An activity based cellular automaton model to simulate land use changes

This paper proposes a spatial explicit model for land use dynamics, based on activities, which can represent for example population or jobs. In land use models based on constrained cellular automata (CA), the total area per land use is defined exogenously, while the model computes the allocation based on transition rules. The activity based model is a CA model, but it is constrained by activities instead of areas. Each time step, activities are distributed over cells based on transition rules. The CA transition rules comprise the effects of the activity in the neighbourhood, the land use of a cell, externalities and a stochastic perturbation term. Land use then is computed based on the activity distribution. Hence land use and activities are mutually dependent and each cell has two values: a land use state and an amount of activity. The activity based model is applied to simulate population dynamics and land use changes in Spain. Simulation results show that the model can produce realistic land use dynamics. Moreover, we argue that the inclusion of activities closer resembles the process of real world land use dynamics and offers good opportunities for integrated modelling.