclimate of the North Carolina coastal plain

Amino acids extracted from Quaternary mollusks represent a preserved residue of the original biomineralization protein. The extent of racemization of these fossilized amino acids is often used as a stratigraphic or chronologic tool. With suitable calibration, racemization data can also be used to construct a history of temperature for the samples in question. Understanding diagenetic amino acid racemization (AAR) kinetics requires a combination of laboratory kinetic experimentation and testing of kinetic models with natural field samples that have independent chronostratigraphic control. These models must also incorporate information on the temperature dependence ("apparent" or "bulk" activation energy) as derived from either elevated temperature laboratory experiments or data from field samples with known ages and diagenetic temperatures. Over the past two decades, several well-calibrated datasets and a variety of kinetic models have become available: foraminifera results provide calibration for low temperature kinetics (e.g., Muller, 1984) and paired radiocarbon-AAR results provide calibration at higher surface temperatures (e.g., Miller et al., 1997). The most recent approaches to modeling of AAR kinetics include a molecular approach (Collins and Riley, 2000) that links racemization to polypeptide breakdown reactions and a transformation approach (Manley et al., 2000) that builds upon earlier "parabolic" (Mitterer and Kriausakul, 1989) or "non-linear" models (Wehmiller et al., 1988). The parabolic model, while useful over limited time or D/L ranges, has the disadvantage that it does not predict correct equilibrium D/L values. The other three models, if calibrated to a common value, predict consistent patterns of racemization that are in good agreement, and the molecular approach has been used to extrapolate foraminifera kinetics (Muller, 1984) to estimate ages and diagenetic temperatures of Pleistocene mollusks from North Carolina, predicting a late Pleistocene cooling of ca. 7 deg. C. for the region, based on a simple square-wave model of temperature change. AAR results for seven amino acids in a suite of radiocarbon-dated Mercenaria samples from the North Carolina coastal plain expand this approach to paleotemperature estimation. Samples have been obtained from beaches, the inner shelf, and from subsurface drilling. 14C ages range from ~1 kyr to >52 kyr, with the older (radiocarbon dead) likely representing "last interglacial" (80-130 kyr) ages. Additionally, the observed AAR kinetics for the Holocene samples from different latitudes (temperatures) allow estimates of "effective activation energies" that compare well with those determined by experimental heating. These effective activation energies can then be used in combination with numerical models of racemization to predict D/L values as a function of time and optional temperature histories for the region. Full glacial temperature reductions predicted in this manner are model-dependant and range from 6 to 11 deg. C.