Idiosyncratic changes in spring arrival dates of Pacific Northwest migratory birds.

Shifts in the timing of bird migration have been associated with climatic change and species traits. However, climatic change does not affect all species or geographic locations equally. Climate in the Pacific Northwest has shifted during the last century with mean temperatures increasing by 1 °C but little change in total annual precipitation. Few long-term data on migration phenology of birds are available in the Pacific Northwest. We analyzed trends in spring arrival dates from a site in the Oregon Coast Range where nearly daily inventories of birds were conducted in 24 of 29 years. Several species showed statistically significant shifts in timing of first spring arrivals. Six of 18 species occur significantly earlier now than during the initial phase of the study. One species arrives significantly later. Eleven show no significant shifts in timing. We associated trends in spring migration phenology with regional climatic variables, weather (precipitation and temperature), traits of species such as migration strategy, foraging behavior, diet, and habitat use, and regional trends in abundance as indexed by Breeding Bird Survey data. We found no set of variables consistently correlated with avian phenological changes. Post hoc analyses of additional climate variables revealed an association of migratory arrival dates across the 18 species with rainfall totals in northern California, presumably indicating that songbird arrival dates in Oregon are slowed by spring storm systems in California. When only the six species with the most strongly advancing arrival dates were analyzed, winter maximum temperatures in the preceding three winters appeared consistently in top models, suggesting a possible role for food availability early in spring to promote the survival and successful reproduction of the earliest-arriving birds. However, additional data on food availability and avian survival and reproductive success are required to test that hypothesis. Despite the appearance of some climate variables in top models, there remains a mismatch between strongly advancing arrival dates in some songbirds and a lack of clear directional change in those climate variables. We conclude that either some previously unrecognized variable or combination of variables has affected the timing of migration in some species but not others, or the appearance of statistically significant directional changes over time can occur without being driven by consistent environmental or species-specific factors.