Global challenges in the 21st century involve complex interacting systems that require diverse assemblies of investigators working collaboratively. This symposium will explore the use of community models, defined as multiagent programmable simulators capable of organic development within the user community, to address coupled problems of land, water, food, and fuel. Issues to be considered in this context include global climate change, air quality, and land use. The session will present an overview of community models as a resource for scientific collaborations and provide a current example of application of community modeling to improve predictive ability of climate-change models. It will follow with an innovative approach to convening and initiating creative collaborative research and problem-solving and discuss opportunities for further development of the community modeling approach to address complex global problems relating to agricultural production and sustainable use of land and water resources. Issues and challenges associated with transferring knowledge gained through hard science models to inform policy decisions around sustainable agriculture and resource use will also be presented.