Search

This paper presents new chronological data applied to the problem of providing a date for the construction of a prehistoric building, with a case study from the Old Scatness Broch, Shetland. The innovative methodology employed utilises the combination of radiocarbon and optically stimulated luminescence (OSL) dates with the archaeological information, which includes the stratigraphic relationships of sampled deposits, context information, and evidence relating to the formation of the deposit. This paper discusses the scientific validity of the dates produced, and the advantages that the methodology employed at this site offers for archaeological interpretation. The combined dating evidence suggests that the broch at Old Scatness is earlier than the conventionally accepted dates for broch construction. More broadly it shows the value of integration of the specialists at the planning stages of the excavation. The application of a Bayesian statistical model to the sequences of dates allowed investigation of the robustness of the dates within the stratigraphic sequences, as well as increasing the resolution of the resulting chronology. In addition, the value of utilising multiple dating techniques on the same deposit was demonstrated, as this allowed different dated events to be directly compared as well as issues relating to the formation of the sampled deposit. This in turn impacted on the chronological significance of the resulting dating evidence, and therefore the confidence that could be placed in the results.

In this paper, we illustrate the ways in which Bayesian statistical techniques may be used to enhance chronological resolution when applied to a series of OSL sediment dates. Such application can achieve an optimal chronological model by incorporating stratigraphic and age information. The application to luminescence data is not straightforward owing to the sources of uncertainty in each date, and here we present one solution to overcoming these difficulties, and introduce the concept of "unshared systematic" errors. Using OSL sediment dates from the site of Old Scatness Broch, Shetland Isles, UK, many measured with a high degree of precision, we illustrate some of the ways in which Bayesian techniques may be applied, as a tool for assessing systematic errors when combined with independent chronological information, and to determine the optimum chronological information for specific events and contexts. We provide a detailed procedure for the application of Bayesian methods to OSL dates using the widely available radiocarbon calibration programme OxCal.