I examined the roosting and perching habitat preferences of a nonbreeding population of bald eagles in North Carolina during 1986 and 1987. I characterized roosting habitat at 2 scales; those of forest stands and individual roost trees. Eagles chose roost areas that were less dense, had less canopy cover, were closer to forest edges, and had larger trees than random forest areas (P < 0.05). Within roost areas eagles choose trees that were larger (height and dbh) than random trees. Additionally. eagles roosting at the Morgan Creek roost preferred dead hardwoods close to the forest edge and eagles at the Mason Point roost preferred trees farther from a frequently used dirt road within the roost. Suitable perch trees were the most important attribute of perching habitat. Eagles preferred loblolly pines and trees with leafless crowns (P < 0.05), which relates to their accessible crown structures. Perch trees were larger (height and dbh, P < 0.05) than adjacent trees along the shore. Eagles utilized the bottom of tree crowns during summer but used treetops during fall and winter. I found no evidence that eagles selected perches in relation to forest stand characteristics within 20 m of perch trees, forest cover types in 1 ha blocks surrounding perches, or habitat disturbances. Management recommendations include techniques to enhance bald eagle habitat on the study area. Primary emphasis should be toward managing for roosting habitat because of its apparent scarcity. Perch trees are plentiful but long-term management is desirable. Future nesting seems likely and management techniques for potential nesting habitat are suggested. / Master of Science

I studied the effect of human activities on bald eagle (Haliaeetus leucocephahus) distribution and abundance at Jordan Lake and Falls Lake, North Carolina in 1986 and 1981. Eagles used most of the area available on Jordan Lake, but 63% of the use occurred in the northern 25% of the lake. Eagle use at Falls Lake was restricted to a few areas in the northern section of the lake. Jordan Lake had 1.2 times as many eagle observations as did Falls Lake. Data from radio-tagged eagles and timing of population fluctuations suggest that eagle populations at Jordan and Falls Lakes were principally migrating eagles from southern states. The peak in eagle numbers in May 1981 may have represented a migratory wave, whereas the decrease in June and July may have been the result of some eagles continuing northward. Eagles returning south from the Chesapeake Bay and other northern areas may account for the slight increase observed in August. Two eagle roosts were located and monitored throughout the study at Jordan Lake. Human activities at both lakes peaked during summer months. Boating was the predominant activity during summer. Sixty-three intentional disturbances by motor boats produced a mean eagle flush distance of 131.2 m. Only 8% of the eagles flushed when the approaching boat was > 250 m from shore. Loglinear analysis revealed that human use of the shoreline and eagle use of the shoreline were related. Shoreline segments (250 m) used by humans were used less frequently by eagles than would be expected under a model of complete independence. I saw more eagles and fewer humans on weekdays than on weekends during boat surveys of selected Jordan Lake sections, suggesting that human use in certain sections on weekends displaced eagles. The lake section north of the Farrington Bridge showed the largest difference between eagle numbers on weekdays versus weekends. I developed a regression model that predicted the threshold density of disturbance Within this section to be 0.5 boats/km². On most days during the summer, this threshold level of boating traffic is surpassed in lake sections south of the Farrington Bridge. Primary management objectives should be to reduce human activities within high-eagle use areas, specifically the northern end of Jordan Lake, and to promote the bald eagle as a recreational benefit rather than a management problem. / Master of Science