Skip to main content
U.S. flag

An official website of the United States government

Modeling reservoir operators as autonomous agents: balancing flood control with water supply

Presentation Date
Monday, December 10, 2018 at 11:50am
Marriott Marquis Marquis 1-2



Modeling water resource management necessitates integration of complex social and hydrologic dynamics. Simulation of these socio-ecological systems requires characterization of the decision-making process of relevant actors, the mechanisms through which they exert control on the biophysical system, their ability to react and adapt to regional environmental conditions, and the plausible behaviors in response to changes in those conditions. Agent based models (ABMs) are a useful tool in simulating these complex adaptive systems because they provide relatively straightforward ways to dynamically couple hydrological models and the behavior of decision-making actors. However, the variability in behavior of water management actors across systems makes characterizing agent behaviors and relationships challenging. Agent typologies group together individuals and/or agencies with similar functional roles, management objectives, and decision-making strategies, thereby simplifying the representation of water management across river basins and increasing transferability and scaling of ABMs. Here we present a framework for identifying and classifying major water actors and apply this modeling framework to the Boise River Basin in southwest Idaho. Precipitation in the upper basin supplies 90% of the surface water used in the basin, thus managers of the reservoir system (located in the upper basin) must balance flood control for the metropolitan area with water supply for downstream agricultural and hydropower use. We modeled reservoir outflows based on operational targets set by the U.S. Army Corps of Engineers Water Control Manual and found that based on observed storage levels, the dam is filled, on average, 50 days later than it could be. Characterizing this bias is imperative for modeling coupled natural-human systems where upstream reservoir management dictates the day of allocation and associated curtailments of junior water rights. This modeling exercise demonstrates how to represent autonomous reservoir management decision-making and the potential for agent typologies to increase the transferability and scaling of water management ABMs in the western US.

    Funding Program Area(s)