Spatial ecology of chinook salmon (Oncorhynchus tshawytscha) in the Pacific Northwest: An empirical study of the effects of environmental variability and dispersal on spawning populations

This dissertation presents empirical and simulation models of chinook salmon (Oncorhynchus tshawytscha) population dynamics in the Pacific Northwest, with an emphasis on quantifying spatiotemporal patterns. The results demonstrate that variability among salmon populations arises across a range of scales. Much of the spatial variability in abundance is locally determined, with stream and watershed properties governing the potential productivity of individual populations. Despite this variability, there is a tendency for nearby populations to exhibit greater synchrony in spawner abundance. Simulation studies suggested that dispersal rates are too low to generate observed levels of synchrony. Rather, distance-dependent sharing of environmental influences is probably a more important source of synchrony. This can involve either spatial correlation in factors affecting fish in geographically distinct spawning and rearing areas, or shared experiences by fish entering the same migratory corridors and estuarine habitats as they move downstream. An example of the former, rainfall was found to be associated with variability in local salmon recruitment. However, differences in the responses of individual streams to rainfall, even for neighboring streams, coupled with the fact that rainfall remains positively correlated even at large scales, make it unlikely that this factor is the primary cause of distance-dependent decay in population synchrony. Instead, synchrony probably arises largely during portions of the life cycle when salmon from different populations occupy the same physical habitats; this conclusion is supported by the observation that interior populations that spend more time in shared portions of the freshwater system exhibit greater synchrony. Finally, year-effects on recruitment estimated over all populations reveal coherent large-scale fluctuations suggestive of a regional climatic influence, although the population-level data do not exhibit interdecadal climate-related shifts in marine productivity observed in previous studies. It is possible that some or all of the estimated year-effects correspond to El Nino or related sub-decadal climate events. When developing plans for conserving chinook populations, managers must account for factors affecting populations at all scales, consider the impact that management actions will have on relationships among populations, and recognize inherent differences in the level of synchrony for populations of different life histories and geographic regions.