ELF Communications System Ecological Monitoring Program: Bird Species and Communities: Annual Report 1989-1990

Abstract

This investigation was designed to isolate effects of electromagnetic (EM) fields produced by extremely low frequency (ELF) antenna systems on bird species breeding in or migrating through Wisconsin and Michigan. Specifically, we seek to determine if bird species richness and abundance differ between areas that are close to the antenna and those that are far enough away to be unaffected by the antenna. We are pursuing this question at both the community and species level. Characteristics examined include total species richness and abundance, abundances of common bird species, and abundances of birds within selected guilds. Our monitoring program has included bird censuses over a five month period from May to September (1986-1989). Additional data were collected in both states in August-September of 1984 and June of 1985. Research in Wisconsin was completed in 1989 but has continued in Michigan. A final report summarizing work in Wisconsin has been completed (Hanowski et al. in press). Here we summarize results of our 1990 research activities in Michigan. The Michigan transmitter began 150 amp tuning and testing intermittently in the first part of May 1989. On 14 May, the transmitter began continuous 150 amp operation for 16 hrs/day on weekdays and all day on weekends. On 7 October 1989, the Michigan transmitter began continuous operation at full power. We therefore consider 1990 to be the first full impact year. Overall, bird abundance and species diversity were highest and approximately the same during May, June, and July. Species diversity was significantly greater on control areas during June and September, but no other differences in community level parameters were significant. Considerable annual variation in numbers of individuals and species was noted. Particularly abundant species (all seasons included) included the Ovenbird, Black-capped Chickadee, and Nashville Warbler. Other common species included Red-eyed Vireo, White-throated Sparrow, Golden-crowned Kinglet, Hermit Thrush, and Black-throated Green Warbler. The most abundant species present on treatment and control segments varied among seasons. Among "abundant" species (>1 individual observed/500 m segment), five of 24 comparisons (21%; all seasons combined) revealed a significant difference between treatment and control segments in Michigan; two indicated a greater abundance on control segments. Previous analyses of vegetation on Michigan study sites (Blake et al. 1988) revealed differences between treatment and control plots. The difference most likely to influence bird populations was distribution of coniferous and deciduous habitats. Treatment segments supported more coniferous and lowland habitats than did control segments. It is important to note that habitat differences that exist between treatment and control areas will not affect our analysis of antenna effects. The Michigan study is designed as a before-and-after experiment; we can compare changes in bird abundance over time on treatment segments and on control segments. If electromagnetic fields produced by antenna operation affect bird distribution patterns, we expect to detect a change in patterns of abundance between treatment and control areas. Such changes, if they occur, would be independent of already present habitat differences. Seventeen of 114 comparisons (15%) of common species (based on prominence values, see page 7) between treatment and control segments (all segments) in Michigan were significant. Values were higher on control segments in 10 cases. Few species were consistently and significantly more abundant on either treatment or control segments among seasons within a year or within seasons among years. Differences between treatment and control segments were most likely due to habitat differences. Species were classified into guilds on the basis of foraging behavior and preferred breeding habitat. Few significant differences in abundance of birds within different guilds were found between treatment and control segments. Differences were most consistent for habitat categories (e.g., birds that prefer deciduous forest were more abundant on control segments in 4 of 5 months), providing further evidence that habitat differences were responsible for many of the observed differences in bird distribution patterns between treatment and control segments.