Open Collections

Abstract

This thesis investigated juvenile freshwater and smolt-to-adult survival (φSAS) rates of threatened Upper Willamette River spring-run Chinook salmon (Oncorhynchus tshawytscha) and other spring- and summer-run Chinook salmon stocks across the Columbia River Basin. In Chapter Two, I analyzed Passive Integrated Transponder (PIT) tag data from 143 releases between 2002 and 2016 to estimate φSAS and release-smolt survival (φRSS) rates for Upper Willamette River spring-run Chinook salmon. I used a Bayesian Cormack-Jolly-Seber model with a minimum design configuration that included two detection locations: one for juvenile smolts before entry to the Columbia River estuary and one for adults on their return to the Willamette River. Sensitivity analysis results found a minimal effect of prior choice on model output results. Estimates of φSAS were consistently below the 2% recovery benchmark for the Columbia River Basin. Additionally, φRSS through river segments varied across Willamette River sub-basins, with the lowest estimates observed from release locations requiring longer migrations through multiple dams. In Chapter Three, I used published estimates of φSAS from the Snake, Middle Columbia, and Upper Columbia River and evaluated survival trends across these and the Willamette River. Common shared year effects revealed interannual variation; however, when separated, trends were not uniform across all regions, albeit with higher uncertainty. Univariate covariate analyses revealed negative associations that did not overlap with zero between φSAS estimates and the abundance of pink salmon in the Northeast Pacific Ocean, with adjusted R² values ranging from 0.02 to 0.11 across regions. Additionally, the Pacific Decadal Oscillation and El Niño-Southern Oscillation were linked to reduced survival, while the North Pacific Gyre Oscillation showed positive associations. Although these drivers showed the same associations with survival across all regions, their relative importance differed. Multivariate model covariate analyses explained up to 36% of the observed variation in survival, but found that no singular covariate appeared as a top driver across all regions. These results suggest that accounting for regional differences in response to environmental and ecological stressors may be beneficial for conservation and recovery strategies for spring and summer Chinook salmon populations in the Columbia River system.