Metrics and sampling designs for detecting trends in the distribution of spawning pacific salmon (Oncorhynchus spp.)

Abstract

The distribution of individuals among populations and in space may contribute to their resilience under environmental variability. Changes in distribution may indicate the loss of genetically distinct subpopulations, the deterioration of habitat capacity, or both. The distribution of Pacific salmon (Oncorhynchus spp.) among spawning locations has recently been recognized as an important component of status assessment by USA and Canadian management agencies, but metrics of spawning distribution have not been rigorously evaluated. We evaluated three metrics of spawning distribution and four sampling designs for their ability to detect simulated contractions in the production of coho salmon (Oncorhynchus kisutch). We simulated population dynamics at 100 sites using a spawner-recruit model that incorporated natural variability in recruitment, age-at-maturity, dispersal, and measurement error in observations of abundance. Sensitivity analyses revealed that high observation error and straying of spawners from their natal streams may mask changes in distribution. Furthermore, monitoring only sites with high spawner abundance, as is often practiced, failed to capture the simulated contraction of production, emphasizing the importance of matching monitoring programs with assessment objectives.