ASSESSMENT OF AMS RADIOCARBON DATING RESULTS ON POLLEN SAMPLES PURIFIED BY FLOW CYTOMETRY: RELIABILITY FOR BUILDING LAKE-SEDIMENT AGE MODELS

A new method for preparing pollen samples for radiocarbon dating, using refined chemistry followed by flow-cytometry sorting, was recently published (Tunno et al. 2020, Front. Ecol. & Evol. doi: 10.3389/fevo.2021.668676), continuing the quest for an efficient, reliable procedure for pollen dating. Here we present the radiocarbon results for variety of pollen samples prepared by this method, and consider the utility of the dates in context. Modern pollen collected from fresh pine cones was measured to test for addition of ¹⁴C-free carbon during the processing, and returned consistently near-modern ages. Surface samples were collected in the Mono Lake basin (eastern Sierra, California) to assess the potential age of pollen deposited on the landscape, from the lake shoreline to high elevations on the Sierran front. The sediments were characterized by variable grain size, moisture content, and vegetation cover, and yielded widely variable amounts of pollen. Radiocarbon measurements on pollen from these surface samples generally validated the conventional wisdom that pollen is stored for at most a few decades in near-surface sediments. Comparison of fifteen pairs of macrofossil and pollen dates from Fallen Leaf Lake, a small, oligotrophic lake south of Lake Tahoe in the Sierra Nevada, show nearly perfect overlap, with one pollen date slightly older than the macrofossil, and two pollen dates significantly younger (~300 yr) than the corresponding macrofossils. Three pairs of macrofossil-pollen dates from Late Holocene sediments at Mono Lake also show good overlap within 2-sigma uncertainties. Additional pollen dates have been measured from deglacial and Early to Mid-Holocene sediments in Mono Lake cores, whose age models are mostly constructed from macrofossil dates. In some cases the pollen dates provide age information for intervals which otherwise have no age control, and the ages fall reasonably within Bayesian-derived age models; in one case a pollen date provides a better estimate than a nearby macrofossil, which had a large uncertainty due to its very small size. Further dating of co-occurring macrofossils and pollen dates in a wider age range and in different depositional settings is required to increase confidence in unpaired dates, but these initial results are promising.