Environmental and geospatial factors drive juvenile Chinook salmon distribution during early ocean migration

Migrating animals rely on a variety of cues to guide them, but the relative importance of those signals may vary with size, life stage, or location. During their initial ocean migration, yearling Chinook salmon (Oncorhynchus tshawytcha) from the Columbia River have stock-specific spatial distributions that shift through time. We used a two-process mixture model to examine how the distribution of yearling migrants from three Chinook salmon stocks varies as a function of geospatial (e.g., latitude and distance from shore) and environmental (e.g., chlorophyll a and temperature) covariates. In this framework, one process described the probability of being inside the spatial, temporal, and environmental boundaries of the migration route, and one process described the patchy distribution of salmon abundance within that route. We found that both environmental and geospatial covariates explained substantial portions of observed spatial patterns in abundance, suggesting that these stocks responded to multiple cues during migration. However, model selection criteria indicated that fish distributions were more affected by geospatial than by environmental covariates. We conclude that during migration, behavioral responses to environmental variation are secondary to responses to geospatial variation, sometimes resulting in suboptimal environmental conditions. This may have sublethal effects on growth and could ultimately influence stock-specific responses to broad-scale climate changes.