The Kentucky warbler is a long-distance migrant that breeds in mature moist deciduous forests of the Southeast. This species is considered a forest interior specialist, mainly because of its low productivity and survival in edge and early successional habitats (Morse and Robinson 1999, Robinson and Robinson 2001). Kentucky warblers occupy fragments as small as 2.4 ha (Blake and Karr 1987), but tracts >500 ha are considered the minimum size necessary to support sustainable populations (McDonald 1998). A dense understory is a common feature of nesting sites. Ground cover averaged 46 percent in Kentucky warbler territories in Missouri (Wenny and others 1993), and vegetation <1.5 m was denser around nests than random sites in South Carolina (Kilgo and others 1996). Dense vegetation 0.3–1 m was also associated with higher numbers of Kentucky warblers in Maryland (Robbins and others 1989). Mesic sites are universally selected (McShea and others 1995, McDonald 1998, Gram and others 2003).
The habitat suitability model for Kentucky warblers contains five parameters:
- successional age class
- small (<2.5 cm d.b.h.) stem density
- forest patch size
- percent forest in the landscape.
The first suitability function combines landform, landcover, and successional age class into a single matrix (SI1) defining unique combinations of these classes
. We relied on relative habitat quality associations reported by Hamel to assign suitability index scores to these combinations. However, we increased suitability index scores for shrub-seedling stands based on data from Thompson and others (1992).
Kentucky warblers nest at the base of shrubs and occur in habitats containing high small stem densities (SI2). We used Kentucky warbler relative abundance data from Wenny and others (1993), Kilgo and others (1996), and Annand and Thompson (1997) to derive a logistic function
that predicts habitat suitability from small stem density
We used a logarithmic function
to quantify the relationship between forest patch size (SI3) and habitat suitability based on observations by Hayden and others (1985) and Robbins and others (1989;
However, the suitability of a specific forest patch is also influenced by its landscape context (SI5). Because Kentucky warblers are particularly sensitive to fragmentation (Lynch and Whigham 1984), we utilized a 10-km window to characterize the landscape. We assumed landscapes with <30 percent forest were poor habitat (suitability index score≤ 0.100), and landscapes with >70 percent forest were excellent habitat (suitability index score ≥ 0.900;
. We fit a logistic function
through a dataset reflecting these assumptions to predict how habitat suitability varied with landscape composition.
To calculate the overall suitability index score, we determined the geometric mean of suitability index scores for functions relating to forest structure (SI1 and SI2) and landscape composition (SI3 and SI4) separately and then the geometric mean of these means together.
Overall SI = ((SI1 * SI2)0.500 * (SI3 * SI4)0.500)0.500