Establishing a Bayesian approach to determining cosmogenic nuclide reference production rates using He-3

Abstract

Production rates are a cornerstone of in situ cosmogenic nuclide applications, including surface exposure dating, erosion rate/denudation rate estimates, and burial dating. The most common approach for estimating production rates is to measure cosmogenic nuclide samples from sites with independently well-constrained exposure histories. In addition, while researchers attempt to minimize the effects of erosion through careful site and sample selection, it can be present at some unknown level in certain sites. We present a general Bayesian methodology for combining information from the nuclide concentrations, the exposure history, and the possibility of erosion, to determine the production rate at a given site. Then, we use another Bayesian approach to combine the results from the various sites.Cosmogenic 3He is an ideal test-bed for our Bayesian approach. It has the most calibration sites of the commonly measured cosmogenic nuclides, and there is evidence for the effect of erosion on some of the sites. Our approach largely reconciles previous discrepancies between sites of widely varying age, even at latitudes where geomagnetic effects are significant. With the canonical Lal/Stone scaling scheme, we derive a global sea level high latitude 3He production rate of 118 ± 2 atoms g−1 yr−1 when considering olivine and pyroxene together. Using the Lifton–Sato–Dunai scaling scheme yields a similar rate of 121 ± 2 atoms g−1 yr−1. Uncertainties associated with these values are improved over previous studies, due to both reduced scatter among the sites and an approach to combining sites which deemphasizes outliers.