From individual fuzzy cognitive maps to agent based models: modelling multi-factorial and multi-stakeholder decision-making for water scarcity


Policymaking for complex Social-Ecological Systems (SESs) – systems in which human and natural environment have constant interactions – is a process that must consider multiple factors and stakeholders. Therefore, proper policy simulation models in an SES should consider both the dynamic behaviour of social and ecological factors and the influence of different stakeholders’ interventions on the system. This requires integrated methodological approaches.

The authors develop an integrated modelling methodology combining an Agent-Based Model with Fuzzy Cognitive Mapping. They use this methodology to simulate impacts of policy options addressing the problems of water scarcity for a farming community in Rafsanjan, Iran. Rafsanjan is among the top producers and exporters of pistachios in the world but it is located in a semi-arid region and production has been severely threatened by water scarcity in recent years.

The results suggest that among the four policies suggested by the local government, a policy of facilitating farmers’ participation in the management and control of their groundwater use leads to the highest reduction of groundwater use and would help to secure farmers’ activities in Rafsanjan.

Key points for decision-makers

  • The pistachio farmers in Rafsanjan depend entirely on groundwater to irrigate their orchards. In the face of severe water scarcity, this study aims to support policymaking for the community.
  • Agent-Based Modelling (ABM) is an example of an actor-based/individual-level approach that represents decisions, behaviours and interactions of stakeholders, while Fuzzy Cognitive Mapping (FCM) is a factor-based/system-level approach that represents changes in factors (variables) of a system and their interactions.
  • First, the behavioural rules of farmers and the causal relations among environmental variables are captured with FCMs that are developed with both qualitative and quantitative data, i.e. farmers’ knowledge and empirical data from studies.
  • An ABM is then developed to model decisions and actions of farmers and simulate their impacts on overall groundwater use and emigration of farmers in this case study.
  • Finally, the impacts of different policy options are simulated and compared with a baseline scenario.
  • The authors find that reducing the farms’ overall size provides a high incentive policy for farmers to reduce their irrigation areas and thus decrease pressures on aquifer and groundwater use. However, due to high rates of emigration of farmers in this scenario, it is not a satisfactory policy from a socioeconomic perspective.
  • Rather, a policy to facilitate farmers’ participation in the management and control of their groundwater use has the highest impact in reducing overall groundwater use, and it reduces emigration.
  • Surprisingly, adopting new irrigation technologies does not have any significant impact on reducing overall groundwater use in the region.
  • The integrated methodology takes into account aspects of complex SESs that cannot be fully covered by either modelling approach if used individually. The approach is particularly useful for the ex-ante analysis of policy options.