A Bayesian Approach to Inferring Depositional Ages Applied to a Late Tonian Reference Section in Svalbard

Abstract

Increasing application of high precision uranium-lead (U-Pb) and rhenium-osmium (Re-Os) geochronology to the ancient geological record has resulted in massive improvement in age control and calibration of key Proterozoic stratigraphic successions and events. Nevertheless, some successions and time intervals remain poorly dated. Insufficient age constraints are particularly problematic for successions that are otherwise rich in geochemical, fossil, or other data with high potential to illuminate our understanding of Proterozoic Earth history. The latter Tonian succession in northeastern Svalbard is one such example. The ca. 820–740 Ma Akademikerbreen and lowermost Polarisbreen groups contain important microfossils and well-established carbon- and strontium-isotopic records, but they remain poorly dated. Here we use radioisotopic dates correlated from other Tonian successions across the globe using carbon isotope chemostratigraphy to calibrate a Tonian composite section in Svalbard by integrating Bayesian inference with a simple 1D thermal subsidence model. This approach allows us to assign realistic ages and uncertainties to all stratigraphic heights in a Akademikerbreen-lower Polarisbreen composite reference section. For example, the Bayesian age-height model yields ages for the onset and end of the Bitter Springs negative carbon isotope anomaly of 808.7 +3.3/−3.5 Ma and 801.9 +3.2/−3.3 Ma, respectively, and a total duration of 6.9 ± 0.2 Ma. These age and duration estimates can be applied to calibrate other Tonian successions that capture the Bitter Springs anomaly assuming that this anomaly is globally correlative.