Abstract:
The 2.5 percent annual decline of the Golden-winged Warbler (Vermivora chrysoptera) over the
past 39 years has concerned staff from various wildlife agencies and conservation organizations,
prompting increased research to better identify and protect habitat for this neo-tropical migratory
warbler. To aid the North Carolina Audubon Society in identifying lands for further research or
survey efforts, the objectives of this project included: 1) Using a combination of geographic
information systems (GIS) spatial analysis and three statistical modeling approaches – a
Classification and Regression Tree (CART), a Generalized Linear Model (GLM), and MaxEnt –
to predict suitable habitat for the Golden-winged Warbler across the study area; 2) Combining
the results of all three modeling techniques to generate a conservative estimate of predicted
habitat; 3) Mapping the results in a GIS, and; 4) Refining and prioritizing the predicted habitat.
The results of this statistical analysis suggested that the variables of elevation, distance from
developed open space, and the focal mean value of Montane Oak forest may be the most useful
in predicting habitat for the Golden-winged Warbler in western North Carolina. To refine the
raw, three model combined output, patches less than 10 hectares were omitted. Also, to
prioritize patches, three scenarios involving lands managed for conservation and open space and
North Carolina Audubon Important Bird Areas (IBAs) were compared by area, number of
patches, and county in which the predicted habitat occurred. Finally, several factors may limit
the practical use of this modeling effort. Thus, the mapped predicted habitat should only be used
to guide further research and is not intended as a basis for management recommendations.
