The absolute chronology of the Late Bronze Age settlements Kammenyi Ambar and Ust’ye I in the Southern Trans-Urals and its Bayesian analysis

Abstract

In this paper, we examine the series of AMS radiocarbon measurements (N = 52) obtained from the Late Bronze Age settlements of Kamennyi Ambar and Usty’e I in the Southern Trans-Urals, Russia. The exploratory data analysis applied to uncalibrated dates allows for the batches and outliers isolation. Furthermore, based on the facts of stratigraphy and application of the Bayesian statistics, we reconstruct the chronology, estimate spans of habitation, and discuss issues of the existing samples. As the first step of the analysis, we consider archaeo-logical contexts of the measurements and statistically identify apparent outliers. Despite the small sample size, the dataset from Ust’ye I obtained in the way that allows to date stratigraphically isolated construction/utilization episodes and thus are highly reliable. At least five measurements from Kamennyi Ambar date the natural events before the settlement construction and serve as upper limits in models. On the second stage of the analysis, the Bayesian models of the sites’ chronology constructed with OxCal 4.3. For Ust’ye I, we designed a three-phase model that allows to date two early habitational phases during which the walls and ditches were built around the settlement. The third phase defines the lower limit of the previous period, as this phase consists of a single radio-carbon measurement. The model suggests the existence of the gap between the two earliest stages, associated with the Sintashta and Petrovka ceramic types. While the sample size is small, this hypothesis agrees well with the site’s stratigraphy, as the settlement was re-modeled entirely at least once. We designed two models for Kamennyi Ambar. The first model includes all available data and consists of three phases: 1) natural events be-fore the settlement construction; 2) the early «walled» phase; 3) the late «unwalled» phase. The second model incorporates data on the wells’ stratigraphy and uses only measurements from the wells. The models almost the same for the early habitational phase, but the modeled chronology of the late phase differs drastically, as the wells-based model condenses the phase and pushes it earlier. Further, the models suggest that Kamennyi Ambar existed only for about 50 years, and Usty’e I for nearly 100 years. The habitational phases within the same peri-ods are partially desynchronized, and possible gaps in the habitation suggested by the models. One possible explanation of the short-term habitation spans and differences is that people needed to resettle regularly as re-source depletion made the areas inhospitable for the communities of the livestock breeders.