A Bayesian network approach to model farmers' crop choice using socio-psychological measurements of expected benefits of ecosystem services

Abstract Models of ecosystem management typically measure the benefits of ecosystem services in terms of ecological or biophysical variables, which are influenced by management decisions and biophysical/ecological conditions. This study uses farmers' expected benefits of ecosystem services as input variables to model their decision between planting rice, annual crops or perennial crops. Based on the theory of planned behavior, a Bayesian network is constructed to model crop choice depending on attitudes toward the ecosystem services of biomass production, reduction of soil erosion, and water quality improvement. The relative importance of these decision-making criteria is quantified using the Analytical Hierarchy Process. Results indicate that Bayesian networks can use socio-psychological measurements to model decision-making. Especially as an extension to biophysical or economic models, they can serve as a powerful tool for grasping the more abstract socio-psychological dimensions of benefits of ecosystem services, and how they translate into the decisions of ecosystem managers.