Abstract
Human agency is a critical determinant of the dynamics of agroecosystems. However, the
manner in which agency is represented within different approaches to agroecosystem modeling is
largely contingent on the scales of analysis and the conceptualization of the system of interest. While
appropriate at times, narrow conceptualizations of agroecosystems can preclude consideration for
how agency manifests at different scales, thereby marginalizing processes, feedbacks, and constraints
that would otherwise affect model results. Modifications to the existing modeling toolkit may therefore
enable more holistic representations of human agency. Model integration can assist with the
development of multi-scale agroecosystem modeling frameworks that capture different aspects of
agency. In addition, expanding the use of socioeconomic scenarios and stakeholder participation can
assist in explicitly defining context-dependent elements of scale and agency. Such approaches,
however, should be accompanied by greater recognition of the meta agency of model users and the
need for more critical evaluation of model selection and application.
