What affects the magnitude of change in first arrival dates of migrant birds?

We analysed which among four factors (mean first arrival date, migration distance, changes in population size, detectability of species) influenced the magnitude of change (regression coefficient) in the first arrival dates of 30 migrant bird species in western Poland during 1983–2003. An examination suggested that several of these factors could be important: the regression coefficient was positively related to mean first arrival date (early species advancing their arrival date more) and negatively with change in population size (species in decline changing less). Moreover, significant differences in regression coefficient were detected between short and long distance migrants and between low detectable and highly detectable species. Undertaking a principal components analysis on the four factors produced an axis explaining 59% of the variance and whose positive values were associated with late arriving, long distance and low detectable species which were more likely to be in decline. However, the multi-collinearity of these factors is a problem that cannot be resolved here and we recommend that further work from different areas is needed to tease apart these effects.