Human-induced trait change in a recently collapsed salmon population complex

Background/Question/Methods

Recent research highlights the importance of population diversity in generating a variance-buffering (portfolio) effect. Implicit is the idea that populations adapted to different conditions will differ in traits that affect their relative productivity. This bottom-up effect of inter-population trait variability on portfolio dynamics has heretofore not been explored explicitly. We address this issue through a focus on a recently collapsed salmon population complex: the fall Chinook stock complex originating from California’s Central Valley. Recent work suggests that the collapsed stock complex is comprised of multiple populations with synchronous dynamics and, thus, only a weak portfolio effect. Moreover, the strength of the portfolio effect has weakened in recent decades, possibly due to hatchery management practices. Here, we focus our attention on the timing at which salmon transition from freshwaters to the ocean, a known bottle neck period in the salmon life cycle. We postulate that hatcheries have altered the mean and variance in the timing of this transition by releasing fish over a narrow range of dates and sites. To test this idea, we have compiled data on hatchery releases across five decades from the five Central Valley hatcheries to characterize among and within hatchery stock variation in this fitness-related trait.

Results/Conclusions

Our results from one hatchery (Feather River) suggest several patterns. First, beginning in the late-1970’s, an increasing proportion of fish were released into the San Francisco Estuary, as opposed to the river itself, which affects their ocean arrival patterns as downstream (estuary) releases are released closer to the ocean than upstream (river) releases. Second, by comparing the mean and variability in release timing between those fish released to the estuary and those released in river, we found that releases to the estuary generally occurred later in the year and over a narrower range of dates than releases to the river. These results suggest dramatic changes in management practices over the last 50 years at the Feather River Hatchery, which have reduced the variability in a known fitness-related trait. Our ongoing work is exploring these same questions using data collected from the remaining four hatcheries. Our long-term goal is to understand whether this human-induced trait change has ecological consequences for the stock complex by asking whether hatchery release practices have homogenized variation in outmigration timing and intensified match-mismatch dynamics in this system.