Estimating the scale-dependent influence of natural terrestrial corridors on the positioning of settlements: A multi-scale study of Roman forts in Wales

Natural terrestrial corridors have been shown to have influenced the positioning of past settlements. The scale at which this pattern-process relationship operates is often un-estimated and thus remains unclear. This paper proposes the comparison of multiple point process models as an approach for estimating the optimal scale at which this relationship is strongest. With this approach, it is revealed that the positioning of Roman forts used during the conquest of Wales was most influenced by natural terrestrial corridors at a scale of 1,100m. At this scale, the Roman army stationed at these forts could control natural corridors – both via on-the-ground response as well as through overseeing movements by the native communities of Wales. Comparing multiple scenarios, it is also shown that the control of river-systems did not influence the positioning of Roman forts at the expense of controlling these natural terrestrial corridors used by those on foot. With archaeological interpretations susceptible to change as a result of the scale at which this pattern-process relationship is measured, the estimation of the optimal scale is pivotal for non-biased inferences on the processes thought to have influenced the positioning of settlements in the past.