Research Needs for Understanding and Predicting the Behavior of Managed Ecosystems: Lessons from the Study of Agroecosystems

Abstract

Managed ecosystems are complex, dynamic systems with spatially varying inputs and outputs that are the result of interrelated physical, biological, and human decision-making processes. To gain an adequate understanding of these systems and predict their behavior, we believe that it is necessary to move beyond stylized theoretical models and loosely coupled disciplinary simulation models to what we describe as “fully integrated models.” Herein we present a conceptual framework for a more integrated approach to the study of managed ecosystems using the example of agricultural ecosystems. We then propose the implementation of a research agenda that fosters coordinated disciplinary research aimed at a better understanding and quantification of linkages across disciplinary models. Key research issues include the effects of spatial scale, the assessment of uncertainty in coupled models, and methods for collecting and analyzing spatially referenced data.