The worm-eating warbler is a Neotropical migrant that breeds in forest interiors of the eastern United States (Hanners and Patton 1998). Minimum area requirements range from 21 ha in the mid-Atlantic (Robbins and others 1989) to >800 ha in Missouri (Wenny and others 1993). Worm-eating warblers nest on the ground along moderate to steep slopes (≥20º) containing dense (≥48 percent) shrub understories in mature deciduous and mixed deciduous-coniferous forests (Gale and others 1997). Both Artman and others (2001) and Blake (2005) found worm-eating warblers less abundant in recently burned stands due to the loss of leaf litter, a preferred nesting and foraging substrate. Canopy closure exceeded 95 percent in both Missouri (Wenny and others 1993) and Connecticut (Gale and others 1997).
The model for worm-eating warbler habitat suitability contains six parameters:
- successional age class
- forest patch size
- percent forest in the landscape, slope
- small (<2.5 cm d.b.h.) stem density
The first suitability function combines landform, landcover, and successional age class into a single matrix (SI1) that defines unique combinations of these classes
. We directly assigned habitat suitability scores to these combinations based on habitat associations reported in Hamel (1992)
We included slope (SI2) in our model because of the prevalence of steep slopes in the territories of worm-eating warblers. We defined slope classes based on data from Gale and others (1997) who identified the relative preference of various slopes for worm-eating warblers
We also included forest patch size (SI3) as a model parameter to account for the preference of worm-eating warblers for forest interiors. We fit a logistic function
to data from Robbins and others (1989) to quantify the relationship between forest patch size and habitat suitability
. The suitability of a forest patch is influenced not only by its size, though, but also its landscape context (SI4). In predominantly forested landscapes, small forest patch sizes not otherwise suitable may be used due to their proximity to a large forest block (Rosenberg and others 1999). We built a logistic function
based on the assumptions that 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;
. The maximum suitability index score of either SI3 or SI4 was assigned to each site to account for the higher suitability of small forest patches in heavily forested landscapes.
We relied on data from Wenny and others (1993) and Annand and Thompson (1997;
to quantify the relationship between habitat suitability scores and small stem density (SI5;
. We assumed worm-eating warblers occurred in forests with low stem densities, but these habitats had lower suitability scores than sites with well-developed understories characterized by dense stems.
To calculate the overall suitability index score, we determined the geometric mean of SI scores for forest structure (SI1 and SI5) and landscape composition (Max(SI2,SI3) and SI4) separately and then the geometric mean of these means together.
Overall SI = ((SI1 * SI5)0.500 * (Max(SI2, SI3) * SI4)0.500)0.500