PARENT SESSION
Oral Session 102: Modeling II: Populations and Aquatic Systems.
Presiding: V Gertseva
Thursday, August 5, 8:00 AM to 11:30 AM, Meeting Room D 133.

CEL Agent IBMs: Decoding and forecasting 3-D fish response to hydrodynamic and water quality pattern.

Goodwin, R. Andrew^{*,1, 2}, Nestler, John², Anderson, James³, Weber, Larry⁴, ¹ Cornell University, Ithaca, New York² US Army Engineer Research & Development Center, Vicksburg, Mississippi³ University of Washington, Seattle, Washington⁴ University of Iowa, Iowa City, Iowa

ABSTRACT- Downstream passage of outmigrating juvenile salmon (migrants) at hydropower dams on the Columbia and Snake Rivers has not been entirely successful. Research in the region has generated considerable data, but inconclusive explanations of factors influencing migrant guidance and passage. To develop a quantitative biological hypothesis of migrant movement behavior of sufficient accuracy to assist engineering design, a coupled Eulerian-Lagrangian agent individual-based model (CEL Agent IBM) was developed. A CEL Agent IBM integrates the three primary theoretical frameworks for studying animal movement, outputs model results in the same form as collected field data, and yields a mechanistic, biologically-tractable interpretation of observed movement amenable to forecast simulation. CEL Agent IBMs are based on well-established principles in game and foraging theories, water quality and hydrodynamic modeling, psychology, and numerical methods. A prototype CEL Agent IBM, the Numerical Fish Surrogate, was developed to decode the hydrodynamic stimuli eliciting observed 3-D migrant movement and passage patterns at Lower Granite Dam, Snake River, WA. The resulting Strain-Velocity-Pressure (SVP) hypothesis describes 74% (r² = 0.74) of observed variation in passage pattern at Lower Granite Dam for one calibration and 10 validation data sets associated with different dam configurations. The Numerical Fish Surrogate provides a new, innovative means to (1) quantitatively interpret 3-D fish movement response to patterns in hydrodynamics, temperature, dissolved oxygen, salinity, and other physicochemical constituents and (2) objectively implement hypotheses of movement behavior to support water resource management. In the Pacific Northwest, the US Army Corps of Engineers uses the Numerical Fish Surrogate, and the embodied SVP hypothesis, to assist in project selection and engineering design of alternatives to improve migrant passage at Corps hydropower dams.

Key words: management, movement, IBM, fish