The prairie warbler breeds in shrubby vegetation under an open canopy (Nolan and others 1999). Typical associations in the CH and WGCP include shrubby southern pine forest, pine barrens, scrub oak barrens, abandoned fields and pastures, regenerating forest, abandoned orchards, grassland-forest edge, Christmas tree farms, and reclaimed strip mine spoils. Prairie warblers use a variety of landforms from xeric uplands in Arkansas to palustrine swamps in Virginia. In contrast to other early successional warblers, prairie warblers occupy sites with fewer dense shrubs than blue-winged warblers, more dense vegetation and drier areas than yellow warblers, and less dense vegetation and higher vegetation strata than common yellowthroats or yellow-breasted chats (Nolan and others 1999).
Prairie warblers nest in shrubs and small trees >20 m from a field-forest edge (Nolan and others 1999, Woodward and others 2001). However, prairie warblers in eastern Texas are typically found in narrow riparian zones, with numbers decreasing quickly at widths increase (Conner and others 2004). Mean territory size varies inversely with population density, ranging from 0.2–3.5 ha in Indiana (Nolan and others 1999). Territory size also varies with forest patch shape, being larger in more linear patches. Although males do not limit their movements to their defended territory, a female’s home range is usually contained within a male’s defended territory. The species is a cowbird host; and though parasitism has little effect on hatching success, it can significantly reduce fledging rates.
Our model for prairie warbler habitat suitability contains seven parameters:
- successional age class
- forest patch size
- small (<2.5 cm d.b.h.) stem density
- edge occurrence
- canopy cover
The first suitability function combined landform, landcover, and successional age class into a single matrix (SI1) to define unique combinations of these classes
. We directly assigned suitability index scores to these combinations based on prairie warbler habitat associations documented in Hamel (1992).
Both Woodward and others (2001) and Rodewald and Vitz (2005) observed edge avoidance by prairie warblers. Thus we used a 3 × 3 window to identify early successional habitats (i.e., grass-forb, shrub-seedling, or sapling successional age class forest) adjacent to mature forest stands (i.e., pole or sawtimber successional age class) and reduced the suitability of locations adjacent to edges by half (SI2;
We also included early successional forest patch size (SI3) as an explanatory variable in our model because prairie warblers do not occur in small clearings or edge habitats. We used data compiled by Larson and others (2003;
to fit a logistic function
that characterized the relationship between habitat suitability and early successional patch size.
We also included small stem density (SI4) as a model parameter because prairie warblers are associated with dense understory vegetation. We used the point count and habitat data reported by Annand and Thompson (1997;
to derive a logistic function
that predicted habitat suitability for prairie warblers from small stem density.
Lastly, we used data from Sheffield (1981) to inform an inverse logistic function
that discounted habitat suitability index scores at increasingly high canopy closures (SI5;
To calculate the overall suitability index score, we determined the geometric mean of SI scores for forest structure attributes (SI1, SI4, and SI5) and landscape composition (SI2 and SI3) separately and then the geometric mean of these means together.
Overall SI = ((SI1 * SI4 * SI5)0.333 * (SI2 * SI3)0.500)0.500