Ecosystem effects on harvested populations: Lower trophic level dynamics in the Northeast Pacific and its implications on sockeye salmon (Oncorhynchus nerka) survival

Almost all epipelagic fish species in the Northeast Pacific show an increase in population size between the late 1950s and the 1980s. The complexity of pelagic ecosystems makes speculations on the causes of these increases easy to justify, and thus various conjectures on the chain of events leading to increased fish survival have been put forward.; In this thesis I try to explain the variability in cohort survival, abundance and distribution of sockeye salmon (Oncorhynchus nerka)---the fish species that has experienced the largest increase in abundance and biomass of all epipelagic fish species in the Northeast Pacific between the late 1950s and the 1980s---by ecosystem effects. I assumed that sockeye salmon total survival rate is largely determined in early marine life due to exposure to predators, which is set by the time at risk of predation, itself a function of sockeye prey, i.e. mesozooplankton, abundance. I then developed two simple food chain models with three and four trophic levels, respectively, which include lower trophic level dynamics but not fish itself. Both population models were calibrated and tested for two locations in the Northeast Pacific through mean field simulations driven by abiotic environmental forcings. Using a 4-hour time step from 1950 to 1990, both calibrated population models were then run as spatially-explicit simulations with a resolution of one degree latitude and longitude for the whole area of the Northeast Pacific, a total of 1240 open ocean fields. To assess the relative importance of biological processes versus physical advection both population models were simulated with and without surface currents.; I have tried to design the best models within reason utilizing the best information on environmental forcings and biological processes available at the time. Simulation results do not suggest a clear linkage between prey density in the oceanic environment and sockeye salmon cohort survival. However, there are two fundamental lessons to be learned from this modeling exercise: First, categorization of ecosystem components into trophic levels with no regard of the many life history strategies is one of the worst aggregation errors in ecology, one that implicitly includes errors of hierarchical organization as well as of spatio-temporal stability. And second, the complexity of ecosystems will always make results from trophodynamic simulations interpretable, even if these results bear no relationship to the natural system.