Quantifying the Number of 14C Determinations Required to Improve Dating Accuracy for Lapita Deposits

Summary

The use of radiocarbon dating to calculate the dates of Lapita deposits remains largely a single-step, ad hoc procedure. The accuracy of dating results can be greatly improved through Bayesian modeling. However, this depends on the number and stratigraphic distribution of radiocarbon determinations and the shape of the calibration curve. To evaluate these issues, we used Oxcal 4.2 to simulate, through the process of back-calibration, radiocarbon determinations that we could expect to receive as measurements on a series of samples of "known" Lapita dates in Remote Oceania. We then ran a series of Bayesian models in which the number of radiocarbon determinations was incrementally increased, stratigraphy altered, and temporal parameters varied. Because the radiocarbon determinations are based on "known" calendar dates we were able to evaluate when a particular Bayesian model provided results that were accurate and precise and at what point the addition of further age determinations did not affect the results. Our analyses demonstrate that dating of Lapita deposits must be approached as an iterative process, requiring a sampling strategy that allows for an assessment of redundancy in results that is indicative of a high degree of accuracy in the calibrated dates.

This Resource is Part of the Following Collections

• Society for American Archaeology 82nd Annual Meeting, Vancouver, BC (2017)   •
• Sailing at the Edge of Time: Global Perspectives on Island Colonization

Keywords

Spatial Coverage

Record Identifiers

Abstract Id(s): 15837