A new clustering approach to determine single-grain equivalent-dose populations in partly mixed samples
Abstract. Single grain luminescence dating has many advantages, including the identification of individual equivalent dose (De) populations, the exclusion of outliers and the possibility to account for beta dose-rate heterogeneity. However, there is no single age model to deal with all these factors in a sample independent manner: this presents challenges for the objective analyses of sample suites with high variability in De distributions and overdispersion (OD). We observed these issues in samples from lake shoreline dunes in southeastern Australia. To address the problem, we propose a workflow to determine one or more De populations based on hierarchical clustering using the Bhattacharyya distance between the grains, which are considered to represent lognormal distributions. For each individual determined De population, we applied the Bayesian Outlier Model (BOM) (Li et al., 2022) with fixed prior information for the entire sample suite, sharing similar depositional assumptions. This approach enables us to deal objectively with samples of varying degrees of mixing regardless of whether they are composed of one or multiple populations with the same or different percentage OD. We demonstrate that the workflow can be robustly applied across a large sample suite, although assumed De distributions may need sample-dependent adjustment for young samples. We detail how for four individual samples, the results strongly align with stratigraphic context and are in agreement with existing age models while providing improved insights.
Competing interests: We declare that one of the co-authors is in the editorial board of Geochronology. There are no other competing interests by the authors.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.