The role of life history and the environment in population dynamics of sockeye salmon
(Oncorhynchus nerka)

Processes linking the environment and life histories are central to our understanding of population dynamics. This thesis combines life history theory and environmental variation to explain recruitment dynamics among populations in Fraser River sockeye salmon (Oncorhynchus nerka). I first explore relationships between spawning stream characteristics and spawning densities and show that streams with more cover have higher spawning densities. Next, I use a 21-year time series for three of these populations to explore hypotheses about how maternal life history traits and migration conditions, experienced during upstream migrations to their spawning grounds, influence reproductive investment. Maternal body size is strongly linked to total reproductive investment and both egg mass and fecundity; however, migration difficulty only influences egg mass and not fecundity. Using the same dataset, I show that egg mass and incubation temperatures influence juvenile fitness-related traits including length, mass and emergence timing. The main finding from these analyses, that warmer incubation temperatures result in lighter juveniles that emerge earlier, led to hypotheses about how incubation temperature might select for egg size among populations. I tested these hypotheses by comparing 16 populations and confirmed the prediction that in streams with warmer water, fish would produce heavier eggs. I then asked if these same maternal traits and environmental conditions would relate to adult recruitment dynamics. Populations spawning in streams with deeper water had higher maximum population growth rates and less variable recruitment. In addition, populations in streams with larger gravel exhibited stronger density-dependence. Finally, I develop a novel framework for evaluating how habitat data, combined with the cost of collecting such information, can be used in developing cost-effective surveys. I demonstrate this general framework with a simple example using the relationships between stream characteristics and sockeye densities, considering the costs and effectiveness of stream variables. Overall, this demonstration of the joint role of maternal traits and environmental conditions in recruitment dynamics supports the potential use of such variables as indicators of population dynamics in the absence of long-term demographic data. Furthermore, it supports the development of cost-effective surveys, which is important as human impacts on populations increase, and as monitoring resources decline.