Chronological uncertainty severely complicates the identification of cyclical processes in radiocarbon-dated time-series

Cycles are widely considered to be an important feature of environmental and human history over the last 50,000 years. However, there is an overlooked problem in the investigation of cyclicity in this time period-the standard statistical methods for identifying cycles assume that observations are precisely dated, but the main relevant dating technique, radiocarbon dating, often yields dates with large and highly irregular uncertainties. Here, we present the results of a massive simulation study that explored the impact of radiocarbon dating uncertainty on our ability to identify cycles in time-series. Our results suggest there is indeed a problem. We found that, at best, we could correctly identify known cycles only 42% of the time and that the false-positive rate was as high as 90%. This indicates that an individual analysis of a single time-series is very likely to return false-positive cycles. One implication of this is that many of the environmental and sociopolitical cycles that have been identified may not be real. Consequently, a program of reassessment is needed.