Oral Session #39: Conservation: Landscapes and Reserve Design.
Presiding: D. Peters
Tuesday, August 6. 1:00 PM to 4:45 PM. Graham Meeting Room, TCC.

Geographic boundary analysis of black-throated blue warbler territory distributions in heavily- and less-browsed northern forests.

Hall, Kimberly*,1, 1 University of Michigan, Ann Arbor, MI

ABSTRACT- Our best opportunity for conserving migratory songbirds in northern forests is to include them in landscape-scale management plans. This requires predicting species distributions across landscapes, which can vary in forest type, management history, and browse pressure from deer. In Michigan's hardwood forests, overstory composition and management history are typically used to group forest stands. However, for shrub-nesting species like black-throated blue warblers (Dendroica caerulescens), heterogeneity in the height, species composition, and density of understory vegetation can strongly influence distributions within stands of the same "type." I used spatially agglomerative clustering, a form of geographic boundary analysis, to describe heterogeneity in forest understory in two heavily-browsed and two less-browsed hardwood stands. This tool delineates boundaries in multivariate, spatial data by grouping locations based on proximity and similarity of variable values. In heavily-browsed forests, territories were concentrated where dense patches of hardwood seedlings (kept at 0.5-0.75 m tall by browsing) were adjacent to clusters dominated by browse resistant conifers, suggesting that birds selected a combination of two understory types. In less-browsed forests, patches of dense hardwood saplings 1-3 m tall were the most often occupied by warblers. If applied to leaf-off remotely-sensed images, spatially agglomerative clustering could be used at larger scales to greatly improve prediction of shrub-nesting bird distributions in forests where browse-resistant understory conifers provide "refuge" habitat.

KEY WORDS: Dendroica caerulescens, spatial analysis