Kendra E Kaiser, Alejandro Flores, Vicken Hillis

Modeling the coupled social and biophysical dynamics of water resource systems is increasingly important due to expanding population, fundamental transitions in the uses of water, and changes in global and regional water cycling driven by climate change. Models that explicitly represent the coupled dynamics of biophysical and social components of water resource systems are challenging to design and implement, particularly given the complicated and cross-scale nature of water governance. Agent based models (ABMs) have emerged as a tool that can capture human decision-making and nested social hierarchies. The transferability of many agent-based models of water resource systems, however, is made difficult by the location-specific details of these models. The often ad-hoc nature of the design and implementation of these models also complicates integration of high fidelity sub-models that capture biophysical dynamics like surface-groundwater exchange and the influence of global markets for commodities that drive water use. A consistent, transferable description of the individuals, groups, and/or agencies that make decisions about water resources would significantly advance the rate at which ABMs of water resource systems can be developed, enhance their applicability across ranges of spatiotemporal scales, and aid in the synthesis and comparison of models across different sites. We outline here a framework to systematically identify the primary agents that influence the storage, redistribution, and use of water within a given system. Reviewing previous studies that apply ABMs to water resources, we propose eight water resources agent types that capture the operational roles that modify the water balance. This typology characterizes common actors in water management systems but can be modified to represent the particularities of specific systems when more detailed information about specific actors is available (e.g. social networks, demographics, learning and decision-making processes). Application of the proposed typologies will support the systematic design and development of transferable scaleable water resources ABMs and facilitate the dynamical coupling of social and biophysical process modeling. To demonstrate, we show the conceptual development of an ABM that describes the interaction of agents within the Boise River Basin in the western United States and illustrate how those agents interact with the biophysical system.