Investigating Michigan prairie warbler breeding territory habitat composition, within-territory resource use, and factors influencing occupancy

Recent Prairie Warbler (Dendroica discolor) observations in Michigan are almost exclusively from Lake Michigan early-successional dune ecosystems in the Lower Peninsula. However, data describing breeding locations, population abundance, and habitat composition influencing breeding territory selection are largely unknown and thus critical information needs for this state-endangered, Neotropical migratory songbird. Understanding Prairie Warbler habitat selection during the breeding season will better inform conservation and management planning. Avian point counts were conducted in 9 study sites along the Lake Michigan coast in 2004 and 2005 to locate Prairie Warblers. Breeding territories were mapped using GPS and GIS. Plant composition was quantified within breeding territories and in unoccupied areas to determine whether geographic variation existed in vegetation associated with Prairie Warbler breeding areas from north to south. Logistic regression was used to develop a predictive model of Prairie Warbler occurrence. Using 1 m resolution imagery, distances from edge of territories and unoccupied areas to 7 environmental covariates were used as variables to forecast Prairie Warbler occupancy in the west Michigan dune ecosystems. Resource utilization functions (RUF) were developed to explore resource use by Prairie Warblers within highest-use areas of their breeding territories, thus relating geographic space use to environmental attributes as portrayed by 30 m land cover data. Comparisons were made across all breeding territories and for southern and northern sites exclusively to examine the potential for a north-to-south gradient in resource use. Thirty-eight Prairie Warblers were detected, with high detection probability, in 5 of the 9 sites surveyed. Breeding areas tended to be clustered within sites, with breeding territory boundaries often adjacent to each other, suggesting potential for conspecific attraction. Results generally support previously observed relationships between Prairie Warbler occurrence and a mixture of herbaceous and low-growing woody vegetation. However, my study was the first to document a latitudinal gradient in Prairie Warbler habitat use related to vegetation composition; with northern territory space use correlated with denser, woody vegetation and southern territory space use associated with more open, grass dominated sites. Plant composition in areas occupied by Prairie Warblers was different from plant composition in unoccupied areas. However, those differences were not consistent from north to south. The best indicators of Prairie Warbler occupancy were proximity to other Prairie Warblers and proximity to shrub cover. I found an 80% probability of occupancy when another Prairie Warbler was within 91 m of an existing territory. I found a 39% probability of occupancy when shrubs were within 1 m. There were no significant environmental covariates from RUFs, however, four observations warrant attention: (1) northern sites have more woody vegetation cover than southern sites, and southern sites seem to be more open, (2) grass cover appears to be important on southern sites, potentially because shrubs are less ubiquitous, (3) higher soil moisture in northern territories was suggested as an important indicator of habitat use, and (4) lack of significant RUF coefficients suggest that the spatial resolution of land cover data used (30 m) may not have been appropriate to describe within territory space use for Prairie Warblers. Key among management recommendations, conservation professionals are encouraged to maintain existing, large areas of contiguous habitat with well dispersed shrubs to accommodate multiple breeding territories and thus maintain coastal Prairie Warbler population viability.