The northern parula is common in the bottomland hardwood and riverine forests of the southeastern United States (Moldenhauer and Regelski 1996). The species also occurs in mixed pine-hardwoods, although at lower densities (Moldenhauer and Regelski 1996). Northern parulas have two competing habitat requirements: a preference for canopy gaps and large forest blocks. Moorman and others (2001) found northern parulas more abundant near canopy gaps than forest-interior sites with an unbroken canopy in bottomland hardwoods, and Annand and Thompson (1997) observed highest northern parula densities in forests containing canopy gaps resulting from single-tree selection practices. However, the probability of detecting northern parulas increases with riparian buffer width (Kilgo and others 1998) and forest patch size (Robbins and others 1989).
Northern parulas forage in the mid- to upper canopy layers (Moldenhauer and Regelski 1996). Therefore, it is not surprising they prefer microsites with high basal area (Robbins and others 1989), high canopy cover and tall canopies (James 1971) and avoid areas with dense understories (often associated with open canopies; Torres and Leberg 1996). In the Southeast, northern parulas nest almost exclusively in Spanish moss (Moldenhauer and Regelski 1996). However, no studies have identified Spanish moss as limiting.
The northern parula habitat suitability model contains six parameters:
- successional age class
- forest patch size
- percent forest in a 1-km radius
- canopy cover
The first suitability function combines 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 the northern parula habitat associations reported in Hamel (1992) for the Southeast.
We derived a logarithmic function
from data on northern parula occupancy rate in forest blocks of varying size (SI2; Hayden and others 1985, Robbins and others 1989;
to predict habitat suitability from patch area. However, small forest patch in predominantly forested landscapes may provide habitat due to their proximity to large forest blocks (Rosenberg and others 1999). Thus, we included the percentage of forest in the local (1-km radius) landscape as a model parameter to adjust suitability index scores for these sites. We assumed landscapes (1-km radius window; SI3) with <30 percent forest provided poor habitat (suitability index score ≤ 0.100), and landscapes >70 percent forest were excellent habitat (suitability index score ≥ 0.900;
. We used the maximum suitability index score from either S2 or SI3 to account for small patches in predominantly forested landscapes.
We included canopy cover (SI4) in our model to capture the preference of northern parulas for interior edges. We assumed northern parula habitat suitability was low for sites with either extremely open or closed canopies and fit an inverse quadratic function
to data demonstrating this relationship
To calculate the overall suitability index score, we determined the geometric mean of SI scores for forest structure attributes (SI1 and SI4) and then calculated the geometric mean of this value and landscape composition (Max of SI2,SI3).
Overall SI = ((SI1 * SI4)0.500 * Max(SI2,SI3))0.500