Inference for population dynamics in the Neolithic period
Andrew Golightly, Newcastle University
Venue: A54, Postgraduate Statistics Centre, Lancaster University LA1 4YF
We consider a spatio-temporal model to describe the spread of the Neolithic across Europe. The arrival time of farming at a collection of radiocarbon-dated sites is modelled by a numerical solution to an advancing wavefront model. Our solution allows for uncertainty in the radiocarbon data, lack-of-fit from the deterministic model and uses a spatial Gaussian process to smooth over any locally geographic anomalies in the data. The analysis makes inferences for the parameters of the wavefront model by using fast Gaussian process emulator approximations to the wavefront model at each site, and gives insights into the importance of spatial scale.