Prof Christopher Ramsey

Abstract:

The Bayesian statistical method of the OxCal v 4.1 program is used to construct an age-depth model for a set of accelerator mass spectrometry (AMS) radiocarbon ages of organic debris collected from a late Holocene Dead Sea stratigraphic section (the Ein Feshkha Nature Reserve). The model is tested for a case where no prior earthquake information is applied and for a case where there is incorporation of known ages of 4 prominent historical earthquakes as chronological anchor points along the section. While the anchor-based model provided a tightly constrained age-depth regression, the "nonanchored" model still produces a correlation where most of the 68% or 95% age ranges of the 52 seismites can be correlated to historical earthquakes. This presents us with the opportunity for high-resolution paleoseismic analysis and comparison between various sites. © 2010 by the Arizona Board of Regents on behalf of the University of Arizona.

Abstract:

Under certain environmental conditions, post-depositional diagenetic loss of bone collagen can severely reduce the number of bones from a particular archaeological site that are suitable for stable isotopic analysis or radiocarbon dating. This study examined nearly 300 bones from 12 archaeological sites across southern England known to yield poor or variable preservation to try to identify one, or more, pre-screening technique(s) that would indicate suitable collagen preservation for radiocarbon dating. The most reliable method was shown to be the percent nitrogen (%N) of whole bone powder, which has an 84% chance of successfully predicting whether or not a bone will yield sufficient (i.e. >1% weight) collagen for dating. © 2009 Elsevier Ltd.

Abstract:

During the laboratory pretreatment of samples for radiocarbon dating, small amounts of carbon may be added to a sample. Contamination can be incorporated at any stage: during chemical pretreatment, combustion to CO2, graphitization, or accelerator mass spectrometry (AMS) measurement. Such carbon contamination is often modern in age, and so can have an especially severe effect on samples older than ~25 ka BP. During the extraction of collagen from bone using the ultrafiltration protocol at the Oxford Radiocarbon Accelerator Unit (ORAU), small amounts of young carbon are added to the sample. Currently, this contamination is poorly characterized when less than 10 mg of collagen is extracted from a bone. Demand to date small collagen samples with 14C concentrations that approach the detection limit of AMS measurement has increased recently with the growing interest in, for example, directly dating Neanderthal remains and Upper Paleolithic bone artifacts. This paper aims to reduce the minimum collagen sample size required to produce a reliable date from 10 to 5 mg by re-examining the combustion background and subsequently the pretreatment background for bone. The average of 136 measurements of directly combusted nylon suggests that 0.0007 ± 0.001 mg of modern carbon is added to each sample, although the distribution is positively skewed. Regression analysis of the measurements of 52 collagen samples extracted from a bone of background age results in a background of just less than 50,000 BP for bone treated at ORAU. © 2010 by the Arizona Board of Regents on behalf of the University of Arizona.

Abstract:

Calibration is a core element of radiocarbon dating and is undergoing rapid development on a number of different fronts. This is most obvious in the area of 14C archives suitable for calibration purposes, which are now demonstrating much greater coherence over the earlier age range of the technique. Of particular significance to this end is the development of purely terrestrial archives such as those from the Lake Suigetsu sedimentary profile and Kauri tree rings from New Zealand, in addition to the groundwater records from speleothems. Equally important, however, is the development of statistical tools that can be used with, and help develop, such calibration data. In the context of sedimentary deposition, age-depth modeling provides a very useful way to analyze series of measurements from cores, with or without the presence of additional varve information. New methods are under development, making use of model averaging, that generate more robust age models. In addition, all calibration requires a coherent approach to outliers, for both single samples and where entire data sets might be offset relative to the calibration curve. This paper looks at current developments in these areas.