Large-scale climate fluctuations, such as the North Atlantic Oscillation (NAO), have a marked effect on the timing of spring migration of birds. It has however been suggested that long-distance migrants wintering in Africa could respond less to NAO than short-distance migrants wintering in Europe, making them more vulnerable to climatic changes. We studied whether migratory boreal and arctic bird species returning from different wintering areas show differences in responses to the NAO in the timing of their spring migration. We used data on 75 species from two bird observatories in northern Europe (60°N). By extending the examination to the whole distribution of spring migration and to a taxonomically diverse set of birds, we aimed at finding general patterns of the effects of climate fluctuation on the timing of avian migration. Most species arrived earlier after winters with high NAO index. The degree of NAO-response diminished with the phase of migration: the early part of a species’ migratory population responded more strongly than the later part. Early phase waterfowl responded strongest to NAO, but in later phases their response faded to non-significant. This pattern may be related to winter severity and/or ice conditions in the Baltic. In the two other groups, gulls and waders and passerines, all phases of migration responded to NAO and fading with phase was non-significant. The difference between waterfowl and other groups may be related to differences between the phenological development of their respective macrohabitats. Wintering area affected the strength of NAO response in a complicated way. On average medium distance migrants responded most strongly, followed by short-distance migrants and partial migrants. Our results concerning the response of long-distance migrants were difficult to interpret: there is an overall weak yet statistically significant effect, but patterns with phase of migration need further study. Our results highlight the importance of examining the whole distribution of migration and warrant the use of data sets from several sampling sites when studying climatic effects on the timing of avian life-history events.