Refining Carbon Dating

But levels of carbon-14 in the atmosphere vary from year to year, so scientists need to calibrate their estimates using long-running records of radiocarbon levels. The shells of marine creatures provide one such record, but it represents the level of carbon-14 in the oceans, which does not exactly reflect the amount in the atmosphere. Cave formations like stalactites and stalagmites, which get their carbon-14 from groundwater, run into the same problem.

Trees provide more accurate readings, since they get their carbon-14 directly from the atmosphere and they lay new visible rings every year. But tree ring data only go back 13,000 years, and thus cannot be used to calibrate older dates. “The hope has always been that we’d find records that we could use for the whole period of radiocarbon dating,” said Bronk Ramsey.

Lake Suigetsu in Japan provided the answer. Due to yearly changes in the lake’s surrounding vegetation, different types of organic material settled on its bottom in summer and winter. These changes are visible in the sediment as alternating dark and light bands known as “varves.” “It’s not unusual to have lakes with varves for short periods, but to have one that extends to the last ice age is unusual,” Bronk Ramsey said.

The sediments are full of plant remains that, like tree rings, took their carbon-14 directly from the atmosphere, and can be accurately matched to a specific year using the varves as a mineral calendar. “This dataset is the only continuous atmospheric record beyond the end of the tree rings,” said Paula Reimer, an archaeologist from Queen’s University Belfast in Northern Ireland who was not involved in the study. It extends over virtually the entire timespan for which carbon-dating is used—as far back as 60,000  years or so, when the the carbon-14 in the sample has decayed to unreliable levels.

Hiroyuki Kitagawa from Nagoya University and Johannes van der Plicht from the University of Groningen found the annual varves in the 1990s. They extracted a core (a column of sediment), did some radiocarbon testing, and published their analysis in Science in 1998. But their single core had missing segments, and because they counted the varves visually, they ended up with a timeline that did not coincide with other records.

Takeshi Nakagawa from Newcastle University decided to revisit the lake in 2006. His team took three cores that overlap in several places, and used two different approaches to count the varves: they looked at them under a microscope and also tracked the chemical changes along them using X-rays. Finally, they compared their data with previous records, including tree rings and cave samples, to account for any uncertainties due to ambiguous layers.

“The authors have done an excellent job in reconstructing the chronology of the Lake Sugietsu cores,” said A. J. Timothy Jull from the National Science Foundation’s Arizona Accelerator Mass Spectrometry Laboratory. “We must exercise some caution about any lake sediment record as it's always possible that there are missing layers. However, this team seems to have done a good job in minimizing these possible effects.” 

Bronk Ramsey said the new data could reveal that current date estimates for many ancient items—any that were dated using carbon-14 calculations—are off by up to a few hundred years. Such errors are not huge, but they matter when trying to understand, for example, how prehistoric people were responding to changing climates. “There won’t be completely radical changes,” he said, “but I think everything from this time frame will be looked at again.”

The Lake Suigetsu data could also be compared to other records to compare how atmospheric changes in carbon-14 match up to oceanic levels. “Having both allows you to look at how the atmosphere and the ocean are responding to each other, with important implications for understanding how the ocean was operating in the last Ice Age,” said Bronk Ramsey.

The data will now be added to IntCal09—an internationally recognized calibration curve that combines several carbon data sets, including marine sediments, cave formations, and tree rings. Reimer says that the update will be completed by early 2013. 

C. Bronk Ramsey et al., “A complete terrestrial radiocarbon record for 11.2 to 52.8 kyr B.P.,” Science, 338: 370-74, 2012.