Different models, different outcomes? A comparison of approaches to land use modeling in the Dutch limes

Current advances in spatial simulation bring unprecedented possibilities for spatio-temporal modeling. In this paper, we focus on modeling the impact of settlement on land use in the Roman period in the Dutch river area, on the northern frontier of the Roman Empire. During this period, the area witnessed a strong population increase that put more demands on the available land to produce food, not only for the local population, but also for the soldiers stationed on the frontier and the citizens of the newly founded towns. We compare an agent-based model (ABM) of agricultural production in the region (ROMFARMS), and a model using the Past Land Use Scanner (PLUS. Both were used to estimate the effects of increased agricultural demand through simulations of food production, taking into account the available workforce and the productivity and availability of suitable land. However, how should we evaluate the model outcomes? What are the advantages and limitations of each? We discuss issues of scale, temporal resolution and model inputs, together with questions of technical implementation and validation. In this way, we aim to point the way to future researchers to implement these approaches effectively in other contexts.