Stock assessments are based on many assumptions about a species’ life history characteristics and population dynamics. Many of these assumptions cannot be directly observed (e.g., natural mortality rate, average lifespan, or reproductive success) and so are estimated, and in some cases probably estimated poorly. Given the inherent errors, stock assessment models and the harvesting guidelines derived from them should be relatively insensitive to small deviations in life history and population dynamics parameters. The project examines whether or not this assumption about model sensitivity is valid under several scenarios for Pacific salmon. In work to date, the fellow has used a life-history simulator for Pacific salmon to test the sensitivity of the cohort reconstruction (stock assessment model) to a range of natural mortality rates. Results show that current assumptions about salmon mortality from predation and disease tend to produce overestimates of stock size by as much as 20 percent on average.
Assessing the Robustness of the Salmon Stock Assessment Process via a Life History Simulator