Replicating measurements to quantify smaller perturbations in tree-ring 14C records

Rapid increases in cosmogenic isotopes in tree-ring and ice core records point to rare, extreme perturbations in the Heliosphere over the past millennia. Identifying additional radiocarbon spikes-so-called Miyake events-and replicating or refuting published candidate events have since become an international priority. Yet, at present there are large gaps, both in magnitude and understanding, between more frequent and lower-energy events identified in space-age measurements and larger rare events identified over the past 15 ka from palaeo archives. Here, we generate additional annually-resolved radiocarbon data for a few short intervals between 3-4 ka before present (BP, i.e. 1950 CE), where initial radiocarbon data suggest either an increase in radiocarbon production or alternatively significant noise arising from the scale of error on the radiocarbon measurement continuum. By performing multiple high-precision radiocarbon measurements per analysed tree-ring year on the same cellulose samples, we reduce uncertainties and thus increase the sensitivity for detecting potential events. In addition to refining the structure of radiocarbon time-series during the Holocene, our findings contribute to the identification and characterization of a likely Miyake-type event in approximately 3481 BP/1532 BCE, and suggest further work is needed to either confirm or refute possible events at approximately 3763 BP/1814 BCE (moderate confidence) and approximately 3060 BP/1111 BCE (low confidence). Initial carbon cycle modelling of these event candidates suggests that they are two times smaller than any other of the Miyake events identified and confirmed in paleoenvironmental data. We also discuss sources of variability and uncertainty in the tree-ring radiocarbon measurement continuum. Using conventional dendrochronological statistics on radiocarbon time-series, we show that dramatic increases in replication (50-100x) would allow datasets to capture 95% of the theoretical variance in the inter-annual to medium frequency domains and, thereby, advance quantification of solar, geomagnetic and carbon cycle variation in Holocene tree-ring records, such as potentially distinguishing high versus low amplitude Schwabe solar cycles over the past millennia. This calls for a big-science approach-more typical for the space sciences-to produce and analyse such high-precision paleoenvironmental records. This article is part of the Theo Murphy meeting issue 'Radiocarbon and cosmic radiation events'.