Winter Microclimate of Bald Eagle Roosts on the Northern Chesapeake Bay

Abstract

From 11 November 1988 to 1 April 1989, we studied the microclimate and nightly energy budgets at three separate Bald Eagle (Haliaeetus leucocephalus) communal roost sites; one used in winter, one used in summer, and another used year-round. We compared the roost sites with the microclimate and nightly energy budgets at randomly selected inland forested sites and eagle shoreline perch sites on the northern Chesapeake Bay. Mean and minimum air temperatures were similar among all sites. Mean and maximum wind speeds were greater at the shore than at other sites. Wind speed did not differ between roosts and inland sites. Among roost sites, mean and maximum wind speeds were lowest at the winter roost. The year-round roost and summer-roost winds did not differ. Mean net radiation was greater at inland sites than at the shore sites, whereas mean net radiation of roost and inland and of roost and shore sites did not differ. Minimum net radiation was greatest at inland sites, whereas roost and shore minimum net radiation did not differ. We calculated that roosting eagles would have expended 205.9 kcal per night at traditional roost sites, 206.6 kcal per night at shoreline perches, and 203.7 kcal per night at inland sites. Calculated energy expended on the 10 coldest nights was similar among roost, shoreline, and inland sites. Adding the estimated cost of transport from shoreline perches to roosts (x = 5.3 kcal/round trip) did not produce significant differences in nightly energy expenditure between eagles roosting in communal roosts vs. those roosting on the shore.

Full text

American Ornithologists' Union
is collaborating with JSTOR to digitize, preserve and extend access to
The Auk.
http://www.jstor.org
Winter Microclimate of Bald Eagle Roosts on the Northern Chesapeake Bay
Author(s): David A. Buehler, Timothy J. Mersmann, James D. Fraser and Janis K. D. Seegar
Source:
The Auk,
Vol. 108, No. 3 (Jul., 1991), pp. 612-618
Published by: American Ornithologists' Union
Stable URL: http://www.jstor.org/stable/4088102
Accessed: 12-08-2015 14:14 UTC
REFERENCES
Linked references are available on JSTOR for this article:
http://www.jstor.org/stable/4088102?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.
This content downloaded from 160.36.178.25 on Wed, 12 Aug 2015 14:14:25 UTC
All use subject to JSTOR Terms and Conditions

WINTER
MICROCLIMATE
OF BALD EAGLE
ROOSTS ON THE
NORTHERN CHESAPEAKE
BAY
DAVID A. BUEHLER,1 TIMOTHY J. MERSMANN, 13 JAMES D. FRASER,' AND
JANIS K. D. SEEGAR2
'Department
of Fisheries
and Wildlife Sciences, Virginia
Polytechnic Institute and State University,
Blacksburg,
Virginia
24061 USA, and
2Chemical
Research, Development, and Engineering
Center, U.S. Army,
Aberdeen
Proving Ground,
Maryland 21010 USA
ABsTRAcr.-From 11 November 1988 to 1 April 1989, we studied the microclimate and
nightly energy budgets at three separate Bald Eagle (Haliaeetus leucocephalus)
communal roost
sites; one used in winter, one used in summer, and another used year-round. We compared
the roost sites with the microclimate and nightly energy budgets at randomly selected inland
forested sites and eagle shoreline perch sites on the northern Chesapeake Bay. Mean and
minimum air temperatures were similar among all sites. Mean and maximum wind speeds
were greater at the shore than at other sites. Wind speed did not differ between roosts and
inland sites. Among roost sites, mean and maximum wind speeds were lowest at the winter
roost. The year-round roost and summer-roost winds did not differ. Mean net radiation was
greater at inland sites than at the shore sites, whereas mean net radiation of roost and inland
and of roost and shore sites did not differ. Minimum net radiation was greatest at inland
sites, whereas roost and shore minimum net radiation did not differ. We calculated that
roosting eagles would have expended 205.9 kcal per night at traditional roost sites, 206.6 kcal
per night at shoreline perches, and 203.7 kcal per night at inland sites. Calculated energy
expended on the 10 coldest nights was similar among roost, shoreline, and inland sites.
Adding the estimated cost of transport from shoreline perches to roosts (x?
= 5.3 kcal/round
trip) did not produce significant differences in nightly energy expenditure between eagles
roosting in communal roosts vs. those roosting on the shore. Received 7 June 1990, accepted 10
January 1991.
PASSERINE
birds may experience thermal stress
during cold winter nights (King 1972), which
sometimes results in significant mortality (Odum
and Pitelka 1939). Consequently, specific pat-
terns of habitat use may reduce thermoregula-
tory costs (Walsberg 1983, 1985, 1986). In con-
trast, cold stress may be less important in the
survival of larger-bodied birds, including rap-
tors such as the Bald Eagle (Haliaeetus leucoce-
phalus;
Newton 1979:212). Hayes and Gessaman
(1980) were unable to induce cold stress in Red-
tailed Hawks (Buteo
jamaicensis)
and Golden Ea-
gles (Aquila chrysaetos) with microclimatic con-
ditions that exceeded average conditions at
many Bald Eagle wintering areas in the coter-
minous United States (- 17?C, 13.47 m/s wind,
and 0.0 W/m2 net radiation). Reports of cold-
induced mortality in Bald Eagles are anecdotal
and come from the northernmost extremes of
the eagle's range (Sherrod et al. 1976). More-
over, two previous studies differed on the im-
3Present address: U.S. Forest Service, Homochitto
National Forest, Gloster, Mississippi 39638 USA.
portance of roost-site selection to energy bud-
gets in eagles. Stalmaster and Gessaman (1984)
reported that eagles in Washington State saved
energy by roosting in protected inland conifers
rather than in deciduous trees adjacent to for-
aging areas. In contrast, Keister et al. (1985) con-
cluded that energy savings at communal roost
sites in the Klamath Basin, Oregon, did not off-
set the cost of flying > 10 km between roost sites
and foraging areas. In both cases, eagles in co-
niferous roost sites had lower nightly heat bud-
gets than if they roosted at more exposed for-
aging areas. The distance from the foraging area
to the roost determined whether eagles moving
to the roosts incurred a net energy gain or loss.
On the northern Chesapeake Bay, eagles
roosted only in deciduous trees and roosted close
(x = 0.18 km) to foraging areas (Buehler et al.
1991). Moreover, we suspected that thermal dif-
ferences between northern Chesapeake roosts
and western roosts exist because the northern
Chesapeake lacks conifers and substantial to-
pographic relief. We tested the hypothesis that
Bald Eagles on the northern Chesapeake Bay
selected roost sites that resulted in reduced en-
612 The Auk 108: 612-618. July 1991
This content downloaded from 160.36.178.25 on Wed, 12 Aug 2015 14:14:25 UTC
All use subject to JSTOR Terms and Conditions

July 1991] Winter
Microclimate
of Eagle Roosts 613
ergy expenditures by comparing the microcli-
mate and energy expended at roosts with the
microclimate and energy expended at shoreline
perch sites and randomly selected inland for-
ested sites.
STUDY AREA AND METHODS
Study area.-We studied Bald Eagle roosts on the
U.S. Army Aberdeen Proving Ground, a 350-km2 mil-
itary installation on the northern Chesapeake Bay,
north of Baltimore, Maryland. Study-area vegetation
consisted of mature coastal lowland oak-gum (Quercus
spp.-Liquidambar styraciflua) forests bordering the
Chesapeake Bay.
Roost, shoreline, and inland sites.-We located roost
sites during 1984-1989 by tracking radio-tagged ea-
gles until they roosted in the evening (Buehler et al.
1991). We studied the microclimate at a communal
roost used year-round (peak use = 30 eagles/night),
a communal-summer site (peak use = 35 eagles/night),
and a communal-winter site (peak use = 32 eagles/
night). All 3 roosts were used by eagles that foraged
on the Chesapeake Bay and Romney Creek, a shallow
tidal creek that flows into the bay. Year-round, sum-
mer, and winter roost sites were 2.89 km, 2.24 km,
and 1.00 km, respectively, from the Chesapeake Bay,
and 1.76 km, 0.82 km, and 1.59 km from their closest
access to Romney Creek. At each roost site, we mon-
itored 4 regularly used trees that were dispersed
throughout the site, to ensure sampling a represen-
tative array of microclimatic conditions. We identified
shoreline perch trees in the foraging areas by tracking
94 radio-tagged eagles from fixed-wing aircraft dur-
ing 1985-1988. We randomly selected 2 of 8 Chesa-
peake Bay shoreline perch trees and 2 of 12 Romney
Creek shoreline perch trees identified within 3 km
of all 3 roosts for microclimate monitoring. We mon-
itored the microclimate at 4 randomly selected, con-
tinuous-canopy inland forested sites within 3 km of
the foraging areas. These areas had trees within the
range of roost-tree heights (15.5-46.6 m).
Microclimate
monitoring.-We monitored wind speed,
temperature, and net long-wave radiation with 3-cup
anemometers (? 1.5%
accuracy, minimum wind speed
detection = 0.22 m/s, Campbell Scientific, Inc., Logan,
Utah), thermistor probes housed in 8- x 12-cm ther-
mal shields (?0.4?C accuracy, Campbell Scientific, Inc.,
Logan, Utah), and net radiometers (?10% accuracy,
Radiation Energy Measurement Systems, Seattle,
Washington). We mounted all sensors on a custom
pvc-pipe frame and suspended the frame from a tree
limb 20-25 m above the ground, within the range of
Bald Eagle roosting heights on the northern Chesa-
peake (Buehler et al. 1991). We used Campbell Sci-
entific CR10 modules to record sensor output every
6 s, to compute averages for input variables every 2
min, and to transfer the average values to a storage
module. We analyzed data collected between sunset
and sunrise.
Eagle nightly energy requirements.-We used a Bald
Eagle heat budget model (Keister et al. 1985) to es-
timate the metabolic heat production rate and total
energy expended by a roosting eagle for each sam-
pling night in each site.
We calculated the mean metabolic heat production
rate (M) in kcal/h over every 2-min period of each
night. Energy expended per 2-min period was cal-
culated by multiplying M by the time period (2/60
of an hour). We calculated the total energy expended
per night by summing all 2-min energy increments.
Statistical analyses.-We used the Chi-square ap-
proximation to the Kruskal-Wallis test (Hollander and
Wolfe 1973) to test for differences in microclimate
and heat budget parameters among roost, shoreline,
and inland sites, because most of the variables were
nonnormally distributed (Kolmogorov-Smirnov test,
P < 0.05). If the Kruskal-Wallis test was significant
(P < 0.05), we used Wilcoxon rank-sum tests for pair-
wise comparisons, with a = 0.05. We tested the night-
ly means (average of all 2-min means per night) for
temperature, wind speed, net radiation, metabolic rate,
and the total nightly energy expended. To determine
if sites differed in nightly microclimate extremes, we
compared the nightly minimum 2-min mean tem-
perature, the nightly maximum 2-min mean wind
speed, the nightly minimum 2-min mean net radia-
tion, and the nightly maximum 2-min mean metabolic
production rate among sites. To determine if variation
among nights obscured differences among sites, we
compared the microclimate at the roost, shoreline,
and inland sites that were monitored simultaneously
(n = 42 nights). To examine the similarity among
roosts, we compared the 3 roost sites. Because roosts
were not monitored simultaneously, we also com-
pared shoreline and inland sites as controls for these
sample periods. To test for effects of extreme nights,
we compared roost, shoreline, and inland sites on the
10 nights with the lowest mean shoreline tempera-
ture, the 10 nights with the greatest mean shoreline
wind speeds, the 10 nights with the lowest mean net
radiations, and the 10 nights with the greatest mean
metabolic rates along the shoreline.
To estimate the energy cost of flying to and from
roosts from the foraging areas, we used assumptions
identical to Stalmaster and Gessaman (1984); flight
speed = 45 km/h, energy for flapping flight = 12.5
x basal metabolic rate (BMR),
energy for soaring flight
= 3.5 x BMR, BMR = 12.47 kcal/h, and flight to and
from roosts = 50%
flapping and 50%
soaring. We add-
ed flight costs into the total nightly energy budgets
to compare energy expended at roosts vs. shoreline
perch sites.
RESULTS
MICROCLIMATE
All roost vs. all shoreline vs. all inland sites.-
Nightly temperature over the winter ranged
This content downloaded from 160.36.178.25 on Wed, 12 Aug 2015 14:14:25 UTC
All use subject to JSTOR Terms and Conditions

614 BUEHLER
Er AL. [Auk, Vol. 108
from -8.3?C to 20.0?C. Most nights (118/126,
93.7%)
had mean temperatures below the 10.60C
lower limit of the Bald Eagle thermal neutral
zone (Stalmaster and Gessaman 1984). Nightly
mean and nightly minimum roost, shoreline,
and inland-site temperatures did not differ (P
= 0.79, 0.64, respectively; Table 1). Mean night-
ly winds ranged from 0.2 m/s to 7.5 m/s and
were greatest at shoreline sites, whereas roost
and inland-site winds did not differ. Maximum
wind speeds also were greatest at shoreline sites,
whereas roost and inland-site maximum winds
did not differ. Mean nightly net radiation ranged
from -69.4 W/m2 to -4.1 W/m2 and was great-
er at inland sites than at shoreline sites. Roost
and shoreline minimum net radiation were less
than net radiation at inland sites.
Individual roost vs. shoreline vs. inland sites.
-
Mean and minimum nightly temperatures did
not differ among individual roosts and the si-
multaneously monitored shoreline and inland
sites (P = 0.78, 0.64, mean and minimum tem-
perature, respectively, year-round roost; P =
0.95, 0.91, summer roost; P = 0.89, 0.82, winter
roost; Table 1).
Shoreline mean and maximum winds were
greater than year-round roost and inland-site
winds (Table 1). Shoreline mean and maximum
winds were greater than inland-site mean and
maximum winds, but shoreline winds did not
differ from summer-roost mean and maximum
winds, and summer-roost mean and maximum
winds did not differ from inland-site winds.
Shoreline mean and maximum winds were
greater than winter-roost and inland-site mean
and maximum winds.
Mean nightly net radiation did not differ
among any individual roost, shoreline, and in-
land-site set (Table 1). Minimum nightly net
radiation also did not differ among the year-
round roost, shoreline, and inland sites. How-
ever, during monitoring of the summer roost,
the inland-site minimum net radiation was
greater (smaller negative value ) than summer-
roost and shoreline values. During monitoring
of the winter roost, the inland-site minimum
net radiation was greater than radiation at
shoreline sites, but inland and roost-site mini-
mum net radiation did not differ, and shoreline
and roost-site minimum net radiation did not
differ.
Comparisons
among roosts.-Mean and maxi-
mum nightly temperatures did not differ among
the 3 roosts (P = 0.29, 0.34, respectively; Table
1). The summer-roost mean and maximum winds
were greater than winter-roost winds, but did
not differ from year-round-roost winds. The
year-round-roost mean net radiation was great-
er than the mean net radiation at summer and
winter roosts. The year-round-roost minimum
net radiation was greater than the summer-roost
minimum net radiation, whereas the winter-
roost minimum net radiation did not differ from
minimum net radiation at either of the other
roosts.
Similar separate analyses of shoreline and in-
land sites showed no differences for the 3 sam-
pling periods for mean and minimum temper-
ature, mean and maximum wind speed, and
mean and minimum net radiation variables (P
> 0.05), except that mean net radiation differed
among the inland sites (P = 0.03).
EAGLE
HEAT BUDGETS
Roost, shoreline, and inland sites. -There were
no differences in mean or maximum metabolic
heat production rates or total energy among
roost, shoreline, and inland sites (Table 2). When
individual roosts were compared with shore-
line and inland sites monitored simultaneously,
mean and maximum metabolic production rates
and total energy did not differ. Similarly, when
roosts were compared with each other, mean
and maximum metabolic production rates and
total energy did not differ. We also detected no
differences in mean and maximum metabolic
rates and total energy expended among all roost,
shoreline, and inland sites on the extreme nights
of winter, including the coldest, windiest, low-
est net radiation, and greatest metabolic rate
nights (Table 3).
Flight cost effects.-At 2.2 kcal/km of flight,
eagles expended 7.7 kcal to fly round-trip be-
tween the closest foraging area (Romney Creek)
and the year-round roost, 3.6 kcal to fly round-
trip between Romney Creek and the summer
roost, and 4.4 kcal to fly round-trip between the
bay shoreline and the winter roost. After these
values were added into the total nightly energy
expended, mean nightly energy expended for
the year-round roost still was not different from
energy expended at shoreline sites (*?
= 206.6
kcal, 199.9 kcal, respectively, P = 0.28). Simi-
larly, when flight costs were included, summer-
roost nightly energy increased to 217.1 kcal, but
not significantly over the 212.9 kcal expended
at the shoreline site (P = 0.59). Winter-roost
This content downloaded from 160.36.178.25 on Wed, 12 Aug 2015 14:14:25 UTC
All use subject to JSTOR Terms and Conditions

TABLE 1. Mean and minimum values (x ? SE) of temperature and net radiation, and mean and maximum values (x ? SE) of wind speed at roost sites (n = 3),
shoreline perches (n = 4), and randomly selected inland forested sites (n = 4), northern Chesapeake Bay, Maryland, November 1988 to April 1989. Individual
roosts, shoreline, and inland sites were monitored simultaneously on 42 nights. Within-column section comparisons with similar letters were not different,
based on pair-wise Wilcoxon rank-sum tests (P > 0.05).
Minimum
Mean temperature temperature Mean wind Maximum wind Mean net radiation Minimum net radiation
(OC) (OC) (m/s) (m/s) (W/m2) (W/m2)
Year-round roostad 3.70 ? 0.65A 1.10 ? 0.67A 1.38 ? 0.14A 3.27 ? 0.20A -25.66 ? 2.05A -40.77 ? 2.41A
Shorelinea 4.12 ? 0.60A 1.88 ? 0.60A 1.98 ? 0.25B 4.51 ? 0.37B -29.48 ? 2.63A -43.14 ? 2.72A
Inlanda 3.87 ? 0.59A 1.62 ? 0.60A 1.34 ? 0.12A 3.07 ? 0.16A -24.82 ? 1.99A -36.90 ? 2.23A
Summer roostbd 2.36 ? 0.83A -0.13 ? 0.80A 1.50 ? 0.13AB 3.49 ? 0.21AB -33.13 ? 2.68A -47.63 ? 2.84A
Shorelineb 2.58 ? 0.82A 0.03 ? 0.78A 1.75 ? 0.15A 4.10 ? 0.26A -33.22 + 2.67A -46.79 ? 2.75A
Inlandb 2.31 ? 0.82A -0.25 ? 0.79A 1.15 ? O.lOB 2.87 ? 0.16B -26.90 ? 2.08A -38.32 ? 2.26B
Winter roostcd 2.86 ? 0.80A 0.24 ? 0.79A 1.11 ? 0.12A 2.83 ? 0.25A -34.39 ? 2.13A -46.17 ? 2.34AB
Shorelinec 3.10 ? 0.75A 0.69 ? 0.73A 1.62 ? 0.16B 3.84 ? 0.29B -36.94 ? 2.28A -49.62 ? 2.47A
Inlandc 2.93 ? 0.72A 0.43 ? 0.73A 1.20 ? 0.12A 2.72 ? 0.17A -31.90 ? 1.85A -42.59 ? 2.04B
All roostse 2.97 ? 0.44A 0.40 ? 0.44A 1.33 ? 0.08A 3.20 ? 0.13A -31.06 ? 1.36AB -44.86 ? 1.48A
All shorelinee 3.27 ? 0.42A 0.87 ? 0.41A 1.79 ? O.llB 4.15 ? 0.18B -33.22 ? 1.48A -46.51 ? 1.54A
All inland' 3.04 ? 0.41A 0.60 ? 0.41A 1.23 ? 0.07A 2.89 ? 0.1OA -27.87 ? 1.16B -39.27 ? 1.27B
Mean and maximum wind speed differed among the year-round roost, shoreline, and inland sites (Kruskal-Wallis tests, P = 0.03, 0.001, respectively).
Mean and maximum wind speed and minimum radiation differed among the summer roost, shoreline, and inland sites (Kruskal-Wallis tests, P = 0.01, 0.001, 0.001, respectively).
Mean and maximum wind speed and minimum radiation differed among the winter roost, shoreline, and inland sites (Kruskal-Wallis tests, P = 0.02, 0.001, 0.001, respectively).
d Mean and maximum wind speed and mean and minimum radiation differed among the year-round, summer, and winter-roost sites (Kruskal-Wallis tests, P = 0.02, 0.01, 0.01, 0.03, respectively). Winter-roost mean
and maximum winds were less than winds at the year-round and summer roosts; year-round-roost mean radiation was greater than radiation at summer and winter roosts; and year-round-roost minimum net radiation
was greater than at the summer roost (P < 0.05).
eMean and maximum wind speed and mean and minimum radiation differed among roost, shoreline, and inland sites (Kruskal-Wallis tests, P = 0.001, 0.001, 0.03, 0.001, respectively).
0"
This content downloaded from 160.36.178.25 on Wed, 12 Aug 2015 14:14:25 UTC
All use subject to JSTOR Terms and Conditions

616 BUEHLER
ET AL. [Auk, Vol. 108
TABLE
2. Mean and maximum values (x ? SE) of metabolic heat production rate and total nightly energy
expended for roost sites, shoreline perches, and randomly selected inland forested sites, northern Chesa-
peake Bay, Maryland, November 1988 to April 1989. Individual roost, shoreline, and inland sites were
monitored simultaneously on 42 nights.
Mean metabolic rate Maximum metabolic rate Total nightly energy
(kcal/h) (kcal/h) (kcal/night)
Year-round roosted 14.56 ? 0.31 15.84 ? 0.33 198.94 ? 4.53
Shoreline, 14.63 ? 0.30 15.84 ? 0.31 199.89 ? 4.37
Inlanda 14.47 ? 0.29 15.66 ? 0.31 197.61 ? 4.18
Summer roostbd 15.39 ? 0.40 16.69 ? 0.40 213.46 ? 6.75
Shorelineb 15.36 ? 0.39 16.68 ? 0.38 212.94 ? 6.51
Inlandb 15.10 ? 0.38 16.42 ? 0.38 209.28 ? 6.42
Winter roost'" 15.10 ? 0.36 16.35 ? 0.37 205.41 ? 6.11
Shorelinec 15.22 ? 0.34 16.45 ? 0.34 206.88 ? 5.91
Inlandc 15.02 ? 0.32 16.20 ? 0.33 204.28 ? 5.70
All roostse 15.02 ? 0.21 16.29 ? 0.21 205.94 ? 3.41
All shorelinee 15.07 ? 0.20 16.32 ? 0.20 206.57 ? 3.28
All inlande 14.86 ? 0.19 16.09 ? 0.20 203.72 ? 3.19
'Mean and maximum metabolic rates and total energy did not differ among year-round roost, shoreline, and inland sites (Kruskal-Wallis tests,
P = 0.88, 0.82, 0.91, respectively).
b Mean and maximum metabolic rates and total energy did not differ among summer roost, shoreline, and inland sites (Kruskal-Wallis tests, P
= 0.79, 0.81, 0.79, respectively).
' Mean and maximum metabolic rates and total energy did not differ among winter roost, shoreline, and inland sites (Kruskal-Wallis tests, P =
0.84, 0.70, 0.87, respectively).
dMean and maximum metabolic rates and total energy did not differ among year-round, summer, and winter roosts (Kruskal-Wallis tests, P =
0.18, 0.24, 0.18, respectively).
e Mean and maximum metabolic rates and total energy did not differ among roost, shoreline, and inland sites (Kruskal-Wallis tests, P = 0.66,
0.57, 0.69, respectively).
nightly energy plus flight costs equaled 209.8
kcal, not different from the 206.9 kcal expended
at the shoreline site (P = 0.67).
DISCUSSION
We do not believe that energy conservation
is an important factor in eagle roost-site selec-
tion on the Chesapeake Bay. Although Stal-
master and Gessaman (1984) reported signifi-
cant energy savings in western Washington,
Keister et al. (1985) reported energy savings at
Oregon roosts were negated by flight costs, and
we found no energy savings under any circum-
stances. There are several possible explanations
for these differences.
It is possible that topographic differences be-
tween the Chesapeake and the western Wash-
ington and Oregon study areas explain the ob-
served differences in energy savings. Greater
topographic variability in Washington and Or-
egon, compared with the relatively flat Chesa-
peake region, may provide eagles with a greater
range of microclimates from which to choose.
The presence of suitable stands of conifers for
roosting in the western areas contributes to mi-
croclimate variability as well. Energy expended
at foraging areas and roost sites may differ sig-
nificantly because of this greater range of mi-
croclimates.
Alternatively, the heat budget model may be
an inaccurate estimate of eagle heat budgets.
Some of the parameters estimated in the model
have been experimentally measured for other
avian species and agree fairly well with model-
derived values (Robinson et al. 1976, Mahoney
and King 1977, Hayes and Gessaman 1982). We
are unaware, however, of any field tests of the
entire model for large-bodied avian species.
Wind speed and net radiation were 2 microcli-
mate factors that varied between shoreline and
roost sites on the northern Chesapeake. Un-
derestimation of the effects of these parameters
could produce the observed results that suggest
that roost sites were not energetically favorable.
It also is possible that eagles select roosting
sites for microclimatic benefits accrued, not on
average winter nights but on catastrophically
stormy nights. The studies to date, which con-
centrated on average or "typical" conditions,
were unlikely to detect evidence for such be-
havior. To account for this effect, we examined
microclimate differences on the 10 most ex-
treme nights of winter but detected no differ-
ences in energy expended. Roost-site selection
under the catastrophic scenario may be more
This content downloaded from 160.36.178.25 on Wed, 12 Aug 2015 14:14:25 UTC
All use subject to JSTOR Terms and Conditions

July 1991] Winter Microclimate of Eagle Roosts 617
TABLE 3. Mean and maximum values (x ? SE) of metabolic heat production rate and total nightly energy
expended for roost sites, shoreline perches, and randomly selected inland forested sites on the 10 coldest
temperature, 10 windiest, 10 lowest net radiation and 10 greatest metabolic rate nights, northern Chesapeake
Bay, Maryland, November 1988 to April 1989.
Mean metabolic rate Maximum metabolic rate Total nightly energy
(kcal/h) (kcal/h) (kcal/night)
Coldest temperaturea
Roosts 19.48 ? 0.33 20.84 ? 0.44 276.03 ? 6.31
Shoreline 19.25 ? 0.31 20.52 ? 0.44 272.72 ? 5.46
Inland 19.05 ? 0.30 20.32 ? 0.45 269.95 ? 5.84
Windiestb
Roosts 15.17 ? 1.16 16.93 ? 1.20 198.59 ? 18.56
Shoreline 15.41 ? 1.10 17.17 ? 1.12 201.55 ? 17.95
Inland 15.08 ? 1.01 16.81 ? 1.06 197.14 ? 16.85
Lowest net radiationc
Roosts 17.29 ? 0.53 18.72 ? 0.59 233.64 ? 9.58
Shoreline 17.42 ? 0.53 18.78 ? 0.58 235.57 ? 9.84
Inland 17.05 ? 0.51 18.50 ? 0.57 230.42 ? 9.39
Greatest metab. ratesd
Roosts 19.42 ? 0.35 20.75 ? 0.48 275.10 ? 6.47
Shoreline 19.25 ? 0.31 20.49 ? 0.45 272.60 ? 5.49
Inland 18.99 ? 0.32 20.24 ? 0.47 268.99 ? 5.99
Mean and maximum metabolic rates and total energy did not differ among roost, shoreline, and inland sites on the 10 lowest temperature
nights (Kruskal-Wallis tests, P = 0.59, 0.63, 0.73, respectively).
bMean and maximum metabolic rates and total energy did not differ among roost, shoreline, and inland sites on the 10 windiest nights (Kruskal-
Wallis tests, P = 0.78, 0.87, 0.84, respectively).
c Mean and maximum metabolic rates and total energy did not differ among roost, shoreline, and inland sites on the 10 lowest net radiation
nights (Kruskal-Wallis tests, P = 0.59, 0.77, 0.79, respectively).
dMean and maximum metabolic rates and total energy did not differ among roost, shoreline, and inland sites on the 10 greatest metabolic rate
nights (Kruskal-Wallis tests, P = 0.64, 0.61, 0.63, respectively).
based on the selective advantages of avoiding
buffeting by strong winds and may not be ther-
moregulatory in nature. Our finding that win-
ter roosts afford greater protection from wind
than summer or year-round roosts is consistent
with this hypothesis. Steenhof et al. (1980) also
suggested that roost-site selection in the Mid-
west may occur to avoid buffeting by winds.
Given the variability of the evidence to date,
hypotheses that explain eagle roost-site selec-
tion in other than thermoregulatory terms may
be more plausible, at least for our study area. A
variety of hypotheses are offered to explain why
some birds roost communally, including pred-
ator avoidance (Lack 1968: 137) or information
transfer (Ward and Zahavi 1973). It is possible
that Bald Eagles select a communal roosting
habitat that facilitates or enhances the benefits
obtained under one or more of these hypoth-
eses.
ACKNOWLEDGMENTS
We thank B. A. Buehler, J. M. T. Cazell, A. K. DeLong,
D. C. DeLong Jr., D. W. Liedlich, M. Roeder, J. M.
Seegar, and T. L. Weller for assisting with fieldwork.
We thank M. R. Fuller, J. P. Ondek, W. S. Seegar, and
F. P. Ward for their help throughout the study. C. P.
Campbell computerized the data. We thank R. L. Kirk-
patrick, D. F. Stauffer, D. J. Orth and E. P. Smith, and
two anonymous referees for reviewing earlier ver-
sions of the manuscript. We thank the U.S. Army-
Chemical Research, Development, and Engineering
Center for funding this project.
LITERATURE CITED
BUEHLER,
D. A., T. J. MERSMANN, J. D. FRASER,
& J. K.
D. SEEGAR.
1991. Nonbreeding Bald Eagle com-
munal and solitary roosting behavior and habitat
use on the northern Chesapeake Bay. J. Wildl.
Manage. 55: 273-281.
HAYES, S. R., & J.
A. GESSAMAN. 1980. The combined
effects of air temperature, wind and radiation on
the resting metabolism of avian raptors. J.
Therm.
Biol. 5: 119-125.
, & . 1982. Prediction of raptor resting
metabolism: comparison of measured values with
statistical and biophysical estimates. J. Therm. Biol.
7: 45-50.
HOLLANDER, M., & D. A. WOLFE. 1973. Nonpara-
This content downloaded from 160.36.178.25 on Wed, 12 Aug 2015 14:14:25 UTC
All use subject to JSTOR Terms and Conditions

618 BUEHLER ET AL. [Auk, Vol. 108
metric statistical methods. New York, John Wiley
and Sons.
KEISTR,
G. P. JR., R. G. ANTHONY, & H. R. HOLBO.
1985. A model of energy consumption in Bald
Eagles: an evaluation of night communal roost-
ing. Wilson Bull. 97: 148-160.
KING, J. R. 1972. Adaptive periodic fat storage by
birds. Pp. 200-217 in Proc. 15th Int. Ornithol.
Congr. (K. H. Voous, Ed.). Leiden, Netherlands,
E. J. Brill.
LACK, D. L. 1968. Ecological adaptations for breed-
ing in birds. London, Methuen & Co., Ltd.
MAHONEY,
S. A., & J.
R. KING. 1977. The use of equiv-
alent black-body temperature in the thermal en-
ergetics of small birds. J. Therm. Biol. 2: 115-120.
NEWTON, I. 1979. Population ecology of raptors. Ver-
million, South Dakota, Buteo Books.
ODUM, E. P., & F. A. PITELKA. 1939. Storm mortality
in a winter starling roost. Auk 45: 451.
ROBINSON, D. E., G. S. CAMPBELL, & J. R. KING. 1976.
An evaluation of heat exchange in small birds. J.
Comp. Physiol. 105: 153-166.
SHERROD,
S. K., C. M. WHITE, & F. S. L. WILLIAMSON.
1976. Biology of the Bald Eagle on Amchitka
Island, Alaska. Living Bird 15: 143-182.
STALMASR,
M. V., & J. A. GEssAMAN.
1984. Ecolog-
ical energetics and foraging behavior of over-
wintering Bald Eagles. Ecol. Monogr. 54:407-428.
STEENHOF, K., S. S. BERLINGER,
& L. H. FREDRICKSON.
1980. Habitat use by wintering Bald Eagles in
South Dakota. J. Wildl. Manage. 44: 798-805.
WALSBERG,
G. E. 1983. Avian ecological energetics.
Pp. 161-220 in Avian biology, vol.7 (D. S. Farner,
J. R. King, and K. C. Parkes, Eds.). New York,
Academic Press.
. 1985. Physiological consequences of micro-
habitat selection. Pp. 389-413 in Habitat selection
in birds (M. L. Cody, Ed.). New York, Academic
Press.
. 1986. Thermal consequences of roost-site
selection: the relative importance of three modes
of heat conservation. Auk 103: 1-7.
WARD, P., & A. ZAHAVI. 1973. The importance of
certain assemblages of birds as "information-cen-
tres" for food-finding. Ibis 115: 517-534.
This content downloaded from 160.36.178.25 on Wed, 12 Aug 2015 14:14:25 UTC
All use subject to JSTOR Terms and Conditions