The future of the past: A robust framework for the upgrade and development of the international radiocarbon calibration/comparison curves

Many researchers in archaeology and the geosciences obtain timescales for their projects by radiocarbon dating plant or animal remains from the preserved deposits with which they work. Radiocarbon dates are not the same as calendar dates, however, and have to be corrected for variations in the radiocarbon content of the atmosphere at the time that the plant or animal lived. This conversion of radiocarbon dates to calendar ages, known as calibration, is not a straightforward correction. Calibration of radiocarbon dates can only be done by comparison to a suitable calibration curve. Such curves are based on measurements of radiocarbon in samples of known calendar age such as tree-rings, or in a less strict sense, on other types of samples where an independent method of dating can be used. For samples which grew in the ocean, such as shells and corals, a separate calibration curve is needed to account for changes in ocean water circulation which may bring up 'old' water from the ocean depths (the reservoir effect). The calibration curves have been refined periodically to provide better estimates of the calendar ages. In 2004, the IntCal Working Group constructed new calibration curves from radiocarbon dated tree-rings back to 12,400 years before present and from independently dated ocean samples, using an estimated correction for the reservoir effect, back to 26,000 years before present. Rather than simply averaging the data, these curves were constructed with statistical tools (models) that allowed for the uncertainty in the calendar ages of the samples used as well as the radiocarbon dates. At that time data beyond 26,000 years before present did not agree so no curve was provided but an estimate of how far the data sets differed from the underlying true curve was given. In the last few years a lot of research has gone into producing radiocarbon datasets from a variety of records. Many of these datasets are now in fairly good agreement so it should be possible to provide curves for estimating calendar ages back to 55,000 years before present. In addition new tree-ring records are becoming available which will improve the precision of the calibration curve. Statistical methods have also been rapidly advancing and so some of the simplifying assumptions that we made about the models in 2004 will no longer be necessary. Working in collaboration with the IntCal Working Group, this project will develop an easily maintainable database of calibration quality radiocarbon data to be used to produce updates to the calibration curves on a regular basis. Advances in statistics will allow us to improve on the previous models to further refine the calibration curves. Measurements of carefully selected coral will help determine what corrections are needed for ocean samples to be used in calibration curves. By improving radiocarbon calibration this project will improve the understanding of the sequence and timing of events in numerous studies in archaeology and in the reconstruction of past environments.