Potential Mechanisms of Population Change in Southwest Alaskan Sockeye Salmon

Abstract

Bristol Bay, Alaska, is home to the world’s largest commercial sockeye salmon (Oncorhynchus nerka) fishery. This fishery not only produces more than half of all sockeye salmon globally, but also supports cultural and economic livelihoods across North America. Over the last century, there have been substantial changes in this fishery due to shifts in climate, biological composition of the ocean, and management practices and sockeye populations fluctuated drastically in response. The freshwater and marine life stages of Pacific salmon are deeply interconnected and although historically the freshwater life stage has been much more thoroughly researched, substantial knowledge gaps exist in both of sectors. In order to thoroughly understand how the dynamics of this valuable resource have shifted through time, it is necessary to fill in the knowledge gaps in both the marine and freshwater systems impacting sockeye salmon population responses. I explore two potential mechanisms of sockeye salmon production through time; marine trophic shifts and density dependence on the spawning grounds. In chapter 1 I use a novel compound specific isotope analysis (CSIA) approach to reconstruct 60 years of trophic position and nitrogen assimilation in the North Pacific Ocean and relate observed changes to regionally relevant environmental shifts. I find that periods of high productivity are associated with truncated food chain lengths and enrichment of the nitrogen isotope baseline of the ecosystem. Although I do not directly assess how food chain length relates to salmon production in the region, I identify this as an area that requires future exploration. In chapter 2 I build a population model to explore how density dependence on the spawning ground impacts recruitment and apply this model to a case study of two small creeks in Southwest Alaska. This model suggests a relationship between environmentally optimal spawn timing in a system and stock-recruit relationships, however when applied to our empirical data, this relationship does not hold up, indicating the need for further investigation.