Integrating actor dynamics with land use cellular automata for modelling climate and environmental policy implementation at regional level

Successful implementation of environmental policies, including climate adaptation and mitigation policies, requires careful consideration of regional and local conditions. Consequently, there is growing understanding that regional models are needed to support climate and environmental policy making. Such models need to take into account the dynamics of geographical space as well as historic and expected future land use change patterns. One relevant geographical modelling approach is based on cellular automata (CA) which has a prominent track record of successful application to a diverse range of geographical problems. Traditionally, CA models are calibrated to reproduce the footprint of actor decision-making manifested in historical land use dynamics, and then projected forward to explore the effect of the observed dynamics on future periods. However, this is a poor representation of the way the world actually works, since policy decisions reflect current needs and priorities, not historic ones. Such a model cannot help us understand how decision-making actors might respond spontaneously to emerging land use outcomes. For these reasons, we believe there is considerable scope for existing CA-based geographical models to be improved by introducing realistic representations of the dynamic behaviour of decision-making actors. We present a modelling approach which retains the well-attested benefits of CA land use models, but which allows greater flexibility in modelling the dynamic behaviour of actors for particular “policy driven” land uses. To implement our approach, we integrate the APoLUS model (APoLUS stands for Actor, Policy and Land Use Simulator) – an open-source, multi-platform model based on geographical CA – with a system dynamics (SD) model describing the actor dynamics. The SD model is tailored to reproduce the dynamics of interaction (and possible conflicting interests) of a number of aggregate actors that might influence regional development in general and might affect (either in positive or in negative way) the implementation of policy under study in particular. In the present paper, we describe new developments in the actor dynamics model family, progressing beyond earlier work in three key ways: (i) incorporating the 2 possible ‘regime shifts’ that might be related, in particular, to election cycles; (ii) describing in more detail the economic drivers of actor dynamics; (iii) introducing the stochasticity in a SD model.