Archaeological Bayesian modelling

Prior knowledge of phases and stratigraphy can be included along with dating evidence in Bayesian models to help constrain the age and time span of phases, gaps, etc. The program OxCal (Bronk Ramsey 2009) was designed for this purpose and is widely used in archaeology.

If you require help constructing and running OxCal age-models for your site(s), Dr Cormac McSparron from the School of Natural and Built Environment has extensive experience in the field and can build and run the models for you. Please contact Cormac at c.mcsparron@qub.ac.uk to discuss options.

A Bayesian way to produce age-depth models constrained by chronological/stratigraphical ordering of the dates. Material deeper down in a core will be older than material further up, so, we know that any age-model should increase monotonously, and not show any “leaps back in time”. We simulate calendar ages from each of the calibrated ¹⁴C ages (blue lines), and construct an age-depth model through these points (linear interpolation between the dated levels). We repeat this process many times. If the age-model shows any negative accumulation rates, we reject that model (red lines). If it increases monotonously, without any negative accumulation rates, we accept the model (green) and store it (grey lines). For the Bayesian model based on the prior knowledge of stratigraphically ordered dates, we will only use the accepted age-models (the growing grey blur).