Do Bayesian methods lead to more precise chronologies? ‘BayLum’ and a first OSL-based chronology for the Palaeolithic open-air site of Mirak (Iran)

Abstract

Bayesian inference has been applied extensively to chronologies in archaeological science since it provides several advantages over the (classic) frequentist approach. One of the most important aspects of applying Bayesian methods is their capacity to consider the stratigraphic relationship between ages. In luminescence dating, a crucial motivation for applying Bayesian modelling is the ability to address the systematic shared uncertainty. The recently deployed R package ‘BayLum’ was developed to ameliorate luminescence-based chronologies by employing Bayesian modelling. Our contribution aims at estimating the impact of stratigraphic order and systematic shared uncertainty on the age results. In this paper, for the first time, we present a comprehensive luminescence-based chronology for the Middle- Upper Palaeolithic site of Mirak. This open-air site is located in the northern fringes of the Iranian central desert, which is considered to be one of the dispersal corridors for hominins (Neanderthal and modern human) living across western and central Asia. We compare chronologies derived by frequentist and Bayesian methods to discuss the effect of stratigraphic ordering and the correlation between samples due to systematic shared uncertainty. Our investigations indicate that applying the stratigraphic order, when age uncertainty overlap one another, plays a fundamental role in reducing the uncertainty. At the site Mirak the obtained Bayesian chronology considering the stratigraphic order for the layer containing predominately Upper Palaeolithic techno- complex results in 21–28 ka. The age of the intermediate layer is in the range of 26–33 ka, and the lower- most layer containing Middle Palaeolithic assemblage gives the age-range of 43–55 ka. These results indicate that Late Pleistocene humans have exploited the site during MIS 3–2. Furthermore, the chronology gives further evidence to the hypothesis that the Iranian central plateau served as a frequently used habitat and dispersal corridor for human groups populating western and central Asia.