Sensitivity of a salmon population model to alternative formulations and initial conditions

Salmon populations in many Pacific coast rivers are in serious decline and in danger of becoming threatened or endangered. A fish population model (SALMOD) that tracks fall chinook salmon has been developed for the Trinity River, California. The model considers the principal environmental factors influencing movement and mortality of young-of-the-year salmon from the time of spawning and egg deposition until they leave freshwater rivers as juveniles. Numbers of salmon produced by alternative managed flow regimes can be estimated using SALMOD. This paper explores the consequences of alternative model construction and formulation choices on model behavior, and the impact of the number of spawners returning from the ocean on development of a robust flow management decision. Results show that SALMOD is responsive to the user's choice of spawner nesting behavior (superimposition), but relatively insensitive to spatial scale describing fish rearing habitat quality. The choice of a suitable managed flow regime is sensitive to the number of adult fish returning to spawn and even more so to their distribution throughout the study area at time of spawning. It would be possible to tailor an adaptive annual flow regime based on monitoring of spawner numbers and their distribution.