ANNUAL VARIATION IN BIRD POPULATIONS OF MIXED-CONIFER NORTHERN HARDWOOD FORESTS

We examined annual variation in breeding bird populations at sites in north-western Wisconsin and in the Upper Peninsula of Michigan from 1986 to 1992 to determine (1) the extent to which different bird species vary in abundance over time and (2) whether or not patterns of variation differ when viewed at local and regional spatial scales. Total abundance of long-distance migrants (species that winter in the tropics) declined during the first few years in both states but subsequent increases returned abundance to levels close to those that occurred at the start of the study. Short-distance migrants (species that winter in temperate regions) showed the greatest similarity between states in the extent and direction of variation in abundance among years; abundance reached a low in both states during 1990. Unlike migrant groups, abundance of permanent residents was not correlated between states. Permanent residents declined in abundance from 1987 to 1988 in Michigan, for example, but increased in abundance in Wisconsin. Fluctuations in abundance generally were poorly correlated among groups within a state. Repeated-measures analysis of variance (one-way) revealed that abundances of 13 and 21 (62%) common, long-distance migrants, 13 of 16 (81%) short-distance migrants, and 7 of 9 (78%) permanent residents varied significantly among years in one or both states. Fluctuations in abundance likely occurred in response to events on the breeding grounds. Factors likely to have affected populations include severe drought, particularly during 1987 and 1988, as well as successional changes in habitat. It is unlikely that events on tropical wintering grounds caused population fluctuations observed in neotropical migrants. Two-way ANOVA indicated that patterns of variation differed between the two study regions (significant interaction effects) for a number of species, particularly permanent residents. Results of this study thus illustrate the importance of considering temporal variation in abundance at more than one spatial scale.