Recipient Organization
NORTHERN ARIZONA UNIVERSITY
(N/A)
FLAGSTAFF,AZ 86011
Performing Department
SCHOOL OF FORESTRY
Non Technical Summary
Understanding the winter ecology of bald eagles has been important in the protection and recovery of this species. Successful managment of a wintering population of bald eagles requires an understanding of the reasons for variability in foraging behavior and habitat use. Results form this study will help forest managers in developing forest management treatments that will retain or enhance habitat for bald eagles.
Animal Health Component
65%
Research Effort Categories
Basic
35%
Applied
65%
Developmental
(N/A)
Goals / Objectives
The objective of the study is to describe winter roosting habitat for bald eagles (Haliaeetus leucocephalus) on the Coconino and Kaibab National Forests and to assess the impacts of forest management on roosts. Specific objectives are: (1) To describe bald eagle night roosts and day perch sites. This shall include characterization of night roosts (frequency of use, tree species used, size, and other features) and areas around roosts and perches using on-the-ground measurements and those obtained using Geographic Information System (GIS; distance to water, roads; vegetation characteristics, vegetation type, patch size, etc.), and (2) To develop a predictive model that will enable resource managers to protect bald eagle roost and perch sites.
Project Methods
Research efforts will be concentrated in areas where bald eagles indicate night roosts in the Kaibab and Coconino National Forests on the Colorado Plateau (Figures 1 and 2). Dominant vegetation includes ponderosa pine (Pinus ponderosa) associations, with mixed conifer associations of Douglas-fir (Pseudotsuga menziesii) and white fir (Abies concolor) at higher elevations, and pinyon-juniper woodland associations (Pinus edulis-Juniperus spp.) at lower elevations. We will locate roost sites using GPS locations provided to us by AGFD. We will categorize night roosts versus day perches by time the location was collected and identify any roosts that used multiple times (e.g., locations used on different days that were < 15 m apart and where only single roost structures were available). We will randomly sample a subset of the potential roosts, prioritizing those roosts that were used >1 time. Searches will be conducted by walking around potential roost trees clockwise, then
counter-clockwise, 2- and 5-m away from the base of the tree or farther if necessary for tree lean. Sign will be removed from the roost tree prior to the next field season. For each identifiable roost, we will measure characteristics of roost trees and roost sites (the patch surrounding each roost). For roost trees, we will record: tree species, dbh, tree height, slope position, aspect, dominance class, tree position (e.g., isolation, interior), and tree vigor classification. We will also collect roost site information. For roost sites, we will sample tree density, height and dbh for trees >40 cm dbh, slope, aspect, canopy closure, number of stumps and shrub density at each site in 15 m radius plots (Buehler et al., 1991). If we cannot identify a specific roost tree (e.g., Figure 3A - there were 3 potential roost trees at this site, all with white wash), we will collect roost site information. In addition, elevation, area, distance to nearest water, distance to potential disturbance
by humans (e.g., roads, buildings, power lines) and distance to forest edge will be sampled using GIS. Random trees and sites will be used for comparison. We visited 4 roost sites to determine how difficult it would be to identify a specific roost and to test our methods for measuring roost sites. At 3 of the 4 roosts, we were able to identify potential roost structures (Figure 3A). We also flushed a mature bald eagle from one of these sites (Figure 3B), which confirmed use in this area. By conducting the work during winter, we are more likely to determine which trees are being used as roosts since bald eagles might be present. The study seems very feasible and there is potential for us to describe a large number of previously unknown wintering bald eagle roosts. Data will be analyzed by Multivariate Analysis Of Variance (MANOVA) using JMP (Statistical Analysis Software). Based on habitat characteristics that are determined to be important, a first approximation CART (Classification
and Regression Trees) model will be developed. By modeling we will estimate the probability that a given area will provide suitable bald eagle roosting habitat.