Quantifying Decision Uncertainty in Water Management via a Coupled Agent-Based Model
Modeling water resource management is a challenge because of the interactions between human decisions, the natural hydrologic cycle, and the impact of risk perception on human decision-making. A study by scientists at Lehigh University, Sandia National Laboratories, and the National Renewable Energy Laboratory (NREL) showed that risk perception can be addressed via the Theory of Planned Behavior, thus improving model representations of how people make water management decisions.
This approach improves on rule-based risk decision making by considering how previous experiences and new information play a role in the decision-making process. Analysis of how farmers manage annual irrigation acreage demonstrates the dynamic nature of decision making, something that is essential to represent in future research regarding evolving natural factors. The approach also allows for a more flexible representation of real-world decision making that can be further expanded to various spatial scales in the future. Results showed that farm location upstream or downstream of a reservoir will affect farmers’ risk perception regarding water availability and influence their behavior about expanding irrigation areas.
Researchers “two-way” coupled an agent-based model (ABM) with a river-routing and reservoir management model (RiverWare) to address the interaction between human-engineered systems and natural processes while quantifying the influence of incomplete/ambiguous information on decision-making processes. The ABM combines Bayesian Inference mapping with a Cost-Loss model to simulate farmers’ psychological risk-based decision processes under evolving socioeconomic conditions. The San Juan River Basin in New Mexico, USA is used to demonstrate the utility of this method. The calibrated model captures the annual variations of historical irrigated areas. The results suggest that the new approach provides an improved representation of human decision-making processes and outperforms the conventional rule-based ABMs that do not consider risk perception. Future studies will focus on modifying the Bayesian Inference mapping to consider farmer interactions, the up-front costs of farmer decisions, and upscaling this method to the regional scale.