A RADAR AND VISUAL STUDY OF NOCTURNAL BIRD AND BAT MIGRATION AT THE PROPOSED PRATTSBURGH-ITALY WIND POWER PROJECT, NEW YORK, FALL 2004

Full text

FINAL REPORT
A RADAR AND VISUAL STUDY OF NOCTURNAL BIRD AND BAT
MIGRATION AT THE PROPOSED PRATTSBURGH–ITALY WIND
POWER PROJECT, NEW YORK, FALL 2004
TODD J. MABEE
JONATHAN H. PLISSNER
BRIAN A. COOPER
PREPARED FOR
ECOGEN LLC
WEST SENECA, NEW YORK
PREPARED BY
ABR, INC.
FOREST GROVE, OREGON

Printed on recycled paper.
A RADAR AND VISUAL STUDY OF NOCTURNAL BIRD AND BAT
MIGRATION AT THE PROPOSED PRATTSBURGH–ITALY WIND
POWER PROJECT, NEW YORK, FALL 2004
FINAL REPORT
Prepared for
Ecogen LLC
950-A Union Rd, Suite 20
West Seneca, New York 14224
Prepared by
Todd J. Mabee
Jonathan H. Plissner
Brian A. Cooper
ABR, Inc.—Environmental Research & Services
P.O. Box 249, Forest Grove, Oregon 97116
March 2005

iEcogen Nocturnal Bird and Bat Migration, 2004
EXECUTIVE SUMMARY
• This report presents the results of a radar and
visual study of bird and bat migration
conducted during 14 August–29 September
2004 at the proposed Prattsburgh–Italy Wind
Power project, located in Yates and Steuben
counties, west-central New York. Radar and
visual observations were conducted for ~6.5
h/night during 45 nights during the fall.
• The primary goal of this study was to collect
information on the migration characteristics of
nocturnally migrating birds, especially
passerines, during the fall-migration period;
the secondary goal was to assess the extent of
use of the area by bats to provide an overall
assessment of potential project-related impacts
to birds and bats. Specifically, the objectives of
this study were to: (1) collect baseline
information on migration characteristics (i.e.,
flight direction, migration passage rates, flight
altitudes) of nocturnally migrating birds and
bats; (2) visually estimate the relative
proportions of birds and bats within the
potential rotor-swept area of the proposed wind
turbines; and (3) determine the number of birds
and bats that would pass within the rotor-swept
area of the proposed wind turbines during the
migratory season.
• In the fall, the mean flight direction of targets
observed on radar was 177°.
• The mean nocturnal passage rate for the fall
season was 200 ± 12 targets/km/h and ranged
among nights between 18 and 863
targets/km/h. Fall passage rates varied among
hours of the night, with lowest mean rates
occurring during the earliest hour of the
evening.
• The mean nocturnal flight altitude for the
entire fall season was 365 ± 3 m agl. Mean
flight altitudes observed on vertical radar were
highly variable among nights and ranged from
202 to 584 m agl. Neither the mean flight
altitude nor the altitudinal distribution of
targets varied among hours within a night.
Nine percent of all targets during fall 2004
were below the maximal height of the
proposed wind turbines (125 m).
• Migration passage rates increased with
tailwinds, calm conditions, easterly
crosswinds, and westerly crosswinds and
decreased with wind speed. Flight altitudes
increased with tailwinds, calm conditions,
westerly crosswinds, and date.
• Assuming an average of 10 nocturnal h/d and
45 d in the fall study, we estimated a turbine
passage rate of 51–362 nocturnal songbird/bat
migrants passing within the area occupied by
each proposed turbine during fall 2004.
• We developed visual sampling methods to
investigate low-altitude migration of birds and
bats. During 14 August–13 September, we
sampled with two, 2,000,000-Cp spotlights
with red lenses and were able to identify as
birds or bats 70% of all targets (n = 20)
occurring within ~75 m agl. During 14–29
September, we sampled with both night-vision
goggles and spotlights and were able to
identify 75% of all targets (n = 106) occurring
~<150 m. Because of our shorter range to
detect birds and bats with only spotlights, we
only used data collected with night-vision
goggles to calculate the proportion of birds and
bats below ~<150 m agl (94% birds 6% bats; n
= 80).
• The key results of our of fall passerine and bat
migration study were: (1) the mean overall
passage rate was moderate (i.e., 200
targets/km/h); (2) mean nightly passage rates
ranged from 18 to 863 targets/km/h; (3) the
percentage of targets passing below 125 m agl
(9%) was similar to that for a small number of
comparable studies; (4) an estimated turbine
passage rate of 51–362 nocturnal migrants
passing within the airspace occupied by each
proposed turbine during the 45-d fall migration
season; and (5) migrants flying below 150 m
agl consisted of ~94% birds and ~6% bats
during the late sampling period (i.e., mid- to
late September).

iii Ecogen Nocturnal Bird and Bat Migration, 2004
TABLE OF CONTENTS
EXECUTIVE SUMMARY ............................................................................................................................ i
LIST OF FIGURES ......................................................................................................................................iii
LIST OF TABLES........................................................................................................................................ iv
LIST OF APPENDICES............................................................................................................................... iv
ACKNOWLEDGMENTS ............................................................................................................................iv
INTRODUCTION ......................................................................................................................................... 1
OBJECTIVES................................................................................................................................................ 1
STUDY AREA .............................................................................................................................................. 1
METHODS .................................................................................................................................................... 3
STUDY DESIGN........................................................................................................................................ 3
RADAR EQUIPMENT............................................................................................................................... 3
DATA COLLECTION ............................................................................................................................... 5
TARGET IDENTIFICATION ON RADAR ........................................................................................... 5
SAMPLING DESIGN.............................................................................................................................. 5
VISUAL OBSERVATIONS OF LOW-ALTITUDE BIRDS AND BATS.............................................6
DATA ANALYSES.................................................................................................................................... 6
RADAR DATA........................................................................................................................................ 6
THE EFFECTS OF WEATHER ON MIGRATION PASSAGE RATES AND FLIGHT ALTITUDES7
RESULTS ...................................................................................................................................................... 7
FLIGHT DIRECTION................................................................................................................................ 7
PASSAGE RATES ..................................................................................................................................... 8
FLIGHT ALTITUDES ...............................................................................................................................8
INTENSIVE SAMPLING PERIODS......................................................................................................... 8
EFFECTS OF WEATHER ON MIGRATION......................................................................................... 12
PASSAGE RATES................................................................................................................................. 12
FLIGHT ALTITUDES........................................................................................................................... 14
TARGETS WITHIN THE PROPOSED TURBINE AREA..................................................................... 14
VISUAL DATA........................................................................................................................................ 16
DISCUSSION.............................................................................................................................................. 16
TIMING OF MIGRATION ...................................................................................................................... 16
PASSAGE RATES ................................................................................................................................... 17
FLIGHT ALTITUDES .............................................................................................................................18
INTENSIVE SAMPLING PERIODS....................................................................................................... 19
MODELING MIGRATION PASSAGE RATES AND FLIGHT ALTITUDES .................................... 19
MIGRATION PASSAGE RATES.........................................................................................................19
FLIGHT ALTITUDES........................................................................................................................... 20
SPECIES COMPOSITION....................................................................................................................... 20
TARGETS WITHIN THE PROPOSED TURBINE AREA..................................................................... 20
CONCLUSIONS....................................................................................................................................... 21
LITERATURE CITED................................................................................................................................ 21
LIST OF FIGURES
Figure 1. Map of the proposed Prattsburgh–Italy Wind Power project in Yates and Steuben
Counties, New York ................................................................................................................. 2
Figure 2. Approximate airspace sampled by Furuno FR–1510 marine radar when operating
in the surveillance mode as determined by field trials with Rock Pigeons. ............................. 4

Ecogen Nocturnal Bird and Bat Migration, 2004 iv
Figure 3. Approximate airspace sampled by Furuno FR–1510 marine radar when operating
in the vertical mode as determined by field trials with Rock Pigeons......................................4
Figure 4. Flight directions of radar targets at the proposed Prattsburgh–Italy Wind Power
project, New York fall 2004 .....................................................................................................8
Figure 5. Mean ± 1 SE nightly passage rates at the proposed Prattsburgh–Italy Wind Power
project, New York, fall 2004 ....................................................................................................9
Figure 6. Percent of seasonal passage rate by hour of the night at the proposed Prattsburgh-
Italy Wind Power project, New York, fall 2004....................................................................... 9
Figure 7. Mean ± 1 SE nightly flight altitudes at the proposed Prattsburgh–Italy Wind Power
project, New York, fall 2004 .................................................................................................. 10
Figure 8. Mean ± 1 SE flight altitude by hour of the night at the proposed Prattsburgh–Italy
Wind Power project, New York, fall 2004............................................................................. 10
Figure 9. Percent of radar targets at each altitude at the proposed Prattsburgh–Italy Wind Power
project, New York, fall 2004, by hour of the night ................................................................11
Figure 10. Variation in passage rate and flight altitude counts by length of sampling session at
the proposed Prattsburgh–Italy Wind Power project, New York, fall 2004........................... 12
Figure 11. Mean number of birds or bats/h observed during visual sampling at the proposed
Prattsburgh–Italy Wind Power project, New York, fall 2004 ................................................17
LIST OF TABLES
Table 1. Nocturnal flight altitudes of radar targets detected at the 1.5-km range at the
Prattsburgh–Italy Wind Power project, NY, fall 2004, by flight-altitude category ...............11
Table 2. Linear-regression models explaining the influence of environmental factors on
migration passage rates of bird and bat targets on surveillance radar at the
Prattsburgh–Italy Wind Power project, NY, fall 2004 ...........................................................13
Table 3. Parameter estimates from the two best-approximating models explaining the
influence of environmental factors on passage rates of bird and bat targets at the
Prattsburgh–Italy Wind Power project, NY, fall 2004 ...........................................................14
Table 4. Linear-regression models explaining the influence of environmental factors on mean
flight altitudes of bird and bat targets on vertical radar at the Prattsburgh–Italy
Wind Power project, NY, fall 2004........................................................................................ 15
Table 5. Parameter estimates from the two best-approximating models explaining the
influence of environmental factors on mean flight altitudes of radar targets at the
Prattsburgh–Italy Wind Power project, NY, fall 2004 ...........................................................16
LIST OF APPENDICES
Appendix 1. Calculation of the turbine passage rate over the entire 45-day fall 2004 study
period at the Prattsburgh–Italy Wind Power project, NY...............................................26
ACKNOWLEDGMENTS
We thank Ecogen for funding this study and Thomas Hagner for helping with field logistics. We thank
Bill Evans for input on study design and for help with the selection of a study site. At ABR, we thank Peter
Sanzenbacher and Corey Grinnell for help with radar sampling, Corey Grinnell for help with report
preparation, and Rich Blaha for figure preparation.

Introduction
1Ecogen Nocturnal Bird and Bat Migration, 2004
INTRODUCTION
Avian collisions with communication towers
have been recorded in North America since 1948
(Kerlinger 2000), with neotropical migratory birds
such as thrushes (Turdidae), vireos (Vireonidae),
and warblers (Parulidae) seeming to be the most
vulnerable to tower collisions during their
nocturnal migrations (Manville 2000). Passerines
also collide with wind turbines (Osborn et al. 2000,
Erickson et al. 2001, 2002), composing >80% of
the fatalities at wind power developments; ~50%
of the fatalities at windfarms involve nocturnal
migrants (Erickson et al. 2001). Studies examining
the impacts of windfarms on birds in the US and
Europe suggest that fatalities and behavioral
modifications (e.g., avoidance of windfarms) occur
in some, but not all, locations (Winkelman 1995,
Anderson et al. 1999, Erickson et al. 2001). Both
the documentation of bird fatalities at most wind
power facilities studied in the US (Erickson et al.
2001) and the paucity of general information on
nocturnal bird migration have generated concern
about the potential of collisions between nocturnal
migrants and the many proposed wind power
developments throughout the country.
Consideration of potential wind power impacts on
nocturnal bird migration is particularly important
because more birds migrate at night than during the
daytime (Gauthreaux 1975, Kerlinger 1995). In
particular, passerines (“songbirds”) may be more at
risk of colliding with structures at night because
these birds tend to migrate at lower altitudes than
do other groups of birds (e.g., waterfowl,
shorebirds; Kerlinger 1995).
Ecogen LLC proposes to build the
Prattsburgh–Italy Wind Power project, an ~80-MW
wind power development in the Finger Lakes
region (Yates and Steuben counties) of west-central
New York (Fig. 1). Each of the 53 wind turbines
will have a generating capacity of up to 1.5 MW.
The monopole towers will be ~80 m (262 ft) in
height, and each turbine will have three rotor
blades. The diameter of the rotor blades and hub
will be 70.5 m (231 ft) or 77 m (253 ft), depending
on the model selected for the project; thus, the total
maximal height of a turbine will be approximately
119 m (389 ft) with a blade in the vertical position.
The proposed development is located within the
glaciated Allegheny Plateau section of the
Appalachian Plateaus physiographic province
(USGS 2003), a well-documented migration
corridor for birds (Bull 1985, Bellrose 1976, Zalles
and Bildstein 2000, Cooper and Mabee 2000,
Cooper et al. 2004).
OBJECTIVES
The primary goal of this study was to collect
information on the migration characteristics of
nocturnally migrating birds, especially passerines,
during the fall-migration period; the secondary
goal was to assess the extent of use of the area by
bats to provide an overall assessment of potential
project-related impacts to birds and bats.
Specifically, the objectives of this study were to:
(1) collect baseline information on migration
characteristics (i.e., flight direction, migration
passage rates, flight altitudes) of nocturnally
migrating birds and bats; (2) visually estimate the
relative proportions of birds and bats within the
potential rotor-swept area of the proposed wind
turbines; and (3) determine the number of birds
and bats that would pass within the rotor-swept
area of the proposed wind turbines during the
migratory season. We also evaluated the influence
of weather on migration passage rates and flight
altitudes.
STUDY AREA
The proposed project is located in the Finger
Lakes region of central New York, in Yates and
Steuben counties (Fig. 1). The Finger Lakes region
is part of the Appalachian Plateaus physiographic
province that was formed when the last Pleistocene
ice cap melted and flooded valleys created by the
advancing glacial ice (USGS 2003). This area is
characterized by moderate valleys with ridges that
range from ~980 ft to ~2100 ft (300–650 m) above
the valley floors.
This proposed development is located ~20
miles (~32 km) south of Canandaigua, NY, and ~7
miles (~11 km) southeast of Naples, NY. The
project area consists primarily of a mix of open
farmland and wooded hillsides with limited
residential development. Virtually all of the land
previously has been logged. Our radar sampling
site (UTM 17S 644523E 4360903N) was located
north of Emerson Road, near the northern portion
of the proposed wind power development (Fig. 1).

Study Area
Ecogen Nocturnal Bird and Bat Migration, 2004 2
Figure 1. Map of the proposed Prattsburgh–Italy Wind Power project in Yates and Steuben Counties,
New York.
!
>
Canandaigua
Canandaigua Lake
Naples
Keuka Lake
!
(
53
£
20
!
>
radar sampling site
4
5051015202.5
km
!>
NY
PA

Methods
3Ecogen Nocturnal Bird and Bat Migration, 2004
METHODS
STUDY DESIGN
We conducted radar and visual observations
on 45 nights between 14 August and 29 September
2004, to overlap with the peak of the passerine bird
and tree-roosting bat migration periods during late
summer and fall (Buffalo Ornithological Society
2002, Johnson 2004). We conducted radar
observations during 41 nights (39 nights for visual
observations); on the remaining four nights, we
were unable to sample with either technique
because of inclement weather (rain). Each night,
we conducted ~6.5 h of radar and visual
observations at the site. Radar and visual surveys
occurred between ~2000 and ~0230, to cover the
peak hours of nocturnal passerine migration within
nights (Lowery 1951, Gauthreaux 1971, Alerstam
1990, Kerlinger 1995).
RADAR EQUIPMENT
Our mobile radar laboratory consisted of a
marine radar that was mounted on the roof of a van
and that functioned as both a surveillance and
vertical radar. When the antenna was in the
horizontal position (i.e., in surveillance mode), the
radar scanned the area surrounding the lab (Fig. 2),
and we manually recorded information on flight
direction, flight behavior, passage rates, and
groundspeeds of targets. When the antenna was
placed in the vertical position (i.e., in vertical
mode), the radar scanned the area in an arc across
the top of the lab (Fig. 3), and we manually
measured flight altitudes of targets with an index
line on the monitor. All data was recorded
manually into a laptop computer. A description of a
similar radar laboratory can be found in
Gauthreaux (1985a, 1985b) and Cooper et al.
(1991), and a similar vertical radar configuration
was described by Harmata et al. (1999).
The radar (Furuno Model FR-1510 MKIII;
Furuno Electric Company, Nishinomiya, Japan) is
a standard marine radar transmitting at 9.410 GHz
(i.e., X-band) through a 2-m-long slotted
waveguide (antenna) with a peak power output of
12 kW. The antenna had a beam width of 1.23°
(horizontal) × 25° (vertical) and a sidelobe of
±10–20°. Range accuracy is 1% of the maximal
range of the scale in use or 30 m (whichever is
greater) and bearing accuracy is ±1°.
This radar can be operated at a variety of
ranges (0.5–133 km) and pulse lengths (0.07–1.0
µsec). We used a pulse length of 0.07 µsec while
operating at the 1.5-km range. At shorter pulse
lengths, echo resolution is improved (giving more
accurate information on target identification,
location, and distance), whereas, at longer pulse
lengths, echo detection is improved (increasing the
probability of detecting a target). An echo is a
picture of a target on the radar monitor; a target is
one or more birds (or bats) that are flying so
closely together that the radar displays them as one
echo on the display monitor. This radar has a
digital color display with several scientifically
useful features, including True North correction for
the display screen (to determine flight directions),
color-coded echoes (to differentiate the strength of
return signals), and on-screen plotting of a
sequence of echoes (to depict flight paths).
Because targets plot every sweep of the antenna
(i.e., every 2.5 sec) and because groundspeed is
directly proportional to the distance between
consecutive echoes, we were able to measure
ground speeds of plotted targets to the nearest 5
mi/h (8 km/h) with a hand-held scale.
Energy reflected from the ground,
surrounding vegetation, and other solid objects that
surround the radar unit causes a ground-clutter
echo to appear on the display screen. Because
ground-clutter echoes can obscure targets, we
minimized their occurrence by elevating the
forward edge of the antenna by ~15° and by
parking the mobile radar laboratory in locations
that were surrounded fairly closely by low trees or
low hills, whenever possible. These objects act as a
radar fence that shields the radar from low-lying
objects farther away from the lab and that produce
only a small amount of ground clutter in the center
of the display screen. For further discussion of
radar fences, see Eastwood (1967), Williams et al.
(1972), Skolnik (1980), and Cooper et al. (1991).
Maximal distances of detection of targets by
the surveillance radar depends on radar settings
(e.g., gain and pulse length), target body size, flock
size, flight profile, proximity of targets in flocks,
atmospheric conditions, and, to some extent, the
amount and location of ground clutter. Flocks of
waterfowl routinely were detected to 5–6 km,

Methods
Ecogen Nocturnal Bird and Bat Migration, 2004 4
Figure 2. Approximate airspace sampled by Furuno FR–1510 marine radar when operating in the
surveillance mode (antenna in the horizontal orientation) as determined by field trials with
Rock Pigeons. Note that the distribution of the radar beam within 250 m of the origin (i.e.,
the darkened area) was not determined.
Figure 3. Approximate airspace sampled by Furuno FR–1510 marine radar when operating in the
vertical mode (antenna in the vertical orientation) as determined by field trials with Rock
Pigeons. Note that the distribution of the radar beam within 250 m of the origin (i.e., the
darkened area) was not determined.

Methods
5Ecogen Nocturnal Bird and Bat Migration, 2004
individual hawks usually were detected to 2–3 km,
and single, small passerines were routinely
detected out to 1–1.5 km (Cooper et al. 1991).
DATA COLLECTION
TARGET IDENTIFICATION ON RADAR
The species composition and size of a flock of
birds or bats observed on the radar usually was
unknown. Therefore, the term “target,” rather than
“flock” or “individual,” is used to describe animals
detected by the radar. Based on the study period
and location, it is likely that the majority of targets
that we observed were individual passerines, which
generally do not migrate in tight flocks (Lowery
1951, Kerlinger 1995); it also is likely that a
smaller number of targets were migratory bats.
Differentiating among various targets (e.g., birds,
bats, insects) is central to any radar study,
especially with X-band radars that can detect small
flying animals. Because bat flight speeds overlap
with flight speeds of passerines (i.e., are >6 m/s;
Tuttle 1988, Larkin 1991, Bruderer and Boldt
2001, Kunz and Fenton 2003; Cooper and Day,
ABR Inc., unpubl. data), it was not possible to
separate bird targets from bat targets based solely
on flight speeds. We were able to exclude foraging
bats based on their erratic flight patterns; however,
it is likely that migratory bats or any bats not
exhibiting erratic flight patterns were included in
our data.
Of primary importance in target identification
is the elimination of insect targets. We reduced
insect contamination by (1) omitting small targets
(the size of gain speckles) that only appeared
within ~500 m of the radar and targets with poor
reflectivity (e.g., targets that plotted erratically or
inconsistently in locations having good radar
coverage); and (2) editing data prior to analyses by
omitting surveillance and vertical radar targets
with corrected airspeeds <6 m/s (following Diehl et
al. 2003). The 6 m/s airspeed threshold was based
on radar studies that have determined that most
insects have an airspeed of <6 m/s, whereas that of
birds and bats usually is 6 m/s (Tuttle 1988, Larkin
1991, Bruderer and Boldt 2001, Kunz and Fenton
2003; Cooper and Day, ABR Inc., unpubl. data).
SAMPLING DESIGN
Each of the six 1-hr nocturnal radar and visual
sampling sessions/night consisted of: (1) one 5–10
min session to conduct visual observations; (2) one
5–10 min session to collect weather data and adjust
the radar to surveillance mode; (3) one 10-min
session with the radar in surveillance mode
(1.5-km range) for collection of information on
migration passage rates; (4) one 10-min session
with the radar in surveillance mode (1.5-km range)
for collection of information on groundspeed,
flight direction, tangential range (minimal
perpendicular distance to the radar laboratory),
transect crossed (the four cardinal
directions—north, south, east, and west), species
(if known), and the number of individuals (if
known); (5) one 10-min session to collect weather
data and adjust the radar to vertical mode; and (6)
one 15-min session with the radar in vertical mode
(1.5-km range) to collect information on flight
altitudes, speed, and direction. After the
completion of the six hourly sessions, we also
conducted an additional 30-min radar session in the
surveillance mode (as above).
For each vertical radar session, the antenna
was oriented parallel to the main axis of migration
(determined by the overall flight direction seen
during the previous surveillance radar session) to
maximize the true flight speed of targets. True
flight speeds of targets can be determined only for
those targets flying parallel to the antenna's
orientation because slower speeds are obtained
when targets fly at an angle to this plane of
orientation. We also randomly selected a session
each night to conduct intensive sampling sessions
(continuous 25-min samples divided into 5-min
periods) both for passage rates (surveillance mode)
and mean flight altitudes (vertical mode) to
determine whether our sampling intensity was
adequate.
Weather data collected twice each hour
consisted of the following: wind speed (collected
with a “OMNI” anemometer in 5-mph [2.2-m/s]
categories); wind direction (in ordinal categories to
the nearest 45°); cloud cover (to the nearest 5%);
ceiling height (in m agl; 1–50, 51–100, 100–150,
151–500, 501–1,000, 1,001–2,500, 2,501–5,000,
>5,000); minimal visibility in a cardinal direction
(in m; 0–50, 51–100, 101–500, 501–1,000,

Methods
Ecogen Nocturnal Bird and Bat Migration, 2004 6
1,001–2,500, 2,501–5,000, >5,000); precipitation
level (no precipitation, fog, drizzle, light rain,
heavy rain, snow flurries, light snowfall, heavy
snowfall, sleet, hail); and air temperature
(measured with a thermometer to the nearest 1°C).
We could not collect radar data during rain because
the electronic filtering required to remove the
echoes of the precipitation from the display screen
also removed those of the targets of interest. We
also obtained weather data (wind speed and wind
direction) from a 50-m high meteorological tower
located near the site.
VISUAL OBSERVATIONS OF LOW-ALTITUDE
BIRDS AND BATS
We conducted visual observations every night
to assess relative numbers of birds and bats flying
within the projected rotor-swept area (i.e., <125 m
agl). During the first hour of surveys (prior to
~2030), observers used 10× power binoculars to
scan for bat activity during crepuscular (twilight)
periods. During subsequent hours, 2-million-Cp
spotlights with red lens filters (to reduce the
attractiveness of the light to insects, birds, and
bats) were used to illuminate targets flying
overhead. One “fixed” spotlight was mounted on a
tripod with the beam oriented vertically, while a
second, handheld light was used to track and
identify potential targets flying through the fixed
beam. For each bird or bat detected visually, we
recorded the taxon (to species when possible),
flight direction, flight altitude, and behavior
(straight-line, erratic, circling). Bats were
classified as “large bats” or “small bats” whenever
possible in an attempt to discriminate the larger
Hoary, Eastern Red, and Silver-haired bats from
smaller species (e.g., Myotis spp.). Prior to 14
September, observers conducted 5–10 min of
unaided visual observations during each hour of
radar surveys. Because of our limited range (i.e.,
from 0 to ~75 m agl) in which to detect birds and
bats while using only spotlights, observers after 13
September used 1X ATN-PVS7 Generation 3
night-vision goggles to enhance the detectability of
targets; we also increased our sampling intensity
by sampling an additional 15 min before the first
radar sampling session and 45 min after the last
radar sampling session.
DATA ANALYSES
RADAR DATA
We entered all radar data into MS Excel
databases. Data files were checked visually for
errors after each night and then were checked again
electronically for irregularities at the end of the
field season, prior to data analyses. All analyses
were conducted with SPSS statistical software
(SPSS 2003). For quality assurance, we
cross-checked results of the SPSS analyses with
hand-tabulations of small data subsets whenever
possible. Radar data were not corrected for
differences in detectability with distance from the
radar unit. The level of significance (α) for all
statistical tests was set at 0.05.
Airspeeds (i.e., groundspeed corrected for
wind speed and relative direction) of
surveillance-radar targets were computed with the
formula:
,
where Va = airspeed, Vg = target groundspeed (as
determined from the radar flight track), Vw=wind
velocity, and θ is the difference between the
observed flight direction and the direction of the
wind vector. Targets that had corrected airspeeds
<6 m/s (16.2% of surveillance data; 34.5% of
vertical data) were deleted from all analyses.
We analyzed flight-direction data following
procedures for circular statistics (Zar 1999) with
Oriana software version 2.0 (Kovach 2003).
Migration passage rates are reported as the mean ±
1 standard error (SE) number of targets passing
along 1 km of migratory front/h (targets/km/h ± 1
SE). Passage rates of targets flying <125 m in
altitude were derived for each hourly period by
multiplying passage rates recorded from
surveillance radar by the percentage of targets on
vertical radar having flight altitudes <125 m. All
flight-altitude data are presented in m agl (above
ground level) relative to a horizontal plane passing
through the radar-sampling site. Actual mean
altitudes may be higher than those reported
because an unknown number of birds fly above the
1.5-km range limit of our radar (Mabee and Cooper
2004).
cosθV2VVVV wg
2
w
2
ga −+=

Results
7Ecogen Nocturnal Bird and Bat Migration, 2004
For calculations of the daily patterns in
migration passage rates and flight altitudes, we
assumed that a day began at 0700 on one day and
ended at 0659 the next day, so that a sampling
night was not split between two dates. We used
repeated-measures ANOVAs with the
Greenhouse-Geisser epsilon adjustment for
degrees of freedom (SPSS 2003), to compare
passage rates and flight altitudes among hours of
the night for nights with data collected during all
six sessions. We examined the effects of session
length on the accuracy of estimating passage rates
and flight altitudes by regressing the different
counts in each 5-min time interval to that obtained
from a continuous 25 min sample. Factors that
decreased our sample size of the various
summaries and analyses included insect
contamination and inclement weather (rain).
Sample sizes therefore sometimes varied among
the different summaries and analyses.
THE EFFECTS OF WEATHER ON
MIGRATION PASSAGE RATES AND FLIGHT
ALTITUDES
We examined the hourly relationships
between passage rates, flight altitudes, and weather
conditions because of the dynamic weather
conditions within a night. This treatment of the
data, however, may violate the assumption of
statistical independence; therefore, our results may
overemphasize the strength of the relationships
presented.
We modeled the hourly influence of weather
and date separately on the dependent variables
passage rates and flight altitudes. We obtained our
weather data (i.e., wind speed and direction) from a
50-m meteorological tower located <1 km from the
radar sampling site. All wind categories except the
calm category had a mean wind speed of ≥2.2 m/s
(i.e., ≥5 mph) and were categorized as the
following: headwinds ESE to SSW (i.e.,
113º–248º), tailwinds WNW to ENE (i.e.,
293º–068º), eastern crosswinds (069º–112º),
western crosswinds (249º–292º), and calm (0–2.2
m/s).
Prior to model specification, we examined the
data for redundant variables (Spearman’s rs>0.70)
and retained five parameters for inclusion in the
model set. We examined scatterplots and residual
plots to ensure that variables met assumptions of
analyses (i.e., linearity, normality, collinearity) and
did not contain presumed outliers (>4 SE). We
used a logarithmic transformation on the dependent
variable "passage rate" to make the data more
normal, whereas the flight altitudes were normally
distributed. We specified 16 models for passage
rates and flight altitudes: a global model containing
all variables and subset models representing
potential influences of four weather variables
(wind speed, wind direction, the presence of fog,
and ceiling height) and date on migration passage
rates and flight altitudes. We analyzed all model
sets with linear regression. Prior to model
selection, we examined the fit of global models
following recommendations of Burnham and
Anderson (1998) that included examining residuals
and measures of fit (R² = 0.39 for passage-rate
models; R² = 0.21, for flight-altitude models).
Because the number of sampling sessions for
both passage rates (n= 241) and flight altitudes
(n= 195) was small relative to the number of
parameters (K) in many models (i.e., n/K < 40), we
used Akaike’s Information Criterion corrected for
small sample size (AICc) for model selection
(Burnham and Anderson 1998). We used the
formulas presented in Burnham and Anderson
(1998) to calculate AICc for our least-squares
(linear regression) methods. We ranked all
candidate models according to their AICc values
and considered the best-approximating model (i.e.,
most parsimonious) to be that model having the
smallest AICc value (Burnham and Anderson
1998). We drew primary inference from models
within 2 units of the minimal AICc value, although
models within 4–7 units may have some empirical
support (Burnham and Anderson 1998). We
calculated Akaike weights (wi) to determine the
weight of evidence in favor of each model and to
estimate the relative importance of individual
parameters (Burnham and Anderson 1998). All
analyses were conducted with SPSS software
(SPSS 2003).
RESULTS
FLIGHT DIRECTION
At night, most radar targets were traveling in
seasonally appropriate directions for fall migration
(i.e., southerly), with a mean flight direction of

Results
Ecogen Nocturnal Bird and Bat Migration, 2004 8
177° for the fall season (mean vector length = 0.39;
n = 3,854 targets; Fig. 4). Most (74%) of the
nocturnal targets were traveling in a southerly
direction, with half (52%) of the flight directions
occurring between SE (135°) and SW (225°).
PASSAGE RATES
The mean nocturnal passage rate for the fall
season was 200 ± 12 targets/km/h (n = 41 nights).
Overall mean nightly passage rates were highly
variable among nights (range = 18–863
targets/km/h) with relatively small numbers of
targets below 125 m agl (Fig.5). Passage rates
varied significantly among hours of the night (F3.6,
93.7 = 4.551; P = 0.003; n = 27 nights; Fig. 6), with
lowest rates occurring during the first hour of
darkness.
FLIGHT ALTITUDES
The mean nocturnal flight altitude for the
entire fall season was 365 ± 3 m agl (n= 6,856
targets; median = 325 m agl). Mean flight altitudes
observed on vertical radar (1.5-km range) were
highly variable among nights and ranged from 202
to 584 m agl (Fig. 7). Mean flight altitudes did not
vary among hours of the night (F5, 95 = 0.298,
P=0.913, n = 35 nights; Fig. 8), and the altitudinal
distribution of targets did not appear to vary among
hours during 41 nights (Fig. 9). The overall
distribution of targets in 100-m categories of flight
altitudes varied from 21.3% in the 201–300 m agl
interval to 0% in the 1,301–1,400 and 1,401–1,500
m agl intervals (Table 1). We determined that
9.2% of all targets flew <125 m, which is the
approximate maximal height of the proposed wind
turbines.
Figure 4. Flight directions of radar targets at the proposed Prattsburgh–Italy Wind Power project, New
York fall 2004.
400
300
200
100
n = 500
0°
90°270°
180°

Results
9Ecogen Nocturnal Bird and Bat Migration, 2004
Figure 5. Mean ± 1 SE nightly passage rates (targets/km/h) at the proposed Prattsburgh–Italy Wind
Power project, New York, fall 2004. Asterisks denote nights not sampled because of rain (n =
4) or technical difficulties (n = 2).
0
100
200
300
400
500
600
700
800
900
1000
Passage rate <125 m agl
n = 246 sessions
14 21 28 4 11 18 25
Overall passage rate
August September
*** *
**
Mean passage rate (targets/km/h)
Figure 6. Percent of seasonal passage rate (± 1SE) by hour of the night (e.g., 2000–2059) at the
proposed Prattsburgh-Italy Wind Power project, New York, fall 2004.
2000 2100 2200 2300 0000 0100
0
5
10
15
20
25
n = 27 nights
Time
Percent of seasonal passage rate

Results
Ecogen Nocturnal Bird and Bat Migration, 2004 10
Figure 7. Mean ± 1 SE nightly flight altitudes (m agl) at the proposed Prattsburgh–Italy Wind Power
project, New York, fall 2004. Asterisks denote nights not sampled because of rain (n = 4) or
technical difficulties (n = 2).
0
100
200
300
400
500
600
700
n = 6,856 targets
*** ** *
28 4 11 18 25
21
14
August September
***** *
Mean flight altitude (m agl)
Figure 8. Mean ± 1 SE flight altitude (m agl) by hour of the night (e.g., 2000–2059) at the proposed
Prattsburgh–Italy Wind Power project, New York, fall 2004.
2000 2100 2200 2300 0000 0100
0
100
200
300
400
n = 35 nights
Time
Mean flight altitude (m agl)

Results
11 Ecogen Nocturnal Bird and Bat Migration, 2004
Figure 9. Percent of radar targets at each altitude at the proposed Prattsburgh–Italy Wind Power project,
New York, fall 2004, by hour of the night (e.g., 2000–2059).
2000 2100 2200 2300 0000 0100
0
10
20
30
40
50
60
70
80
90
100
0-200 m agl
201-400 m agl
401-600 m agl
601-800 m agl
801-1000 m agl
1,001-1,500 m agl
n = 7,729 targets
Time
Percent of targets by altitude
Table 1. Nocturnal flight altitudes of radar targets (% of all targets) detected at the 1.5-km range at the
Prattsburgh–Italy Wind Power project, NY, fall 2004, by flight-altitude category. Total n =
6,856 targets.
Flight altitude (m agl) Percent of radar targets
0–100 6.0
101–200 18.0
201–300 21.3
301–400 18.3
401–500 13.4
501–600 9.0
601–700 5.7
701–800 3.6
801–900 2.6
901–1,000 1.2
1,001–1,100 0.5
1,101–1,200 0.2
1,201–1,300 0.1
1,301–1,400 0.0
1,401–1,500 0.0

Results
Ecogen Nocturnal Bird and Bat Migration, 2004 12
INTENSIVE SAMPLING PERIODS
Over 90% of the variation in a 25-min
passage-rate sampling session was explained by a
5-min count of passage rate, and ~96% was
explained by our current method of conducting a
10-min count (Fig. 10). Nearly 95% of the
variation in a 25-min flight-altitude sampling
session was explained by using our current method
of sampling for 15 min (Fig 10).
EFFECTS OF WEATHER ON MIGRATION
We investigated the importance of weather
(i.e., wind direction, wind speed, fog, ceiling
height) and date on both the passage rates and
flight altitudes of nocturnal migrants by building a
series of models (combinations of the various
weather variables and date), then using a
model-selection technique (AIC) to quantify the
statistical strength of those models. The AIC
method allows one to (1) rank and identify the
“best” model(s) (i.e., the most statistically
supported models) from the full set of models, and
(2) assess the statistical strength and relative
importance of individual variables composing the
“best” models.
PASSAGE RATES
The best-approximating model explaining
migration passage rates of nocturnal migrants
during fall migration was the model containing the
variables wind direction, wind speed, and date
(Table 2). The second-best model contained the
variables wind direction and wind speed
(∆AICc= 0.78), and a third model containing wind
direction, wind speed, date, and fog
(∆AICc= 0.85) also received similar empirical
support (Table 2). The top two models contained
Figure 10. Variation in passage rate and flight altitude counts by length of sampling session at the
proposed Prattsburgh–Italy Wind Power project, New York, fall 2004. The Coefficient of
Variation (R²) expresses the amount of variation in a 25-min count (the maximal time
sampled) explained by counts of shorter duration.
0 5 10 15 20 25
0.6
0.7
0.8
0.9
1.0
Passage rate, n = 31 nights
Mean flight altitude, n = 27 nights
Minutes sampled
Coefficient of Determination (R
2
)

Results
13 Ecogen Nocturnal Bird and Bat Migration, 2004
Table 2. Linear-regression models explaining the influence of environmental factors on migration passage rates of bird and bat targets on
surveillance radar at the Prattsburgh–Italy Wind Power project, NY, fall 2004 (n = 241 sampling sessions). Model weights (wi) were
based on Akaike’s Information Criterion (AIC).
Model RSSa Kb AICc
c ∆ AICc
d wi
e
Wind direction + wind speed + date 29.6 8 –489.17 0.00 0.39
Wind direction + wind speed 29.9 7 –488.39 0.78 0.27
Wind direction + wind speed + date + fog 29.4 9 –488.32 0.85 0.26
Global model: wind direction + wind speed + date + fog + ceiling height 29.4 10 –486.13 3.04 0.09
Wind speed 32.8 3 –474.47 14.70 0.00
Wind direction 32.3 6 –471.69 17.48 0.00
Wind direction + date 32.1 7 –471.06 18.11 0.00
Wind direction + fog 32.2 7 –470.54 18.63 0.00
Wind direction + date + fog 32.0 8 –470.04 19.12 0.00
Wind direction + ceiling height 32.3 7 –469.94 19.23 0.00
Wind direction + date + ceiling height 32.1 8 –469.15 20.02 0.00
Date + fog 43.7 4 –403.60 85.57 0.00
Date 45.4 3 –396.36 92.81 0.00
Fog 45.6 3 –394.98 94.19 0.00
Date + ceiling height 45.4 4 –394.29 94.88 0.00
Ceiling height 47.3 3 –386.32 102.85 0.00
a Residual sum of squares.
b Number of estimable parameters in approximating model.
c Akaike’s Information Criterion corrected for small sample size.
d Difference in value between AICc of the current model versus the best approximating model with the minimal AICc value.
e Akaike weight—probability that the current model (i) is the best approximating model among those being considered.

Results
Ecogen Nocturnal Bird and Bat Migration, 2004 14
the same significant positive associations with
tailwinds, calm conditions, and eastern and western
crosswinds (Table 3). Date and wind speed were
not related to passage rates. The weight of
evidence in favor of the “best” model
(wbest/wsecond best; Burnham and Anderson 1998)
was 1.4 times that of the second-best model
(Burnham and Anderson 1998). The sum of
Akaike weights (Σwi) of parameters across all
models provided evidence for the relative
importance of variables from these models, with
wind direction and wind speed (1.00) being more
important than date (0.73), fog (0.34), and ceiling
height (0.09).
FLIGHT ALTITUDES
The best-approximating model explaining
flight altitudes of nocturnal migrants during fall
migration was the model containing the variables
wind direction and date (Table 4). The next three
best models also received empirical support
(∆AICc < 2; Table 4). The top two models
contained strong positive associations with
tailwinds, calm winds, western crosswinds, and
date (Table 5). Flight altitudes were not related to
eastern crosswinds and fog. The weight of
evidence in favor of the “best” model
(wbest/wsecond best) was 1.7 times that of the second
best model. The Σwi suggested that both wind
direction (1.00) and date (0.91) were more
important than fog (0.33), wind speed (0.29), and
ceiling height (0.20).
TARGETS WITHIN THE PROPOSED
TURBINE AREA
In the fall, the mean passage rate of targets
<125 m was 20.0 ± 3.7 targets/km/h. We made
several assumptions to estimate the turbine passage
rate (i.e., the number of targets that would pass
within the area occupied by each proposed
turbine): (1) the minimal area occupied by the wind
turbine (i.e., side profile), (2) the maximal area
occupied by the wind turbine (i.e., front profile,
including the rotor-swept area), (3) a worst-case
scenario of the rotor blades turning constantly, (4)
Table 3. Parameter estimates from the two best-approximating models explaining the influence of
environmental factors on passage rates of bird and bat targets at the Prattsburgh–Italy Wind
Power project, NY, fall 2004 (n = 241 sampling sessions). Coefficients (B) of the categorical
variable wind direction were calculated relative to headwinds.
Model B SE R²
Wind direction + wind speed 0.368
Intercept 2.269 0.100
Wind direction = tailwind 0.310 0.070
Wind direction = calm 0.295 0.109
Wind direction = E crosswind 0.373 0.167
Wind direction = W crosswind 0.215 0.070
Wind speed -0.047 0.110
Wind direction + wind speed + date 0.375
Intercept 1.537 0.448
Wind direction = tailwind 0.290 0.071
Wind direction = calm 0.254 0.112
Wind direction = easterly crosswind 0.371 0.166
Wind direction = westerly crosswind 0.169 0.075
Wind speed -0.049 0.110
Date 0.003 0.002

Results
15 Ecogen Nocturnal Bird and Bat Migration, 2004
Table 4. Linear-regression models explaining the influence of environmental factors on mean flight altitudes of bird and bat targets on vertical
radar at the Prattsburgh–Italy Wind Power project, NY, fall 2004 (n = 195 sampling sessions). Model weights (wi) were based on
Akaike’s Information Criterion (AIC).
Model RSSa Kb AICc
c ∆ AICc
d wi
e
Wind direction + date 1,842,000.0 7 1799.50
0.00 0.31
Wind direction + date + fog 1,831,310.0 8 1800.54
1.04 0.18
Wind direction + date + ceiling height 1,834,599.0 8 1800.89 1.39 0.15
Wind direction + wind speed + date 1,835,284.0 8 1800.97
1.46 0.15
Wind direction + wind speed + date + fog 1,824,268.0 9 1801.99
2.49 0.09
Global model: wind direction + wind speed + date + fog + ceiling height 1,821,139.0 10 1803.88
4.38 0.03
Wind direction 1,905,267.0 6 1803.94
4.43 0.03
Wind direction + fog 1,892,815.0 7 1804.81
5.31 0.02
Wind direction + wind speed 1,899,937.0 7 1805.54
6.04 0.02
Wind direction + ceiling height 1,900,213.0 7 1805.57
6.07 0.01
Wind speed 2,087,084.0 3 1815.39
15.89 0.00
Date 2,136,473.0 3 1819.95
20.45 0.00
Date + ceiling height 2,132,745.0 4 1821.70 22.19 0.00
Date + fog 2,136,456.0 4 1822.03
22.53 0.00
Ceiling height 2,301,564.0 3 1834.47
34.96 0.00
Fog 2,302,377.0 3 1834.53
35.03 0.00
a Residual sum of squares.
b Number of estimable parameters in approximating model.
c Akaike’s Information Criterion corrected for small sample size.
d Difference in value between AICc of the current model versus the best approximating model with the minimum AICc value.
e Akaike weight—probability that the current model (i) is the best approximating model among those being considered.

Discussion
Ecogen Nocturnal Bird and Bat Migration, 2004 16
45 d in the study, and (5) an average of 10
nocturnal hours/day across the 45-d period. If all
migrants approached the turbines from the side, an
estimated 51 migrants would have passed within
the area occupied by one turbine. If all migrants
approached the turbines from the front, an
estimated 362 migrants would have passed within
the area occupied by one turbine (Appendix 1).
VISUAL DATA
We collected visual data on 39 nights during
the fall field season. We did not observe any bats
moving during crepuscular sessions (~1930–2015).
During our initial period of spotlight observations
(14 August–13 September), we observed low
numbers of birds and bats (mean = 1.98 visual
targets/h; Fig. 11) and were able to identify 70% of
all targets (n = 20) as either birds or bats. By using
night-vision goggles with the filtered spotlights
(14–29 September), we were able to identify 75%
of the targets observed (n = 106). In addition, we
generally observed higher passage rates of birds
and bats with this improved method (mean = 5.23
visual targets/h; Fig. 11). During this time period
(14–29 September), proportions of birds and bats
flying <~150 m agl (our effective sampling
distance with the night-vision goggles) were 94%
birds and 6% bats (n = 80).
DISCUSSION
Predictions of the effects of wind power
development on migratory birds and bats are
hampered by a lack of detailed knowledge about
patterns of the nocturnal migration and behavior of
birds and bats around wind turbines. We have
documented some of the key migration
characteristics that can be used both to assess the
risk of collision with wind turbines and to describe
general properties of nocturnal bird migration at
the proposed project site.
TIMING OF MIGRATION
Understanding the timing of migration at
multiple temporal scales (e.g., within nights, within
seasons, and seasonally within years) allows the
determination of patterns of peak migration that
Table 5. Parameter estimates from the two best-approximating models explaining the influence of
environmental factors on mean flight altitudes of radar targets at the Prattsburgh–Italy Wind
Power project, NY, fall 2004 (n = 195 sampling sessions). Coefficients (B) of the categorical
variables "wind direction" and "fog" were calculated relative to headwinds and fog conditions.
Model B SE R²
Wind direction + date 0.200
Intercept –48.282 131.401
Wind direction = tailwind 79.375 18.211
Wind direction = calm 92.710 22.046
Wind direction = easterly crosswind 54.705 41.618
Wind direction = westerly crosswind 69.989 23.005
Wind direction + date + fog 0.205
Intercept –97.975 139.668
Wind direction = tailwind 80.666 18.248
Wind direction = calm 98.877 22.813
Wind direction = easterly crosswind 54.732 41.607
Wind direction = westerly crosswind 70.258 23.000
Date 1.328 0.529
Fog 54.193 51.731

Discussion
17 Ecogen Nocturnal Bird and Bat Migration, 2004
can be used with other information, especially
weather, to develop predictive models of avian and
bat collision risks. Such models may be useful for
both pre-construction siting decisions and for the
consideration of operational strategies to reduce
fatalities.
Within nights, fall passage rates increased
~1–2 h after sunset, peaked prior to midnight, then
decreased slightly later in the evening. Several
studies have found a pattern similar to this, in
which the intensity of nocturnal migration begins
to increase ~30–60 min after sunset, peaks around
midnight, and declines steadily thereafter until
dawn (Lowery 1951, Gauthreaux 1971, Kerlinger
1995, Farnsworth et al. 2004).
Within seasons, nocturnal migration often is a
pulsed phenomenon (Alerstam 1990; Cooper and
Day, ABR, unpubl. data). In this study,
moderate–large mean nightly passage rates (>300
targets/km/h) occurred on 9 nights: 21 and 31
August and 7, 10, 13, 18, 23, 25, and 28
September. Overall, fall migration peaked at 863
targets/km/h on 10 September. Thus, the migratory
period we studied was characterized by many
migratory pulses throughout the season. In general,
most fall songbird migration in this part of New
York occurs between late August and mid-October
(Cooper and Mabee 2000; Buffalo Ornithological
Society 2002; W. Evans, Old Bird Inc., pers.
comm.).
PASSAGE RATES
Passage rates are an index of the number of
migrants flying past a location; thus, they may be
useful to assess the relative importance of several
sites being considered for wind power
development. The high daily variation in migration
passage rates during the fall (18–863 targets/km/h)
seen in this study illustrates the importance of
Figure 11. Mean number of birds or bats/h observed during visual sampling at the proposed
Prattsburgh–Italy Wind Power project, New York, fall 2004. Asterisks denote nights not
sampled because of rain (n = 4) or technical difficulties (n = 2).
0
2
4
6
8
10
12
Bird
Bat
**** ** *
Night vision gogglesSpotlights
21 4 11 18 2514 28
August September
Mean number of targets/hour

Discussion
Ecogen Nocturnal Bird and Bat Migration, 2004 18
continuous sampling throughout each entire
migration period to identify these few, but
important, migration nights.
In this study we used our passage-rate data in
two ways: (1) to examine the passage rate of all
migrants passing over our study area, and (2) to
examine the passage rate of migrants within the
height of the proposed wind turbines (~125 m).
Although both metrics are useful for comparing the
relative importance of sites, the second metric is
especially well-suited for site comparisons among
wind power developments because of its
altitude-specific nature. This second metric also
can be used as the starting point for a more
in-depth risk assessment.
The observed passage rates in the project area
were comparable to those at other locations in New
York where we have conducted fall migration
studies with similar equipment and methods. The
mean fall nocturnal passage rate in this study was
200 targets/km/h, compared with fall passage rates
of 122 targets/km/h at Harrisburg, NY (located
~190 km northeast of this study site; Cooper and
Mabee 2000); 168 targets/km/h at Wethersfield,
NY (located ~80 km northwest of this study site;
Cooper and Mabee 2000); 225 targets/km/h at
Carthage, NY (located ~200 km northeast of this
study site; Cooper et al. 1995b), and 238
targets/km/h at Chautauqua, NY (located ~190 km
southwest of this study site; Cooper et al. 2004).
Fall passage rates in other locations in the eastern
US were similar to what we recorded here (e.g.,
199–241 targets/km/h at Mt. Storm, WV; Mabee et
al. 2004). In contrast, lower passage rates have
generally been observed in the Midwest (e.g.,
27–108 targets/km/h at four sites in South Dakota
and Minnesota; Day and Byrne 1990) and the West
(e.g., 19–26 targets/km/h at the Stateline and
Vansycle wind power facilities in eastern Oregon;
Mabee and Cooper 2004).
Our estimates of passage rates below the
proposed turbine height in the project area (20.0
targets/km/h flying <125 m agl) were similar to fall
rates at the Chautauqua site in western New York
(20.8 targets/km/h flying <140 m agl; Cooper et al.
2004) and were lower than those rates observed at
the Mount Storm site along an Appalachian
ridgeline in West Virginia (36.3 targets/km/h flying
<125 m agl; Mabee et al. 2004). Unfortunately, we
do not believe that it is appropriate to compare
flight altitudes in this study with those at other
New York sites studied before 2001 (Wethersfield,
Harrisburg, Carthage) because of different
equipment (i.e., a different vertical radar
configuration) used in those studies.
FLIGHT ALTITUDES
Flight altitudes are critical for understanding
the vertical distribution of nocturnal migrants in
the airspace and are another important metric for
assessing the risk of avian fatality events at
proposed wind power development sites. In
general, passerines migrate at lower flight altitudes
than do other major groups of over-land migrants
such as shorebirds and waterfowl (Kerlinger 1995).
Large kills of birds at tall, human-made structures
(generally lighted and guyed communications
towers; Avery et al. 1980) and the predominance of
nocturnal migrant passerines at such kills
(Manville 2000) indicate that large numbers of
these birds fly <500 m agl on at least some nights.
Mean flight altitudes at the proposed project
site were lower (365 m agl) than those at other sites
studied in the fall in New York (Chautauqua, mean
= 532 m agl) and West Virginia (Mt. Storm, mean =
410 m agl). Unfortunately, we do not believe that it
is appropriate to compare flight altitudes in this
study with those at other New York sites studied
before 2001 (Wethersfield, Harrisburg, Carthage)
because of different equipment (i.e., a different
vertical radar configuration) that probably resulted
in a low altitude bias. Similar to our results,
however, other studies that used a variety of radar
systems and analyses have indicated that the
majority of nocturnal migrants fly below 600 m agl
(Bellrose 1971; Gauthreaux 1972, 1978, 1991;
Bruderer and Steidinger 1972; Cooper and Ritchie
1995). Kerlinger (1995) summarized radar results
from the eastern US and concluded that
three-quarters of passerines migrate <600 m agl.
In contrast to these results, other researchers
have found that peak nocturnal densities extend
over a broad altitudinal range up to ~2,000 m
(Harper 1958, in Eastwood 1967; Graber and
Hassler 1962, Nisbet 1963, Bellrose and Graber
1963, Eastwood and Rider 1965, Bellrose 1967,
Blokpoel 1971; Richardson 1971, 1972; Blokpoel
and Burton 1975). We suspect that differences
between the two groups of studies are largely due

Discussion
19 Ecogen Nocturnal Bird and Bat Migration, 2004
to differences in location, species-composition of
migrating birds, local topography, radar equipment
used, and perhaps weather conditions. It has been
suggested that limitations in equipment and
sampling methods of some previous radar studies
may have been responsible for their overestimation
of the altitude of bird migration (Able 1970,
Kerlinger and Moore 1989). For example, the
radars used by Bellrose and Graber (1963),
Blokpoel (1971), and Nisbet (1963) could not
detect birds below 450 m, 370 m, and 180 m agl,
respectively. In contrast, our vertical radar could
detect targets down to ~10–15 m agl, allowing us
to detect low-altitude migrants.
We also examined the percentage of targets
below approximate turbine height (i.e., 125 m agl)
and estimated that ~9% flew <125 m agl at this
study site, compared with 4% <140 m agl at
Chautauqua, NY (Cooper et al. 2004), 13–16%
<125 m agl at Mt. Storm, WV (Mabee et al. 2004),
and 3–9% <125 m agl at the Stateline and Vansycle
wind power facilities in eastern Oregon (Mabee
and Cooper 2004). Based on observations made
during this study, mean flight altitudes and the
proportion of targets flying 200 m agl did not vary
among hours of the night.
Similar to our migration studies elsewhere
(Cooper and Ritchie 1995; Cooper et al. 1995a,
1995b; Cooper and Mabee 2000; Mabee and
Cooper 2004), we recorded large among-night
variation in mean flight altitudes during the fall
migration season, although mean flight altitudes
always were above the proposed turbine heights
(observed minimum = 202 m agl). Daily variation
in mean flight altitudes may have reflected changes
in species composition, vertical structure of the
atmosphere, and/or weather conditions. Variation
among days in the flight altitudes of migrants at
other locations has been associated primarily with
changes in the vertical structure of the atmosphere.
For example, birds crossing the Gulf of Mexico
appear to fly at altitudes where favorable winds
minimize the energetic cost of migration
(Gauthreaux 1991). Kerlinger and Moore (1989),
Bruderer et al. (1995), and Liechti et al. (2000)
have concluded that atmospheric structure is the
primary selective force determining the height at
which migrating birds fly.
INTENSIVE SAMPLING PERIODS
Our current method of sampling for 10 min/h
to obtain a representative passage-rate count
(targets/km/h) appears to be justified, judging from
the high degree of correlation between estimates
based upon 10-min samples and those from 25-min
samples (i.e., R2 = 0.96). Similarly, our current
method of sampling for 15 min/h to obtain a
representative mean flight altitudes also appears
justified, judging by the high degree of correlation
(R2 = 0.95) between 15-min and 25-min samples.
MODELING MIGRATION PASSAGE RATES
AND FLIGHT ALTITUDES
MIGRATION PASSAGE RATES
It is a well-known fact that general weather
patterns and their associated temperatures and
winds affect migration (Richardson 1978, 1990). In
the Northern Hemisphere, air moves
counterclockwise around low-pressure systems and
clockwise around high-pressure systems. Thus,
winds are warm and southerly when an area is
affected by a low to the west or a high to the east
and are cool and northerly in the reverse situation.
Clouds, precipitation, and strong, variable winds
are typical in the centers of lows and near fronts
between weather systems, whereas weather usually
is fair with weak or moderate winds in
high-pressure areas. Numerous studies in the
Northern Hemisphere have shown that, in fall,
most bird migration tends to occur in the western
parts of lows, the eastern or central parts of highs,
or in intervening transitional areas. In contrast,
warm fronts, which are accompanied by southerly
(unfavorable) winds and warmer temperatures,
tend to slow fall migration (Lowery 1951,
Gauthreaux 1971; Able 1973, 1974; Blokpoel and
Gauthier 1974, Richardson 1990). Conversely,
more intense spring migration tends to occur in the
eastern parts of lows, the western or central parts of
highs, or in intervening transitional areas.
We examined the influence of weather (i.e.,
wind speed, wind direction, date, fog, and ceiling
height) on migration passage rates and identified
wind direction and wind speed as the most
important factors. Fall migration passage rates
increased with tailwinds, calm conditions, and
eastern and western crosswinds and decreased with
wind speed. Fog and low ceiling height, however,

Discussion
Ecogen Nocturnal Bird and Bat Migration, 2004 20
occurred rarely during this study (n = 3 nights with
fog; n = 2 nights with ceiling height <150 m agl),
and their influence on passage rates could not be
determined. The variables identified as important
in this study are generally consistent with results of
other studies (Lowery 1951, Gauthreaux 1971;
Able 1973, 1974; Blokpoel and Gauthier 1974;
Richardson 1990; Mabee et al. 2004).
FLIGHT ALTITUDES
Radar studies have shown that wind is a key
factor in migratory flight altitudes (Alerstam
1990). Birds fly mainly at heights at which
headwinds are minimized and tailwinds are
maximized (Bruderer et al. 1995). Because wind
strength generally increases with altitude, bird
migration generally takes place at lower altitudes
in headwinds and at higher altitudes in tailwinds
(Alerstam 1990). Most studies (all of those cited
above except Bellrose 1971) have found that
clouds influence flight altitude, but the results are
not consistent among studies. For instance, some
studies (Bellrose and Graber 1963, Hassler et al.
1963, Blokpoel and Burton 1975) found that birds
flew both below and above cloud layers, whereas
others (Nisbet 1963, Able 1970) found that birds
tended to fly below clouds.
In this study, flight altitudes increased with
tailwinds, calm conditions, western crosswinds,
and date (higher mean altitudes later in the season),
consistent with findings of Alerstam (1990).
Because of the rare occurrence of fog and low
ceiling height during this study we could not
ascertain possible relationships between these
conditions and flight altitudes. The need to
understand how birds respond to foggy conditions
is warranted, however, as the largest single-night
kill for nocturnal migrants at a wind power project
occurred on a foggy night during spring migration,
when 27 passerines fatally collided with a turbine
near a lit substation at the Mountaineer wind power
development in West Virginia (Kerlinger 2003).
Fatality events of this magnitude are rare at wind
power developments, although large kills of
migratory birds have sporadically occurred at
other, taller structures (e.g., guyed and lighted
towers >130 m high) in many places across the
country during periods of heavy migration,
especially on foggy, overcast nights in fall (Weir
1976, Avery et al. 1980, Evans 1998, Erickson et
al. 2001).
SPECIES COMPOSITION
Determination of species-specific risks to
nocturnal migrants requires the identification of
species migrating through the area of interest.
Flight speeds observed on surveillance radar (mean
= 10.1 ± 0.04 m/s) suggested that most of the avian
radar targets we observed in this study were
passerines, rather than faster-flying bird species
such as shorebirds or waterfowl. Furthermore, our
visual observations confirmed the presence of both
passerines and bats in the lower air layers (i.e.,
<150 m agl). The method used during the
beginning of the project (August 14–September 13,
2004), in which we used two 2,000,000-CP
spotlights with red lenses, provided only limited
information for estimating the proportion of birds
and bats within the entire rotor-swept area (i.e.,
<125 m agl), although the data collected during
that early period do provide some information on
the occurrence of birds and bats at lower altitudes
(i.e., <~75 m agl). We believe that the improved
method using night vision goggles from 14 to 29
September provides data that are adequate for
estimating the proportion of birds and bats within
~150 m agl (94% birds and 6% bats) for that time
period.
Most (86%) of the bat fatalities at wind power
developments and other tall structures occur during
mid-July to mid-September and involve long-range
migratory tree-roosting bat species such as Hoary
(Lasiurus cinereus), Eastern Red (Lasiurus
borealis), and Silver-haired (Lasionycteris
noctivagans) bats (Erickson et al. 2002, Johnson et
al. 2003, Erickson et al. 2004, Kerns 2004). Of the
seven bats observed during this study, two
appeared to be tree-roosting bats. In general,
fatality rates of bats are much lower in the central
and western US (Erickson et al. 2002) than in the
eastern US, where substantial bat kills have been
observed along an Appalachian ridgeline in West
Virginia and Pennsylvania (Erickson 2004, Kerns
2004).

Literature Cited
21 Ecogen Nocturnal Bird and Bat Migration, 2004
TARGETS WITHIN THE PROPOSED
TURBINE AREA
We estimated a turbine passage rate of 51–362
nocturnal migrants passing within the area
occupied by each proposed turbine at the
Prattsburgh–Italy Wind Power project during our
45-d fall study period. Our late-season estimate of
96% birds and 4% bats may underestimate the
proportion of bats over the whole fall season, based
upon the timing of bat fatalities at other wind power
sites earlier in the season (Johnson 2004).
Regardless, our estimated turbine passage rate
provides a starting point for developing a complete
avian risk assessment; however, our estimate must
be combined with an estimate of the proportion of
migrants that (1) do not collide with turbines
because of their avoidance behavior and (2) safely
pass through the turbine blades by chance alone —
a proportion that will vary with the speed at which
turbine blades are turning as well as the flight
speeds of individual migrants. Once this
information is known, one may be able to assess the
likelihood of avian and bat fatalities at proposed
wind power projects. The proportion of nocturnal
migrants that detect and avoid turbines is currently
unknown in the US (but see Winkleman 1995 for
studies in Europe), and there are no empirical data
that predict a species’ ability to pass safely through
the rotor-swept area of a turbine (but see Tucker
1996 for a hypothetical model). We speculate,
however, that the values are high for both of these
missing pieces of information, considering the
relatively low avian fatality rates at wind power
developments in the US (Erickson et al. 2002).
CONCLUSIONS
This study focused on nocturnal migration
patterns and flight behaviors during the peak
periods of fall passerine and bat migration at the
proposed Prattsburgh–Italy Wind Power project in
New York. The key results of our of fall passerine
and bat migration study were: (1) the mean overall
passage rate was moderate (i.e., 200 targets/km/h);
(2) mean nightly passage rates ranged from 18 to
863 targets/km/h; (3) the percentage of targets
passing below 125 m agl (9%) was similar to that
for a small number of comparable studies; (4) an
estimated turbine passage rate of 51–362 nocturnal
migrants passing within the airspace occupied by
each proposed turbine during the 45-d fall
migration season; and (5) migrants flying below
150 m agl consisted of ~94% birds and ~6% bats
during the late sampling period (i.e., mid- to late
September).
LITERATURE CITED
Able, K. P. 1970. A radar study of the altitude of
nocturnal passerine migration. Bird-Banding
41: 282–290.
Able, K. P. 1973. The role of weather variables and
flight direction in determining the magnitude
of nocturnal migration. Ecology 54:
1031–1041.
Able, K. P. 1974. Environmental influences on the
orientation of free-flying nocturnal bird
migrants. Animal Behaviour 22: 224–238.
Alerstam, T. 1990. Bird migration. Cambridge
University Press, Cambridge, United
Kingdom. 420 pp.
Anderson, R., M. Morrison, K. Sinclair, and D.
Strickland. 1999. Studying wind energy/bird
interactions: a guidance document. Metrics
and methods for determine or monitoring
potential impacts on birds at existing and
proposed wind energy sites. National Wind
Coordinating Committee, Washington, DC.
87 pp. Available at
http://www.nationalwind.org/pubs/default.htm
Avery, M. L., P. F. Springer, and N. S. Dailey. 1980.
Avian mortality at manmade structures: an
annotated bibliography (revised). U.S. Fish
and Wildlife Service, Biological Services
Program, Report No. FWS/OBS–80/54.
152 pp.
Bellrose, F. C. 1967. Radar in orientation research.
Proceedings XIV International Ornithological
Congress. pp. 281–309.
Bellrose, F. C. 1971. The distribution of nocturnal
migration in the air space. Auk 88: 397–424.
Bellrose, F. C. 1976. Ducks, geese, and swans of
North America. 2nd ed. Stackpole Books,
Harrisburg, PA. 540 pp.

Literature Cited
Ecogen Nocturnal Bird and Bat Migration, 2004 22
Bellrose, F. C., and R. R. Graber. 1963. A radar
study of flight directions of nocturnal
migrants. Proceedings XIII International
Ornithological Congress. pp. 362–389.
Blokpoel, H. 1971. The M33C track radar (3-cm)
as a tool to study height and density of bird
migration. Canadian Wildlife Service Report
Series 14: 77–94.
Blokpoel, H., and M. C. Gauthier. 1974. Migration
of lesser Snow and Blue geese in spring
across southern Manitoba, Part 2: Influence of
weather and prediction of major flights.
Canadian Wildlife Service Report Series 32:
1–28.
Blokpoel, H., and J. Burton. 1975. Weather and the
height of nocturnal migration in east-central
Alberta: a radar study. Bird-Banding 46:
311–328.
Bruderer, B., and P. Steidinger. 1972. Methods of
quantitative and qualitative analysis of bird
migration with a tracking radar. Pages
151–168 in Galler, S.R., Schmidt-Koenig, K.,
Jacobs, G.S., Belleville, R.E., eds. Animal
orientation and navigation: a symposium.
NASA SP262. U.S. Government Printing
Office, Washington, DC.
Bruderer, B., T. Steuri, and M. Baumgartner. 1995.
Short-range high-precision surveillance of
nocturnal migration and tracking of single
targets. Israeli Journal of Zoology 41:
207–220.
Bruderer, B., and A. Boldt. 2001. Flight
characteristics of birds, I. Radar
measurements of speeds. Ibis 143: 178–204.
Buffalo Ornithological Society. 2002. Seasonal
checklist of birds for the Niagara Frontier
Region. Buffalo Ornithological Society, Inc.,
Buffalo, NY. 21 pp.
Bull, J. L. 1985. Birds of New York State. Cornell
University Press, Ithaca, NY. 703 pp.
Burnham, K. P., and D. R. Anderson. 1998. Model
selection and inference: a practical
information-theoretic approach.
Springer-Verlag, New York, NY. 353 pp.
Cooper, B. A., R. H. Day, R. J. Ritchie, and C. L.
Cranor. 1991. An improved marine radar
system for studies of bird migration. Journal
of Field Ornithology 62: 367–377.
Cooper, B. A., and R. J. Ritchie. 1995. The altitude
of bird migration in east-central Alaska: a
radar and visual study. Journal of Field
Ornithology 66: 590–608.
Cooper, B. A., C. B. Johnson, and E. F. Neuhauser.
1995a. The impact of wind turbines in upstate
New York. Pages 607–611 in LGL Ltd. ed.
Proceedings of National Avian-Wind Power
planning meeting II September 1995,
Washington, DC. Available at
http://www.nationalwind.org/pubs/avian95/
default.htm
Cooper, B. A., C. B. Johnson, and R. J. Ritchie.
1995b. Bird migration near existing and
proposed wind turbine sites in the eastern
Lake Ontario region. Unpublished report
prepared for Niagara–Mohawk Power
Corporation, Syracuse, NY, by ABR, Inc.,
Forest Grove, OR. 71 pp.
Cooper, B. A., and T. J. Mabee. 2000. Bird
migration near proposed wind turbine sites at
Wethersfield and Harrisburg, New York.
Unpublished report prepared for
Niagara–Mohawk Power Corporation,
Syracuse, NY, by ABR, Inc., Forest Grove,
OR. 46 pp.
Cooper, B. A., A. A. Stickney, and T. J. Mabee.
2004. A radar study of nocturnal bird
migration at the proposed Chautauqua wind
energy facility, New York, fall 2003.
Unpublished report prepared for Chautauqua
Windpower LLC, Lancaster, NY, by ABR,
Inc.—Environmental Research & Services,
Forest Grove, OR. 26 pp.
Day, R. H., and L. C. Byrne. 1990. Avian research
program for the Over-the-Horizon Backscatter
Central Radar System: radar studies of bird
migration, fall 1989. Unpublished report
prepared for Metcalf & Eddy/Holmes &
Narver, Wakefield, MA, by Alaska Biological
Research, Inc., Fairbanks, AK. 102 pp.

Literature Cited
23 Ecogen Nocturnal Bird and Bat Migration, 2004
Diehl, R. H., R. P. Larkin, and J. E. Black. 2003.
Radar observations of bird migration over the
Great Lakes. Auk 120: 278–290.
Eastwood, E. 1967. Radar ornithology. Methuen
and Co., Ltd., London, United Kingdom.
278 pp.
Eastwood, E., and G. C. Rider. 1965. Some radar
measurements of the altitude of bird flight.
British Birds 58: 393–426.
Erickson, W. P., G. D. Johnson, M. D. Stickland, D.
P. Young, Jr., K. J. Sernka, and R. E. Good.
2001. Avian collisions with wind turbines: a
summary of existing studies and comparisons
to other sources of avian collision mortality in
the United States. National Wind
Coordinating Committee, Washington, DC.
62 pp. Available at
http://www.nationalwind.org/pubs/default.htm
Erickson, W. P., G. D. Johnson, D. P. Young, Jr., M.
D. Stickland, R. E. Good, M. Bourassa, K.
Bay, and K. Sernka. 2002. Synthesis and
comparison of baseline avian and bat use,
raptor nesting, and mortality information from
proposed and existing wind developments.
Unpublished report prepared for Bonneville
Power Administration, Portland, OR, by
WEST, Inc., Cheyenne, WY. 124 pp.
Erickson, W. 2004. Patterns of daily mortality
searches at Meyersdale, Pennsylvania. Talk
presented at the National Wind Coordinating
Committee meeting, “Onshore wildlife
interactions with wind developments:
Research Meeting V”, 3–4 November 2004,
Lansdowne, VA. Available at
http://www.nationalwind.org/events/wildlife/2
00411/presentations.htm.
Evans, B. 1998. Deadly towers. Living Bird
17(2): 5.
Farnsworth, A., S. A. Gauthreaux, Jr., and D. van
Blaricon. 2004. A comparison of nocturnal
call counts of migrating birds and reflectivity
measurements on Doppler radar. Journal of
Avian Biology 35:365–369.
Gauthreaux, S. A., Jr. 1971. A radar and direct
visual study of passerine spring migration in
southern Louisiana. Auk 88: 343–365.
Gauthreaux, S. A., Jr. 1972. Behavioral responses
of migrating birds to daylight and darkness: a
radar and direct visual study. Wilson Bulletin
84: 136–148.
Gauthreaux, S. A. Jr. 1975. Radar ornithology: bird
echoes on weather and airport surveillance
radars. Clemson University Press, Clemson,
SC.
Gauthreaux, S. A., Jr. 1978. Migratory behavior
and flight patterns. Pages 12–26 in M. Avery,
ed. Impacts of transmission lines on birds in
flight. U.S. Fish and Wildlife Service, Office
of Biological Services, Report No.
FWS/OBS–78/48. 151 pp.
Gauthreaux, S. A., Jr. 1985a. Radar,
electro-optical, and visual methods of
studying bird flight near transmission lines.
Unpublished final report prepared for Electric
Power Research Institute, Palo Alto, CA, by
Clemson University, Clemson, SC. 76 pp.
Gauthreaux, S. A., Jr. 1985b. An avian mobile
research laboratory: hawk migration studies.
Pages 339–346 in M. Harwood, ed.
Proceedings of Hawk Migration Conference
IV. Hawk Migration Association of North
America, Washington, CT.
Gauthreaux, S. A., Jr. 1991. The flight behavior of
migrating birds in changing wind fields: radar
and visual analyses. American Zoologist 31:
187–204.
Graber, R. R., and S. S. Hassler. 1962. The
effectiveness of aircraft-type (APS) radar in
detecting birds. Wilson Bulletin 74: 367–380.
Harmata, A. R., K. M. Podruzny, J. R. Zelenak, and
M. L. Morrison. 1999. Using marine
surveillance radar to study bird movements
and impact assessment. Wildlife Society
Bulletin 27: 44–52.
Hassler, S. S., R. R. Graber, and F. C. Bellrose.
1963. Fall migration and weather: a radar
study. Wilson Bulletin 75: 56–77.

Literature Cited
Ecogen Nocturnal Bird and Bat Migration, 2004 24
Johnson, G. D. 2004. A review of bat impacts at
wind farms in the U.S. Presentation at the
National Wind Coordinating Committee
meeting, “Wind energy and birds/bats
workshop: understanding and resolving bird
and bat impacts,” 17–19 May 2004,
Washi n gton, D C .
Johnson, G. D., W. P. Erickson, M. D. Strickland,
M. F. Shepherd, and D. A. Shepherd. 2003.
Mortality of bats at a large-scale wind power
development at Buffalo Ridge, Minnesota.
American Midland Naturalist 150: 332–342.
Kerlinger, P. 1995. How birds migrate. Stackpole
Books, Mechanicsburg, PA. 228 pp.
Kerlinger P. 2000. Avian mortality at
communication towers: a review of recent
literature, research, and methodology.
Unpublished report prepared for U.S. Fish and
Wildlife Service, Office of Migratory Bird
Management, by Curry & Kerlinger LLC,
Cape May Point, NJ. 38 p.
Kerlinger, P. 2003. FAA lighting of wind turbines
and bird collisions. Presentation at the NWCC
Wildlife Working Group Meeting, 18
November 2003, Washington DC. Available
at
http://www.nationalwind.org/events/wildlife/
20031117/presentations/default.htm
Kerlinger, P., and F. R. Moore. 1989. Atmospheric
structure and avian migration. Pages 109–141
in D. M. Power ed. Current Ornithology. Vol.
6.
Kerns, J. 2004. Patterns from daily mortality
searches at Backbone Mountain, West
Virginia. Presentation at the National Wind
Coordinating Committee meeting, “Onshore
wildlife interactions with wind developments:
Research Meeting V,” 3–4 November 2004,
Lansdowne, VA. Available at
http://www.nationalwind.org/events/wildlife/
200411/presentations.htm.
Kovach, W. 2003. Oriana version 2.0. Kovach
Computing Services, Anglesey, Wales, United
Kingdom.
Kunz, T. H., and M. B. Fenton. 2003. Bat ecology.
University of Chicago Press, Chicago, IL.
779 pp.
Larkin, R. P. 1991. Flight speeds observed with
radar, a correction: slow “birds” are insects.
Behavioral Ecology and Sociobiology 29:
221–224.
Liechti, F., M. Klaassen, and B. Bruderer. 2000.
Predicting migratory flight altitudes by
physiological migration models. Auk 117:
205–214.
Lowery, G. H., Jr. 1951. A quantitative study of the
nocturnal migration of birds. University of
Kansas Museum of Natural History 3:
361–472.
Mabee, T. J., and B. A. Cooper. 2004. Nocturnal
bird migration in northeastern Oregon and
southeastern Washington. Northwestern
Naturalist. 85: 39–47.
Mabee, T. J., B. A. Cooper, and J. H. Plissner.
2004. A radar study of nocturnal bird
migration at the proposed Mount Storm
wind-power development, West Virginia, fall
2003. Unpublished report prepared for
Western Ecosystems Technology, Inc.,
Cheyenne WY, and Nedpower US LLC,
Chantilly, VA by ABR, Inc., Forest Grove,
OR. 40 pp.
Manville A. M. 2000. The ABCs of avoiding bird
collisions at communication towers: the next
steps. In Proceedings of the December 1999
workshop on avian interactions with utility
structures, Charleston, SC. Electric Power
Research Institute, Palo Alto, CA. Report No.
1000736.
Nisbet, I. C. T. 1963. Measurements with radar of
the height of nocturnal migration over Cape
Cod, Massachusetts. Bird-Banding 34: 57–67.
Osborn R. G., K. F. Higgins, R. E. Usgaard, C. D.
Dieter, and R. D. Neiger. 2000. Bird mortality
associated with wind turbines at the Buffalo
Ridge Wind Resource Area, Minnesota.
American Midland Naturalist 143: 41–52.
Richardson, W. J. 1971. Spring migration and
weather in eastern Canada: a radar study.
American Birds 25: 684–690.

25 Ecogen Nocturnal Bird and Bat Migration, 2004
Richardson, W. J. 1972. Autumn migration and
weather in eastern Canada: a radar study.
American Birds 26: 10–16.
Richardson, W. J. 1978. Timing and amount of bird
migration in relation to weather: a review.
Oikos 30: 224–272.
Richardson, W. J. 1990. Timing of bird migration
in relation to weather: updated review. Pages
79–100 in E. Gwinner, ed. Bird migration.
Springer–Verlag, Berlin, Germany.
Skolnik, M. I. 1980. Introduction to radar systems.
McGraw-Hill, New York, NY. 581 pp.
SPSS. 2003. SPSS for Windows, version 12.0.
SPSS, Inc., Chicago, IL.
Tucker, V. A. 1996. A mathematical model of bird
collisions with wind turbine rotors. ASME
Journal of Solar Energy Engineering 118:
253–262.
Tuttle, M. D. 1988. America’s neighborhood bats.
University of Texas Press, Austin, TX. 96 pp.
USGS (U.S. Geological Survey). 2003.
Physiographic regions of the US. Available at
http://tapestry.usgs.gov/physiogr/physio.html.
Weir, R. D. 1976. Annotated bibliography of bird
kills at man-made obstacles: a review of the
state of the art and solutions. Canadian
Wildlife Service, Ottawa, ON, Canada. 85 pp.
Williams, T. C., J. Settel, P. O’Mahoney, and J. M.
Williams. 1972. An ornithological radar.
American Birds 26: 555–557.
Winkelman, J. E. 1995. Bird/wind turbine
investigations in Europe. Pages 43–47 and
110–140 in LGL Ltd. ed. Proceedings of
National Avian–Wind Power Planning
Meeting I, Lakewood, CO. Available at
http://www.nationalwind.org/pubs/avian94/
default.htm
Zalles, J. I., and K. L. Bildstein, eds. 2000. Raptor
Watch: a global directory of raptor migration
sites. Birdlife International, Cambridge,
United Kingdom, and Hawk Mountain
Sanctuary, Kempton, PA. Birdlife
Conservation Series No. 9.
Zar, J. H. 1999. Biostatistical analysis. 3rd ed.
Prentice-Hall, Inc., Englewood Cliffs, NJ.
663 pp.

Ecogen Nocturnal Bird and Bat Migration, 2004 26
Appendix 1. Calculation of the turbine passage rate (the number of targets that would pass within the
area occupied by each proposed turbine) over the entire 45-day fall 2004 study period at
the Prattsburgh–Italy Wind Power project, NY...
Calculation parameter
WIND-TURBINE CHARACTERISTICS
(A) Total turbine height (m) 119
(B) Blade radius (m) 39
(C) Height below blade (m) 41
(D) Approximate front-to-back width (m) 6
(E) Minimal (side profile) area (m2) = A × D 714
(F) Maximal (front profile) area (m2) = (C × D) + (π × B2) 5,024
PASSAGE RATE
(G) Mean rate below 125 m agl (targets/km/h) 20.0
(H) Area sampled below 125 m agl = 125 x 1,000 (m2) 125,000
(I) Mean passage rate per unit area (targets/m2/h) = G/H 0.00016
TURBINE PASSAGE RATE
(J) Duration of study period (# nights) 45
(K) Mean number of hours of darkness (h/night) 10
(L) Minimal number of targets/h within turbine area = E × I 0.11424
(M) Maximal number of targets/h within turbine area = F × I 0.80390
(N) Minimal number of targets within turbine area (side profile) during 45-night period = J x K × L 51
(O) Maximal number of targets within turbine area (front profile) during 45-night period = J x K × M 362