Grassland and woodland bird occurrence and habitat selection in the prairie-forest transition zone of Minnesota

Numerous species of woodland and grassland nesting birds have experienced population declines in the midwestern United States. Habitat loss and habitat fragmentation are suspected to be two of the underlying causes for the population declines of woodland and grassland birds. The Minnesota Valley Wetland Management District (WMD) covers 14 counties in southeastern Minnesota in the transition zone between forested and prairie ecoregions. Little information is available relative to which birds are using grassland or woodland habitat within the WMD and how their presence or absence is related to the available habitat conditions. Therefore, my study objectives were to: (1) determine avian species composition within woodland and grassland habitats, (2) determine how selected grassland obligate birds and area sensitive woodland birds are associated with habitat variables that were measured on multiple spatial scales (landscape, patch, and local scales), and to (3) create spatially explicit models using a geographic information system for individual species and/or species assemblages (i.e., grassland obligates and area sensitive woodland birds that were associated with landscape scale variables. Woodland and grassland bird surveys were conducted from 1 June until 15 July during 2003 and 2004. Local scale variables were measured after completion of the point count, while patch and landscape variables were measured using a geographic information system (GIS) developed for the study area. Logistic regression models were used to determine associations between habitat variables and bird presence for selected grassland and woodland species. Linear regression models were used to determine associations between variables and bird species richness. All models were developed a priori and analyzed using an information-theoretic approach. During the study, 188 grassland patches and 178 woodland patches were surveyed. A total of 49 different bird species were detected during grassland counts and 66 different bird species were sampled during woodland counts. Common Yellowthroats (Geothlypis trichas), Song Sparrows (Melospiza melodia), Red-winged Blackbirds ( Agelaius phoeniceus), Sedge Wrens (Cistothorus platensis ), and Bobolinks (Dolichonyx oryzivorus) were present on the most grassland points, while Eastern Wood-pewees (Contopus virens), House Wrens (Troglodytes aedon), Red-eyed Vireos (Vireo olivaceus), Northern Cardinals (Cardinalis cardinalis), and Great Crested Flycatchers (Myiarchus crinitus ) were present on the most woodland points. Species richness of grassland obligates was positively associated with the amount of grass within 400 m of points, while negatively associated with the amount of patch edge and woody perimeter. Species richness of area sensitive woodland birds was positively associated with the amount of woodland within 1600 m of points and woodland patch size. Variables from each spatial scale were important for individual bird species models. My results indicate managers need to plan conservation actions on multiple spatial scales for the benefit of area sensitive species.