As its name implies, the Louisiana waterthrush is associated with water throughout its range (Robinson 1995). Louisiana waterthrushes reach their highest densities along gravel-bottomed first- and second-order streams flowing through large (>350 ha) tracts of mature deciduous forest (Robbins and others 1989, Robinson 1995). Birds also breed at lower densities along mud-bottomed streams in cypress swamps and bottomland hardwood forests (Hamel 1992, Robinson 1995).
Prosser and Brooks (1998) developed and validated a habitat suitability model for Louisiana waterthrushes in central Pennsylvania. The model contained eight variables: canopy cover (>80 percent considered ideal), shrub cover (<25 percent), ratio of deciduous to conifer cover (30–69 percent, mostly reflecting hemlock dominance along streams in the Northeast), herbaceous cover (<25 percent), stream order (1st or 2nd with well-developed pools and riffles), water clarity and substrate (clear and rocky or sandy), nesting cover (uprooted trees or creviced, steep banks), and forest area (>350 ha).
Our Louisiana waterthrush habitat suitability index model contained eight parameters:
- successional age class
- distance to stream
- canopy coverv
- small (<2.5cm d.b.h.) stem density
- forest patch size
- percent forest in a 1-km radius.
The first suitability function combines landform, landcover, and successional age class into a single matrix (SI1) defining unique combinations of these classes
. We directly assigned suitability index scores to these combinations based on the vegetation and successional age class associations of Louisiana waterthrush outlined in Hamel (1992).
We included distance to stream (SI2) as a model variable because waterthrushes use streams and creeks for foraging and nesting. Louisiana waterthrushes restrict their foraging to the stream bed and bank, so we assumed a sharp decline in suitability with increasing distance to a stream
. We used an inverse logistic function to characterize this relationship
We also included canopy cover (SI3) and small stem density (SI4) as model variables, based on the preference of this species for mature forested sites with closed canopies and open understories. We fit logistic
and inverse logistic
functions to data adapted from the habitat suitability model of Prosser and Brooks (1998) for canopy cover
and small stem density
Forest patch size (SI5) affects the occupancy of habitats by Louisiana waterthrush. To predict the habitat suitability from forest patch size, we fit a logarithmic function
to data from Hayden and others (1985) and Robbins and others (1989;
on the detection probabilities of Louisiana waterthrush in forest patches of varying size. Forest patch size alone may not be an appropriate measure of a site’s suitability, though. In predominantly forested landscapes, small forest patches not otherwise suitable may be occupied due to their proximity to large forest blocks (Rosenberg and others 1999). We fit a logistic function
to a dataset based on the assumptions that landscapes 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 S5 or SI6 to ensure small forest blocks in predominantly forested landscapes were assigned an appropriate suitability score.
To calculate the overall suitability index, we determined the geometric mean of SI scores for forest structure (SI1, SI3, and SI4) and landscape composition (Max(SI5, SI6) and SI2) separately and then the geometric mean of these means together.
Overall SI = ((SI1 * SI3 * SI4)0.333 * (Max (SI5, SI6) * SI2)0.500)0.500