Search

The abundance of many Chinook salmon ( Oncorhynchus tshawytscha) stocks has declined despite reductions in harvest. We used state-space models parameterized with data from 57 Chinook salmon indicator stocks, ranging from coastal Oregon to southeast Alaska, to quantify long-term (since 1972 release year) changes in juvenile marine survival rate and mean age-at-maturity, as well as identify stock groupings with coherent dynamics. We found that juvenile marine distribution—rather than freshwater life history, run timing, or adult marine distribution—was the best predictor of trends in both survival and age. Only subyearling stocks that enter the Strait of Georgia showed evidence of transitioning to a low juvenile survival period, other groupings exhibited low and stable or cyclical patterns in survival. Conversely, declines in mean age-at-maturity were widespread and do not appear to have stabilized, suggesting that future declines in Chinook salmon population productivity may be influenced by earlier maturation or increased adult mortality, rather than further reductions in juvenile marine survival.

Recruitment variability in many fish populations is postulated to be influenced by climatic and oceanographic variability. However, a mechanistic understanding of the influence of specific variables on recruitment is generally lacking. Feeding ecology is one possible mechanism that more directly links ocean conditions and recruitment. We test this mechanism using juvenile Chinook Salmon (Oncorhynchus tshawytscha) collected off the west coast of Vancouver Island, British Columbia, Canada, in 2000–2009. Stable isotopes of carbon (δ13C), an indicator of temperature or primary productivity, and nitrogen (δ15N), an indicator of trophic position, were taken from muscle tissues of genetically stock-identified salmon. We also collated large-scale climate indices (e.g., Pacific Decadal Oscillation, North Pacific Gyre Oscillation), local climate variables (e.g., sea surface temperature) and copepod community composition across these years. We used a Bayesian network to determine how ocean conditions influenced feeding ecology, and subsequent survival rates. We found that smolt survival of Chinook Salmon is predicted by their δ13C value, but not their δ15N. In turn, large-scale climate variability determined the δ13C values of salmon, thus linking climate to survival through feeding ecology, likely through qualities propagated from the base of the food chain.