10-5 Estimating Freshwater Life-Stage Production of Hanford Reach Fall Chinook Salmon Using a Spatially Explicit, Individual-Based Model

Brian Bellgraph , Pacific Northwest National Laboratory, Richland, WA
Geoffrey A. McMichael , Pacific Northwest National Laboratory, Richland, WA
Daniel B. Hayes , Department of Fisheries and Wildlife, Michigan State University, East Lansing, MI
Brian M. Roth , Fisheries and Wildlife, Michigan State University, East Lansing, MI
William A. Perkins , Pacific Northwest National Laboratory, Richland, WA
Marshall Richmond , Pacific Northwest National Laboratory, Richland, WA
Todd N. Pearsons , Grant County Public Utility District, Ephrata, WA
Flows at Priest Rapids Dam on the Columbia River, U.S.A. are currently managed to protect the spawning, incubation, and rearing life-stages of Hanford Reach fall Chinook salmon.  However, a thorough understanding of the relationship between flow and freshwater life-stage dynamics limits the ability of fisheries managers to evaluate alternative flow scenarios on fall Chinook salmon productivity.  To understand these relationships, an individual-based model is being created that links the temporal and spatial variability of habitat in the Hanford Reach to the population dynamics of each freshwater life stage of fall Chinook salmon.  The model is divided into spawning adult, in-gravel eggs and alevin, and free-swimming juvenile life-stages.  Life-stages are further divided into sub-models that explicitly delineate critical components of each life-stage.  Habitat inputs (e.g., depth, velocity, bed slope, substrate type, and temperature) needed for each sub-model are obtained through a dynamic link with a 2-dimensional hydrodynamic model with computational cells at 5–10-m resolution.  Outputs include estimates of production and mortality rates by source (e.g., dessication, high temperature, stranding, and predation) for each life-stage.  A critical advantage of this modeling approach is that predictions are available at a variety of scales, which provide critical checkpoints for model validation and calibration, and allow the user flexibility to investigate an array of scenarios.  Upon completion, this model will inform conservation and management of Hanford Reach fall Chinook salmon, which are considered one of the most productive Chinook salmon stocks of North America.  Challenges and innovations associated with model development will be presented to benefit those developing similar models for other species and areas.