TOP   CONTENTS   HOME 

3.1 Radiocarbon Timescale and Calibration

Compensating corrections

Radiocarbon dates calculated from the ratio of modern carbon14 activity and sample activity and the half-life of carbon¹⁴ need to be calibrated to compensate for temporal variations in the concentrations of carbon¹⁴ in the atmosphere' (Klein, Lermann, Damon and Linick 1980). Development of suitable calibration scales has gone on since the late 1950s when the existence of such variations was reported (Suess 1955; and de Vries 1958 and 1959). Early schemes of calibration produced wide differences in calibrated dates, partly because methods of measurement and assumptions differed, and partly because there were comparatively few replicate measurements, among which even fewer involved more than one laboratory, thus allowing cross-checking of results.

Radiocarbon calibration and dendrochronology

However, scientific techniques have improved, more sophisticated methods have been developed, there has been increased awareness of the need to apply error multipliers to results to reflect variations between laboratories (and hence improve comparability of results), and so over the 1970s and 1980s long and previously floating dendrochronological timescales have been fixed and linked. In 1986 the internationally most acceptable calibrated radiocarbon timescales so far derived were published by Stuiver, Pearson and others in Radiocarbon 28, 2B. (Stuiver and Pearson 1986, Pearson and Stuiver 1986, Pearson, Pilcher, Bailey, Corbett and Qua 1986). The timescales cover the period AD1950 to 5210BC. They are based on dendrochronologically dated wood which was measured for carbon¹⁴ at laboratories in Seattle and Belfast. Each laboratory used different methods of measurement and different tree ring sequences from both Europe and America.

The Seattle data was based on samples spanning 10 year intervals and the Belfast data (for the most part) on 20 year intervals. Despite these and other small differences it was possible to average the Seattle carbon¹⁴ ages so that a Seattle bi-decade differed only in a few instances by more than 1.5 years from the mid-point of the corresponding Belfast samples. The wood was collected from Ireland, southern Germany and the west Pacific coastal region of the United States. The carbon¹⁴ ages of wood of the same age were proved to differ by only a few years on average; and the systematic carbon¹⁴ measurement of the wood differed between Belfast and Seattle by only a few years or less.

The results produced a very small off-set of 4 years between the two calibration curves attributable probably to laboratory bias. The laboratories have identical upper limit error multipliers. The points on the two curves have a statistical standard deviation , but compared with such deviations on many earlier calibration curves they are moderate at +/- 14 years over AD1950-2500BC (ad1950-2055bc). The results appear authoritative and provide the best high-precision calibration curves to date.

3.2 Radiocarbon Date Tables

Consequently the Stuiver and Pearson calibration curves have been adopted as the dating bench-mark here, and the Tables at the end of this section of bp, CalBC/AD and CalBP use the data reported by Stuiver, Pearson and others in the 1986 articles. Seattle data is used for the period CalAD 110 to CalBC 670 (ad54-555bc), and Belfast data for CalBC 670-4370 (518-3620bc). The ad/bc column arithmetically converting bp dates was supplied by myself. The column of bp dates includes the standard deviation reported by the laboratories, but for the sake of simplicity and clarity this deviation is omitted for the ad/bc and CalAD/DB/BP columns. Belfast data from 2500-5120BC (2053-4239bc) is in fact subject to a larger error multiplier than the preceding Belfast and Seattle data, and the standard deviation broadens CalBC 2500 (2053bc) to +/- 20 years from the preceding average +/- 14 years mentioned above. This becomes evident from the Tables.

Stuiver and Pearson calibration curves have been adopted as the dating bench-mark here, and the Tables of bp, CalBC/AD and CalBP use the data reported by Stuiver, Pearson and others. Seattle data is used for the period CalAD 110 to CalBC 670 (ad54-555bc), and Belfast data for CalBC 670-4370 (518-3620bc). The ad/bc column arithmetically converting bp dates was supplied by myself. The column of bp dates includes the standard deviation reported by the laboratories, but for the sake of simplicity and clarity this deviation is omitted for the ad/bc and CalAD/DB/BP columns. Belfast data from 2500-5120BC (2053-4239bc) is in fact subject to a larger error multiplier than the preceding Belfast and Seattle data, and the standard deviation broadens CalBC 2500 (2053bc) to +/- 20 years from the preceding average +/- 14 years mentioned above. This becomes evident from the Tables.

The sources for the Tables are as follows:

The bp, BC and BP dates in the Tables come from the above articles, and the ad/bc dates were added by the author. The raw radiocarbon data appears in the first two columns as bp+/-n years which is converted to ad/bc in the third column. This is then calibrated to CalBC/AD and to CalBP in the last two columns.

 TOP   CONTENTS   HOME 

© Internet Archaeology/Author(s)
University of York legal statements | Terms and Conditions | File last updated: Wed Nov 7 2001