|
Wildlife habitat relationships in forested landscapes are commonly based on
associations between wildlife species and selected forest types and age classes
(e.g., pole-sized paper birch). Using data from 360 point-count locations, we
tested simple habitat-relationship models for 53 species of forest-dwelling
birds in the White Mountain National Forest in New Hampshire and Maine.
Misclassification was common, attributed to both omission and commission. We
then hypothesized that many bird species respond to more general features of
the landscape, which led us to evaluate spatially-explicit habitat models. One
set of models tests Ecological Land Types (ELTs) as predictors of avian
occurrence; ELTs describe relatively contiguous areas with similar soil
characteristics, topographical relief, and potential vegetation. A final set
of habitat models reflects coarse-scale landscape charactersitics (e.g.,
configuration and patchiness) and is based on patterns of land cover as
determined by remote sensing. We find that avian species respond to the
landscape at different scales and that traditional habitat-relationship models
are often misleading.
|