Agent-based modelling to simulate farmers' sustainable decisions: Farmers' interaction and resulting green consciousness evolution

Agent-based models (ABMs) have been adopted to simulate different kinds of complex systems, from biological systems to complex coupled human-natural systems (CHANS). In particular, when used to simulate man managed systems, they allow considering human behavioural aspects within the modelling framework. On the other hand, environmental Life Cycle Assessment (LCA) has become an acknowledged tool in research, industry and policy to assess systems' environmental sustainability. More recently, LCA is being applied to assess the potential environmental impacts of large scale policy actions (e.g., actions to combat climate change). This paper describes the application of a coupled ABM-LCA model to simulate cropping activities in the Grand Duchy of Luxembourg. The ABM considers farmers' proneness to risk, which was assessed via a naive Bayesian model trained with the results of a survey distributed to the farmers of the study region. The goal of the study is to assess the effects of the agents' interactions, that can take place in a farmers' social network, on the agricultural activities. Geographic Information System (GIS) information, national statistics and naive Bayesian model are used to parameterize agents' behaviour and interaction rules. We believe such assessment is necessary for the successful design of public adaptation strategies and subsidy schemes since governmental adaptation actions are needed to reduce emissions due to agricultural activities. Two scenarios (with different levels of farmers' environmental awareness) were simulated. The results show that the mean and variance of the distribution of farmers' environmental awareness change due to the effect of the interactions and, as a consequence, farmers' long-term decisions concerning agricultural activities are affected. This is reflected in the environmental impacts generated by such activities.