CRC 1266 - Scales of Transformation

Phase 1 - Research activities 2016-2020


G1: Timescales of Change - Chronology of cultural and environmental transformations



Principal Investigators: Dr. John Meadows, Prof. Dr. Thomas Meier
Staff: Helene Rose
 

 

G1’s work-package 1 (Bayesian chronological modelling) involves creating realistic simulation models before selecting 14C samples and modelling the results. A major component has been the validation or rejection of new or legacy 14C dates, through inter-laboratory replication, quality assurance indicators and sensitivity analyses. 

In 2017, together with Martin Furholt and Nils Müller-Scheessel (subproject C2), we began work on dating the Linearbandkeramik site at Vráble, Slovakia, initially focussing on the internal chronology of the southwestern settlement. By coring geomagnetic anomalies attributed different houses, we aimed to randomly sample the whole settlement, in order to map its development through time (Figure 1).

Vrable sampling
Fig. 1. (above) development of coring strategy for 2017 Vráble field season; curves show that to maximize the number of datable houses, coring strategy should reflect anticipated success rate (% of cores yielding suitable dating material) (J. Meadows); (below) coring in action at Vráble (N. Müller-Scheessel).

Poor preservation of material suitable for dating led to a change in focus in 2018, to understanding chronological relationships between the three adjacent LBK settlements at Vráble, and what they imply in terms of population trends or fluctuations. A major challenge was incorporating dates on poorly preserved bone samples, given the lack of informative prior information which would assist us to detect misfitting results. A chronological model was developed, and a comparison of the Vráble settlement chronology to models from similar LBK sites was published in 2019.

Vrable results
Fig. 2. (above) Vráble LBK settlements and houses (rectangles), dated (bold colours) and undated (faint shading) (N Müller-Scheessel/J Meadows); (below) estimated dates of occupation and abandonment in each settlement, and settlement longevity (Meadows et al. 2019).

G1 began to work on 14C results from the Middle Neolithic gallery grave at Niedertiefenbach, Hesse, when human bone samples dated by Ben Krause-Kyora (subproject  F4) indicated a shorter chronology than those previously dated by Christoph Rinne (subproject D2). G1 funded the replication of 10 samples, confirming the shorter chronology, but the results fall on a calibration plateau, leading to large uncertainties in the dates of individual samples. Simulation modelling showed that archaeogenetic kinship information developed by Alexander Immel (subproject F4) could dramatically improve the site chronology, giving the type of accuracy and precision normally achieved only on more favourable sections of the calibration curve (Figure 3).

Niedertiefenbach simulations
Fig. 3. (above) Summary of modelled simulated potential dates from the gallery grave at Niedertiefenbach, from a model incorporating some kinship information (dark) compared to a model of the same 14C ages without kinship information; (below) comparison of model output using the same kinship information, with simulated 14C ages corresponding to burials with slightly earlier or later calendar date ranges (J. Meadows).

A chronological model, incorporating dates on 40 individuals (co-funded by subprojects D2, F4 and G1), dietary stable isotopes, age-at-death estimates, kinship and stratigraphy is under review. We argue that the burials span about a century, centred on the late 3200s cal BCE, and that some individuals (and thus their genomes) can be dated to shorter intervals within this range.

Working with Dragana Filipović and Wiebke Kirleis (subproject F3), G1 contributed to a preliminary publication on dating the spread of broomcorn millet cultivation in Bronze Age Europe, and to conference presentations and publication texts incorporating new and published 14C ages obtained by the F3 team from over 60 sites across central-eastern Europe. G1 developed a simple spatial-temporal modelling approach to map the rate of dispersal of millet cultivation, by estimating the date of millet arrival in each region (Figure 4).

Millet sites

Fig. 4. Locations of sites with directly-dated pre-Roman broomcorn millet. Colours represent arbitrary regional grouping of sites, not the millet dates themselves (J Meadows).

Helene Agerskov Rose’s PhD project (Bayesian chronological modelling of the early Iron Age in Southern Jutland, Denmark) involves selecting and dating c. 150 14C samples from 3 large urnfield cemeteries; quality assurance of the 14C ages obtained (through laboratory intercomparison and methodological investigation), and using experimentally cremated bone to quantify potential wood-age offsets in cremated bone 14C dates. Papers exploring these technical aspects are already published or accepted, and a final interpretative paper on ‘currency models’, estimating absolute date ranges for the production of typologically diagnostic metal objects found in cremations (such as belt clasps and pins), is in preparation. The rapid turnover of these types, and the shape of the calibration curve, pose significant challenges, but with the data quality and density now available, it is possible to accurately recover the date ranges of different artefact types (Figure 5), and thus to differentiate periods of faster and slower change in material culture. 

Currency model simulations
Fig. 5. Simulation of ‘currency models’ for hypothetical artefact types produced in the Danish early Iron Age. Diamonds represent calendar dates of burials containing the corresponding artefact types, whose 14C ages are simulated and included in a Bayesian chronological model. The coloured polygons are the model’s posterior density estimates of when each type was produced (J. Meadows/H. Rose). 

News
New Publications
Fieldworks
Links