Diversity within populations can help buffer against collapse in a changing environment. This phenomenon, called the portfolio effect, is particularly applicable to salmon stocks because varied conditions in streams can create runs with diverse traits, such as outmigration timing. In the California Central Valley fall-run Chinook system, which is heavily affected by humans through hatcheries, harvest, and habitat change, eroded diversity among runs could have contributed to the recent population collapse. In particular, hatchery release practices can change the amount of exchange among creeks: juvenile fish released farther from the hatchery and closer to the ocean are more likely to stray and return to a non-natal creek when they return to freshwater as adults to breed. Using a quantitative genetic model, we investigate the interactions and feedbacks between hatchery management practices and the portfolio effect, looking at how release practice affects population dynamics and diversity across creeks. We consider two creeks, one with a hatchery and one without, and an optimal ocean arrival time that varies annually. Specifically, we ask whether releasing hatchery fish closer to the ocean can drive the homogenization of outmigration timing across creeks and if so, what the consequences are for population dynamics.
Results/Conclusions
We find that releasing hatchery fish closer to the ocean results in both genetic and demographic effects on the populations in both creeks, not just the creek with the hatchery. Overall, the closer to the ocean hatchery fish are released, the higher the average annual combined population returning to both creeks and the higher the variability of annual total return, presenting a tradeoff between average return population size and stability through time. Additionally, the mean outmigration dates become more similar between the two creeks as hatchery fish are released closer to the ocean. Though the mean traits in both creeks shift towards each other, the average trait of the fish in the creek without the hatchery becomes more like those in the creek with the hatchery. In addition to changes in the mean trait values, the total genetic variance of the whole system also decreases as hatchery fish are released closer to the ocean, and the system becomes less diverse overall. These results indicate that hatchery release practices can drive homogenization among streams, with a tradeoff between genetic distinctiveness of creeks and average total return population size.