Agent-based simulation of effects of social interactions on climate policy outcomes

Project Contact

Nick Gotts, Macaulay Institute
Gary Polhill, Macaulay Institute

Project Summary

Work in this strand so far has used the FEARLUS (Framework for Evaluation and Assessment of Regional Land Use Scenarios) modelling system to examine the potential for harnessing social and informational interactions between farmers to facilitate the reduction of agricultural greenhouse gas emissions. Current models are based on the Upper Deeside catchment, where livestock farming, with the attendant methane production, is the main activity; and on the possibility of using collective payments to farmers who collectively keep their emissions below a specified ceiling. Emissions will be much easier and cheaper to measure or estimate on a multi-farm than a single-farm basis, and there is the possibility farmers seen as risking the collective reward will come under social pressure to reduce emissions for that reason. The agents in current FEARLUS models, developed over several years, are motivationally and cognitively quite sophisticated: they have aspirations both for financial return and for social status with their neighbours; the relative importance attached to these factors by an individual changes in response to specific events. They may imitate successful neighbours, learn from past experience, and ask for advice.

Early results suggest that collective payments alone reduce the overall emissions level, even if farmers are assumed not to care for their neighbours' good opinion; but the effect is considerably enhanced if it is assumed that they do. The threshold below which the collective reward is paid out has an important influence: if it is too low, there is too little opportunity to learn that the reward is achieveable; if it is too high, less than optimal emissions reduction will be gained for any specific outlay in payments.

Future work with FEARLUS will examine likely farmer responses to a wider range of climate policy instruments, in the context of expected trends in Scottish farming over the next few decades.

The strand will shortly be augmented by work on a new agent-based modelling system, which will widen its scope beyond agriculture. CARLESS (Climate And Rural Landuse Enviro-Social Simulator) will be used to investigate the potential of a range of policy instruments to encourage greenhouse gas (GHG) emission reductions in Scottish rural communities. CARLESS will use an innovative approach to agent-based modelling, highly modular, and based on a modelling platform, AMEBON (Approach to Model-Evidence Bridging with an Ontology Network), designed for modularity and ease of understanding and making use of the ontology language OWL. OWL ontologies will be used to link CARLESS models explicitly with the empirical and theoretical motivation and justification for specific modelling choices they embody.

Publications

• Gotts, N.M. and Polhill, J.G. (2007) Using Collective Rewards and Social Interactions to Control Agricultural Pollution: Explorations with FEARLUS-W. In: Interdisciplinary approaches to the simulation of social phenomena. (ed. F. Amblard). Conference of the European Social Simulation, Association, 4th, Toulouse, France, 10th-14th September
  2007, pp253-262.
• Polhill, J.G. and Gotts, N.M. (2006) A new approach to modelling frameworks. Proceedings of the First World Congress on Social Simulation, Kyoto University, Kyoto, Japan, 21-25 August 2006, Vol 1. pp215-222.
• Polhill, J.G. and Gotts, N.M. - (2007) Using social interactions in the control of diffuse agricultural pollution: modelling with FEARLUS-W. Annual International Conference at the Royal Geographical Society with IBG, London, 29-31 August 2007