Content uploaded by Ruben Fijn
Author content
All content in this area was uploaded by Ruben Fijn
Content may be subject to copyright.
97
© Wildfowl & Wetlands Trust Wildfowl (2012) 62: 97–116
Habitat use, disturbance and collision risks for
Bewick’s Swans Cygnus columbianus bewickii
wintering near a wind farm in the Netherlands
RUBEN C. FIJN1*, KAREN L. KRIJGSVELD1, WIM TIJSEN2,
HEIN A.M. PRINSEN1& SJOERD DIRKSEN1
1Bureau Waardenburg, P.O. Box 365, 4100 AJ Culemborg, Netherlands.
2De Dolven 39, 1778 JP Westerland, Netherlands.
*Correspondence author. E-mail: r.c.fijn@buwa.nl
Abstract
Each winter ~ 30% of the Northwest European Bewick’s Swan Cygnus columbianus
bewickii population feeds in Polder Wieringermeer, the Netherlands, on waste crops left
after the harvest. The area has also become important for generating energy as a result
of wind farm development. This study analyses pre- and post-construction data on
Bewick’s Swan distribution, movements and foraging behaviour in the vicinity of a
nine-turbine wind farm site, in order to determine the effects of wind turbines on
wintering swans. The swans’ flight-lines between feeding areas and the roost were
recorded visually and using radar over 10 evenings in good weather conditions. Food
availability on different agricultural plots appeared to be an important factor explaining
swan numbers and distribution in the area. In circumstances with even food availability
early in the season, swans showed a preference for foraging in areas further away from
the turbines, indicating some displacement caused by the turbines. Nevertheless, swans
increasingly fed closer to the wind turbines during the course of the season in response
to food availability. The likelihood that a single Bewick’s Swan passing through the
wind farm will collide with a turbine (collision risk) at the nine-turbine site, determined
from swan movements through the wind farm (number of swan flights per unit length
per unit time) and from regular searches for carcasses, was estimated at 0–0.04% in
winter 2006/2007. Avoidance behaviour was observed, with birds navigating around
and between the lines of turbines. The observed disturbance of foraging birds early in
the season, the acquired knowledge of avoidance responses, and the calculated
collision rates in this study can be used for future assessments during planning and
construction of new wind farms in wintering areas of Bewick’s Swans, especially in
areas where important congregations of world or flyway populations occur.
Key words: barrier effects, Bewick’s Swan, collision, disturbance, Wieringermeer,
wind turbines.
98 Effects of wind farm on Bewick’s Swans
© Wildfowl & Wetlands Trust Wildfowl (2012) 62: 97–116
Government plans to reduce carbon
emissions to slow down global climate
change, include increasing the capacity to
generate energy from renewable sources
such as wind and tide. A primary area for the
production of wind energy in the
Netherlands is Polder Wieringermeer, in the
northwest part of the country. Traditionally
this area was cultivated agricultural land, but
nowadays it is also increasingly used for
generation of renewable energy. By 2010,
several wind farms had been built in the
area, with a total of 54 large turbines (>1
MW, hub height above 70 m) installed, along
with 36 smaller solitary turbines (0.85 MW,
hub height ~ 50 m). These ninety turbines
generate a total of 106 MW of electricity
but new turbines planned for the future will
increase the capacity to 400 MW.
The Northwest European population of
Bewick’s Swans Cygnus columbianus bewickii
has decreased substantially in numbers since
the mid 1990s. It was estimated at 21,500
individuals in January 2005, and national
trend indices indicate a further decline since
then (Rees & Beekman 2010). The swans
breed in arctic Russia and a large proportion
of the population winters in the UK and
the Netherlands. Polder Wieringermeer is an
internationally important wintering area for
the species, with counts indicating that
25–33% of the population use the site each
winter. The polder provides feeding grounds
in close proximity to roosting places, and the
birds are able to feed on crop remains
(mainly sugar beet) left after the harvest,
typically from November onwards (van Gils
& Tijsen 2007).
Previous studies have discussed three
main ways in which wind turbines can affect
bird populations: through the disturbance
and displacement of foraging and resting
birds, by flying birds colliding with the
turbines, and by the turbines potentially
acting as a barrier during flight (Langston &
Pullan 2003; Dirksen et al. 2007; Percival
2007; Drewitt & Langston 2008). Wind
farms are known to have negative effects on
some species (e.g. Madders & Whitfield
2006; Thelander & Smallwood 2007), but
more detailed understanding of species-
specific responses to the turbines is required
for an adequate assessment of the impact
of the turbines on bird populations.
Research into the disturbance and
displacement of birds has mostly focussed
on changes in numbers at turbine locations
(i.e. calculated a species-specific ‘disturbance
distance’, e.g. Winkelman 1989; Schreiber
1993; Kruckenberg & Jaene 1999), but
disturbance of foraging and resting
waterbirds can also result in changes in
physiology, behaviour and habitat choice
(e.g. Orloff & Flannery 1992; Kruckenberg
& Jaene 1999). Swans are potentially at risk
of collisions because Whooper Swans are
known to fly at altitudes of 5–45 m during
commuting flights to feeding areas (Larsen
& Clausen 2002). The collision risk (i.e. the
probability that a given bird flying through
the wind farm will collide with a turbine) is
a combination of the probability of
collision and the movement of birds
through the wind farm area (cf. Desholm et
al. 2006; Band et al. 2007). In general, the
number of birds that collide with a turbine
in a specific wind farm per unit time (i.e. the
collision rate) differs between studies.
Across species and locations, previously
found collision rates range from 3.7–58
Effects of wind farm on Bewick’s Swans 99
© Wildfowl & Wetlands Trust Wildfowl (2012) 62: 97–116
birds per year per turbine (e.g. Winkelman
1989; Winkelman 1992; Everaert & Stienen
2007). This rate depends on a range of
factors including the number of birds flying
through the area, the location and lay-out of
the wind farm, landscape features, and the
behaviour and physiology of the species
(Thelander et al. 2003; Dirksen et al. 2007; de
Lucas et al. 2008; Drewitt & Langston 2008;
Martin 2011). The mortality rate and
collision risk for Bewick’s Swans have been
modelled previously for a wind farm at
Cheyne Court in the UK. Here, collision risk
was estimated at 0.145 % of bird passages,
with a mortality rate of 0.06 swans over 180
days, but it should be noted that the study
used an avoidance rate of 0.9962 from
observations made mainly of gulls (Painter
et al. 1999) which have different flight
characteristics (Chamberlain et al. 2006).
To the best of our knowledge, the study
presented here is the first before/after
assessment of the possible impact of wind
turbines on Bewick’s Swans at a wintering
site. We used pre- and post-construction
data to study whether the installation of
multiple new wind turbines coincided with a
change in Bewick’s Swan numbers,
distribution and habitat choice in the area.
Furthermore, collision risk was assessed for
Bewick’s Swans at the site from a calculated
collision rate and from measures of flight
intensity through the area covered by the
wind farm.
Methods
Study area
Between February 2003 (start of first
building activities) and July 2006 (opening
and first month of full operation), the
Energy Research Centre of the Netherlands
(ECN) built a wind farm in the spring and
summer months in Polder Wieringermeer
(52°49'54"N, 5°04'50"E) in one of the
agricultural areas used by large numbers of
wintering Bewick’s Swans. This farm
consists of two lines of different types of
turbines positioned west–east with a
northern row of five and a southern row of
four turbines. All turbines were rated > 2.3
MW with an average hub height of 90 m
and a rotor diameter of 100 m (i.e. a rotor
sweep area of 40–140 m above ground
level). Turbines in the northern row are on
average 300 m apart whereas turbines in
the southern row are ~ 400 m apart, and
the two rows are 1,600 m apart. Small red
lights shine during darkness on top of the
hub.
The study area (~ 1,860 ha) around the
ECN turbines was divided into two
contiguous parts: the wind farm area (~ 770
ha) in which the new wind farm was built,
and an adjacent unchanged area (~ 1,090 ha)
with no new turbines, hereafter referred to
as the ‘control’ area (see Fig. 4 for an outline
of the study area). Some solitary wind
turbines were present near farms (3 in the
wind farm area; 6 in the unchanged area) in
the study area. These were installed several
years before the study commenced and were
smaller (maximum height reached by the
rotors = ~ 80 m) than the new wind farm
turbines.
Displacement of swans from their
feeding areas
Surveys of the study area were conducted at
around midday on a near daily basis in the
100 Effects of wind farm on Bewick’s Swans
© Wildfowl & Wetlands Trust Wildfowl (2012) 62: 97–116
winter, prior to construction (from 23
October 2000 until 7 March 2001), and
again after construction (from 27 October
2006 until 25 January 2007), to determine
whether the swans were displaced from
some of their feeding areas. The number of
wintering swans present was recorded on
each occasion, together with their
distribution across the site and foraging
behaviour. Swan numbers and distribution
were also recorded in winters 2003/04 to
2005/06 inclusive, but these surveys were
part of the monthly waterbird counts
undertaken in the Netherlands, so were less
frequent than in winters 2000/2001 and
2006/2007. Nevertheless, they provide a
good indication of the numbers of swans
present for each winter between the two
study seasons. The distance from each
group of swans (taken from the centre of
the group) to the nearest turbine was
measured using ArcGIS for each of the
count days.
The swans’ favoured food in the Polder
Wieringermeer (mainly waste sugar beet
and, to a lesser extent, carrots and potatoes)
was available only between harvest and
ploughing, the length of this period being
determined by the farmers (Dirksen et al.
1991; W. Tijsen unpubl. data). Food
availability in the study area was recorded
during 2006/2007 (but not in 2000/2001)
by mapping the different crop types on a
field-by-field basis, keeping track of the
harvest and noting the ploughing dates. By
doing so, the total number of hectares of
sugar beet fields was recorded. From
farming records the total number of
hectares of sugar beet fields in the study
area in 2000/2001 could be determined.
Quantification of Bewick’s Swan
flights
The movements of swans passing the wind
farm area during flights to and from night-
time roosts were recorded visually and with
radar. The use of radar provided precise
information on flight behaviour (flight-
lines) through and around the wind farms,
as well as quantifying the number of flights,
particularly during hours of darkness when
visual observations were not possible. The
radar system used was an X-band marine
surveillance radar with a peak power of 12
kW (Furuno FR1510 MARK–3, X-band
pulse repeat frequency 9,410 ± 30 MHz,
vertical beam width 20°, rotation speed 24
rpm, supplied by Radio Holland Rotterdam)
mounted on a 2 m high tripod. Radar range
was set to 2.8 km to cover the entire study
area. Due to lower detection probability at
the outer limit of the radar range, effectively
a circle around the radar with a radius of 2.5
km (19.6 km2) was sampled. The radar
system was positioned 0.8–1.5 km from the
turbines and the radar thus reached a
minimum of ~ 1 km beyond the turbines.
Bewick’s Swan movements in the study
area were monitored using radar over five
evenings and the following mornings in
winter 2000/01 from four hours around
sunset and four hours around sunrise (two
hours before until two hours after, in each
case), to provide the Environmental Impact
Assessment (EIA) of the proposed wind
farm with baseline data on the flight-paths
taken by the birds. Fieldwork was also
carried out on seven evenings in 2006/2007,
in differing but albeit generally good
weather conditions for the time of year
Effects of wind farm on Bewick’s Swans 101
© Wildfowl & Wetlands Trust Wildfowl (2012) 62: 97–116
(temperature = 6°–15°C, wind direction =
S–SW; wind speed = 3–7 Bft; cloud cover =
4/8 to 8/8; precipitation = dry, with only
occasional showers), from approximately
2 h before to 2 h after sunset. The departure
of different groups of Bewick’s Swans from
the fields to roosting areas on nearby Lake
IJsselmeer was highly synchronised and
occurred over a relatively short period of
time. Observations continued until all
swans, as determined by the swan survey
earlier that day, had left the study area for
the roost. In case of poor visibility (due to
darkness), species identification was
determined from the birds’ flight calls and
the characteristic behaviour (size and speed)
of echoes on the radar screen. If a potential
group of swans seen on the radar was out of
audible range, one of the field observers
was directed towards the flying group to
confirm species identification. Swan
movements were also recorded on three
additional evenings in 2006/2007 at a
second wind farm in Polder Wieringermeer
(‘Waterkaaptocht’; 52°51'46"N, 5°02'22"E;
~ 4 km from the study area), which has eight
similar 2.3 MW turbines in one line (see
Krijgsveld et al. 2009), to increase the
number of flight records.
All bird tracks observed in the field were
digitised and, if positively identified by field
observers, flight-path specifications (i.e. date,
time, species, number of birds and altitude
of flight) were stored in an ArcGIS database.
This database was used to produce maps of
the swans’ flight-lines within and around the
boundaries of the wind farms. The detailed
flight data made it possible to calculate the
proportion of the swans present in the study
area and in adjacent feeding areas that passed
through the wind farm during flights to the
roost. About 30 min after sunset observers
were not able to observe flying swans in the
field; however, at close distances, structures
such as wind turbines might still be visible
to flying swans, especially when some
background illumination is present. As the
exact extent of this phenomenon is
unknown, we decided in this study to set the
boundary between dark and light at 30 min
after sunset, in other words when observers
encountered reduced visibility. A diversion
from the intended flight-path was defined as
occasions when a swan discontinued its
flight direction, in either the horizontal or
vertical plane. These avoidance records were
used in calculating swan movement (‘flux’)
through the wind farm, which in turn was
used to determine collision risk as described
below.
Collision risk
The collision risk for Bewick’s Swans in the
study area was calculated by dividing the
collision rate by the flux (i.e. the number of
Bewick’s Swan per area [m2] within the wind
farm per unit time). Collision rate was
investigated by regular searches for corpses
combined with corpse disappearance rate
experiments. Between 27 November 2006
and 2 February 2007, the area below the
turbines in the study area and also at the
Waterkaaptocht wind farm was searched for
collision victims at 2–3 day intervals. The
additional wind farm was included to
increase the probability of finding collision
victims, as earlier research found that the
frequency with which birds collided with
turbines was low (e.g. Winkelman 1992;
Krijgsveld et al. 2009). The area within a
102 Effects of wind farm on Bewick’s Swans
© Wildfowl & Wetlands Trust Wildfowl (2012) 62: 97–116
radius of 100 m of each turbine was
searched, on the basis that previous studies
found that victims fall within a radius of up
to 1.1 times the hub height of the turbine
(Winkelman 1992; Grünkorn et al. 2009), i.e.
up to 88 m in this study. Swans are expected
to fall at even closer range due to their high
body mass (Krijgsveld et al. 2009). The
searched area (100 m radius) therefore was
considered large enough to include all
potential victims. We included only those
turbines under which vegetation type and
height did not obstruct visibility of potential
victims. Nevertheless, the total searched area
in the winter of 2006/2007 was 15,697,457
m2(98.6%) of a total area of 15,927,874 m2
around the turbines in both wind farms. The
area below a turbine was searched either
with binoculars from the base of the turbine
(ECN wind farm) or by walking in parallel
lines 4–6 m apart (Waterkaaptocht, see
Krijgsveld et al. 2009), depending on
visibility of potential victims. Because swans
are conspicuous, with their large size and
white colour, a detection probability of
100% was assumed. All victims found
during the searches were recorded,
photographed and sent to the Dutch
veterinary laboratory CIDC-Lelystad for post
mortem examination (internally and
externally) to determine the cause of death.
Scavenging predators, such as Common
Buzzard Buteo buteo and Red Fox Vulpes
vulpes, roam the study area and might
remove swan corpses during the study
period, resulting in underestimates of
collision rates. To determine the
disappearance rate, seven defrosted
carcasses were laid out in the study area (1
Brent Goose Branta bernicla, 4 Bewick’s
Swans and 2 Mute Swans Cygnus olor), placed
semi-randomly in all directions at distances
of 1–100 m from the turbines. Turbines
used for the disappearance test were not
used in victim searches, to avoid predators
and scavengers being attracted to the
former, which could lead to an increase in
disappearance of collision victims. Presence
and condition (eaten, moved, buried) of
carcasses were registered for two weeks after
carcasses had been laid out. The probability
that a carcass remained at a location was
calculated as the probability that a carcass
present on day twas still present at day t + 1,
day t + 2, etc. Calculations were similar to
those undertaken by Winkelman (1992) to
facilitate comparison with other studies.
The number of collision victims,
corrected for observer efficiency and
disappearance rate (Nc), was determined by
correcting the number of victims found
(Nf), for the probability that a victim
remains at the location rather than
disappearing through scavenging (Pd), the
probability of finding a victim (Pf), the
fraction of the total area (100 m radius)
underneath the turbine that was searched
(Fs), and for the fraction of days of the
research period that victims were searched
for (Fd). The corrected number of collision
victims used to calculate the collision rate
for swans within the whole wind farm was
thus calculated as follows (following
Winkelman 1992): Nc = Nf / (Pd × Pf ×
Fs × Fd).
Statistical analysis
Data were analysed using SPSS version 15.0.
Changes in swan numbers wintering in the
study area over the years were calculated as a
Effects of wind farm on Bewick’s Swans 103
© Wildfowl & Wetlands Trust Wildfowl (2012) 62: 97–116
proportion of the total number recorded
across Polder Wieringermeer (Fig. 1). The
numbers of swans in the wind farm area and
the adjacent unchanged (‘control’) area did
not follow a normal or a Poisson distribution
(Figs. 1 & 2); non–parametric statistics (Chi-
square test and Spearman Rank correlation)
therefore were used to analyse these data.
Distance to the nearest turbine in relation to
date (Fig. 3) was analysed using a logarithmic
regression. Linear regressions on arcsine
transformed proportionate data were used to
model the carcass disappearance rate (Fig. 5).
Mean values are given ± s.d. unless otherwise
stated.
Results
Swan numbers during the winter
Bewick’s Swans were present in the study
area from 23 October to 7 March in winter
2000/01 and from 1 November to 28
January in winter 2006/07. The maximum
numbers counted in the study area (i.e. in
both the ‘control’ and the wind farm areas)
were significantly lower in 2006/07 than in
2000/01 (χ21= 36.9, P< 0.001; Table 1).
The shorter period that swans were present
in the area, in combination with the lower
peak counts, resulted in fewer swan-days
being recorded in the year following
construction than beforehand (χ21= 128.6,
P< 0.001). The wind farm area and the
adjacent ‘control’ area showed a similar
decrease in the total number of swan-days,
but the seasonal maximum count decreased
more substantially within the wind farm site
(Table 1).
In contrast, the maximum number of
birds present across the whole of Polder
Wieringermeer was higher after construction
(Table 1). Pre-construction, in 2000/2001,
up to 89% of the winter’s maximum
number of swans counted across Polder
Wieringermeer was found in the wind farm
area and 70% in the adjacent ‘control’ area.
Post-construction, in 2006/2007, these
percentages decreased to 24% and 29%
respectively. The proportion of the total
number of birds in Polder Wieringermeer
that visited the study area decreased
significantly in the years between 2000/2001
and 2006/2007 (rs= –0.90, n = 5, P < 0.05;
Fig. 1).
The proportion of the Northwest
European Bewick’s Swan population
wintering in Polder Wieringermeer has
increased during the study, from 5% of the
total population in 2000/2001 to 11% in
2006/2007 (Rees & Beekman 2010; Table
1). In contrast, the study area within Polder
Wieringermeer appears to have become less
attractive with 5% of the Northwest
European population present in 2000/2001
and 3% in 2006/2007.
Swan feeding distribution
There was a within-winter shift in the
distribution of swans across the study area
in relation to variation in the availability of
sugar beet remains following the harvest in
the 2006/2007 season. Numbers in the
‘control’ area correlated significantly with
the number of hectares of fields with
food remains in that area during habitat
assessments (rs= 0.38, n = 85, P < 0.01), but
there was no significant association between
the number of swans and the number of
hectares with food remains in the wind farm
area (rs= 0.24, n = 64, n.s.). In 2000/2001
104 Effects of wind farm on Bewick’s Swans
© Wildfowl & Wetlands Trust Wildfowl (2012) 62: 97–116
such a shift was not observed and Bewick’s
Swans were present in both ‘control’ and
wind farm areas throughout the season.
The number of birds in the wind farm
area increased when the number of hectares
with available food decreased in the ‘control’
area (rs= –2.53, n = 85, P < 0.05; Fig. 2).
These results imply that, when food was
available on fields both with and without
turbines, the swans generally foraged in
the area without the newly-constructed
turbines. Up to 530 birds (95% of the peak
count for the study area in 2006/07) were
recorded on fields within the area with new
Table 1. Numbers of swan-days (sum of number of swans on each day of the field season,
on days when counts were missing, gaps in data were calculated as the average of the two
counts spanning the missing count) and seasonal maximum numbers in the ‘control’ and
wind farm areas in the two study seasons 2000/2001 and 2006/2007. Also shown are the
proportions of the total number of birds wintering in the Wieringermeer and of the total
Northwest European population. Changes in abundance between the two study seasons are
expressed as a percentage.
2000/2001 2006/2007 Change
Total number of swan-days
‘Control’ area 20,714 4,546 – 78%
Wind farm area 34,586 9,526 – 72%
Seasonal maximum count
‘Control’ area 860 550 – 36%
Wind farm area 1,099 530 – 52%
Polder Wieringermeer 1,230 2,233 + 82%
Proportion of Wieringermeer birds
‘Control’ area 0.70 0.29 – 59%
Wind farm area 0.89 0.24 – 73%
NW European population (Rees &Beekman 2010)
23,000 21,500* – 7%
Proportion of NW European population
Polder Wieringermeer 0.05 0.11 + 120%
Entire study area 0.05 0.03 – 40%
*Note that the Northwest European population figure for Bewick’s Swans described in Rees
& Beekman (2010) is based on the census of January 2005 and not 2007.
Effects of wind farm on Bewick’s Swans 105
© Wildfowl & Wetlands Trust Wildfowl (2012) 62: 97–116
turbines on 12 individual days in November,
but these were resting on grass and not
foraging. At the beginning of the 2006/
2007 season, when sugar beet remains were
available in both areas, the swans foraged
predominantly in the ‘control’ area. Later in
the season when most of the fields in the
‘control’ area had been ploughed, the swans
moved to the wind farm area and closer to
the turbines to utilise the sugar beet remains
that were still available there.
Bewick’s Swans foraged significantly
closer to the turbines as the season
progressed (Fig. 3; logarithmic regression of
distance of birds to the nearest turbine
versus date: F1,84 = 65.62, r2 = 0.44,
P < 0.001). This effect was attributable
mainly to a large number of birds feeding at
greater distances from the turbines at the
start of the season. Excluding these birds
from the analysis still resulted in a
significant, albeit smaller, decrease in the
distance of the swans from the turbines
as the winter progressed (F1,77 = 21.05,
r2 = 0.22, P < 0.001). The decrease in
distance was not due to the distribution of
harvested fields as the distance of harvested
fields to the turbines did not decrease
significantly during the course of the season
(linear regression of distance of fields to the
nearest turbine versus date: F1,32 = 0.39,
r2 = 0.01, P = n.s.). The distance between
foraging and resting Bewick’s Swans and the
turbines was on average 560 m (s.e. = 57.9,
n= 86), whereas the minimum recorded
distance was 125 m.
Figure 1. Maximum numbers of swans counted each winter in the study area and across the Polder
Wieringermeer. The proportion of the Bewick’s Swans wintering in Polder Wieringer meer recorded in
the study area is also illustrated. The wind farm was built during the summer months between summer
2003 and summer 2006.
106 Effects of wind farm on Bewick’s Swans
© Wildfowl & Wetlands Trust Wildfowl (2012) 62: 97–116
Swan flights in the study area
Swans flew towards the roosting sites in the
late evening and early night. At least 1,664
Bewick’s Swan flight-paths for 101 groups
flying to the roost were recorded in both
wind farms during eight out of ten
fieldwork evenings in 2006/2007 (flights
were not recorded during two evenings as
swans were absent from the study area and
no swans flew past from adjacent areas).
This is a minimum estimate of the total
number of swan flights as 33 groups were
recorded only as radar tracks in complete
darkness, > 30 minutes after sunset. The
birds giving these tracks could be identified
as Bewick’s Swans on the basis of flight calls
but group size could not be determined. A
minimum group size was estimated on these
occasions, based on the number of birds
counted by the field observer earlier in the
day. There was substantial variation in the
timing of the evening flights to the roost.
Of all groups of swans, 61 ± 41% (range:
0–100%, n= 7 nights, 101 groups) flew after
dark (> 30 min after sunset) each night in
2006/2007. Group size was limited to 16 ±
41 (range 1–300) birds at maximum. Of all
ha in Impact
n in Control
ha in Control
n in Impact
Figure 2. Numbers of swans in the ‘Control’ (no new large wind turbines) and wind farm ‘Impact’
(nine new large wind turbines) sections of the study area in relation to the availability of sugar beet
remains. Early in the winter, when waste sugar beet was available in both areas, Bewick’s Swans fed in
areas away from the turbines.
Effects of wind farm on Bewick’s Swans 107
© Wildfowl & Wetlands Trust Wildfowl (2012) 62: 97–116
individual swans flying towards the roosting
sites, 75 ± 35% (range: 0–100%, n= 7
nights, 1,664 birds) flew after dark. Birds
that flew past the outer edge of the wind
farm adjusted their flight direction at a
distance of a few hundred metres at
maximum (n= 562 birds). Of all swans
present in the area an average per day of
16 ± 22.5% (range: 0–65%) flew through
the wind farm during commuting flights
(Table 2).
In 2000/2001, Bewick’s Swans generally
flew in straight lines from fields where they
had been feeding during the day towards the
roost site (Lake IJsselmeer), although no
fixed flight-paths through the landscape
were identified. Foraging areas were similar
in 2006/2007 (albeit not identical to those
recorded in 2000/01, due to crop rotation
and a decrease in the area of sugar beet
available) and birds were seen to fly in a
similar direction to the roost. In 2006/
2007 birds adjusted their flight-paths to
the presence of the wind turbines during
both light and darkness; however, neither
large deflections around the entire wind
farm nor panic reactions in the air were
observed. Birds avoided turbines by
navigating around individual turbines and
between rows of turbines (as illustrated
for the evening of 24 November 2006 in
Fig. 4).
Figure 3. Distance of Bewick’s Swan flocks in the study area to the nearest turbine during the course
of winter 2006/07, from 1 November onwards (logarithmic regression with r2= 0.44).
108 Effects of wind farm on Bewick’s Swans
© Wildfowl & Wetlands Trust Wildfowl (2012) 62: 97–116
Collision rate estimates
Two Bew ick ’s Swan s we re fo und d ead d uri ng
> 2 months of searching for corpses in the
study area (31 field days, average interval
between searches = 2.3 days). Collision with
the wind turbines could be ruled out as the
cause of death in both cases for the following
reasons: 1) there were no fractures or
dislocations found during post mortem
examinations, 2) the birds were found > 150
m from the turbines, and 3) the birds were
found upwind of the wind farm and the wind
force was strong (4–5 Bft) on both days.
Dissection did not reveal a clear cause of
death and it was assumed that the swans had
died of natural causes or been killed by a
predator. That no swans were found to have
collided with the turbines during the study
period does not however, mean that the
collision rate was zero. In order to consider
the potential consequences of collision–
related mortality, a collision rate was
determined based on the assumption that
Table 2. Bewick’s Swan flights in the study area (ECN) and in the nearby Waterkaaptocht
wind farm (WK), recorded as visual and radar observations of the swans’ flight-paths. The
number of swans that were present in, or flying through the study area is shown; the
percentage of these birds that flew close to or through the wind farm during commuting
flights (% head towards wind farm), and thus potentially at risk of collision, was calculated
(i.e. number flying towards wind farm/number swans*100). Of the birds that flew toward
the wind farm, some avoided the wind farm entirely (% deflecting, i.e. number
deflecting/number swans*100) and some flew through the wind farm (% through wind farm,
i.e. number through wind farm/number swans*100).
Date Location No. swans % head to % deflecting % through
wind farm wind farm
21 Nov 2006 ECN 94 18 14 4
24 Nov 2006 ECN 294 100 98 2
01 Dec 2006 ECN 51 100 100 0
07 Dec 2006 ECN 459 66 1 65
16 Jan 2007 ECN 9 100 100 0
01 Dec 2006 WK 351 43 12 32
03 Jan 2007 WK 227 70 53 16
10 Jan 2007 WK 206 26 17 9
Mean ± s.d. 211 ± 155.1 65 ± 33.6 49 ± 44.0 16 ± 22.5
Effects of wind farm on Bewick’s Swans 109
© Wildfowl & Wetlands Trust Wildfowl (2012) 62: 97–116
one turbine victim was found in this study.
This does not provide an absolute measure of
collision rate, but does give a maximum
estimate of collision rate for the studied
season. This figure can subsequently be used
to estimate maximum collision risk (see next
section). The probability that a victim was
found (Pf) was set to 1 (see Methods section).
The disappearance tests found that seven
carcasses placed in the study area
disappeared at a slow rate (Fig. 5). After four
days, two were scavenged but all were still
present and recognisable. Only one bird, a
Mute Swan, totally disappeared during the
14-day trial; it was found to have been buried
by a Red Fox at the foot of a turbine, six days
after being laid out. The remaining six
carcasses were still present and recognisable
after fourteen days. A scavenging animal
moved two birds, by 1 m and 25 m
respectively. The probability (Pd) that a bird
was still present (after the average search
interval in this study: 2.3 days) was 0.97
(linear regression: Pd = –0.0255* number of
days since placement + 1.026, r2= 0.71, Fig.
5). The proportion of the total area
underneath the turbine that was searched
(Fs) was 0.986. The proportion of days over
the search period that victims were searched
for (Fd) was set to 1 as the mean interval
between searches was smaller than the
quickest disappearance of laid-out corpses.
Figure 4. Map of the study area, showing the wind farm area (eastern part) and the adjacent ‘control’
area (western part), with Bewick’s Swan flight-paths (arrows) from foraging fields to the Lake
IJjsselmeer roost on 24 November 2006. Numbers adjacent to the arrows indicate group size. Insert
shows the location of the study area within the Netherlands.
110 Effects of wind farm on Bewick’s Swans
© Wildfowl & Wetlands Trust Wildfowl (2012) 62: 97–116
Fitting the number of collision victims
found (between zero and one) to the above
parameters, gives an estimated collision rate
of 0–1.05 swans colliding with the turbines
each season for both wind farm sites. The
study season consisted of 1,163 ‘turbine
search days’ (number of turbines * number
of search days) so the estimated collision
rate is 1.05/1,163 = 0.0009 per turbine per
night. This collision rate implies a maximum
of approximately 2–3 victims per winter (15
October – 15 March) in both wind farms
considered in this study.
Collision risk
The near-daily swan counts gave an average of
132 Bewick’s Swans present each evening
during the 2006/2007 winter. Of these 132
birds, 16% flew through the wind farm area
(see the swan flights section above). Assuming
that the route to and from the roosting area is
flown twice per day, and that dusk flights are
as risky as dawn flights (noting that light levels
are low in both cases), an average of 42 swan-
flights pass the turbines every 24 h. With an
estimated maximum collision rate of 0.0009
birds per turbine per night, the maximum
collision risk can be calculated as: (17
(turbines) * 0.0009)/42 = 0.0004 (fraction), or
0.04% of all swans passing the two wind
farms. Because no actual collision victims
were found, this collision risk reflects the
maximum risk; the actual risk estimate is of
0–0.04 % of Bewick’s Swans passing these
particular turbines colliding with them in each
24 h period.
y = -0.026x + 1.026
r
2
= 0.714
y = -0.063x + 1.087
r
2
= 0.825
0%
20%
40%
60%
80%
100%
02468101214
Days since lay-out date
Presence Non-scavenged
Figure 5. Status of seven carcasses for up to 14 days after being placed in the study area. Shown are the
percentage of carcasses still present at the location after x days (black bars, closed line, linear regression
r2= 0.71 and a slope of –0.026) and the percentage of carcasses remaining at the location without being
scavenged by predators (grey bars, dotted line linear regression r2= 0.83 and a slope of –0.063).
Effects of wind farm on Bewick’s Swans 111
© Wildfowl & Wetlands Trust Wildfowl (2012) 62: 97–116
Discussion
Disturbance of foraging swans
The proportion of the total number of
Bewick’s Swans wintering in Polder
Wieringermeer that visited the study area
was significantly lower after construction of
the wind farm (2006/2007) than before it
was built (2000/2001). This decrease was
particularly evident in the wind farm area in
comparison with the adjacent area,
suggesting that the birds had been displaced
by the newly-constructed turbines. Whilst
the installation of the turbines seems to
have made the wind farm area less attractive
to the swans, the birds’ use of the ‘control’
area (without newly-built turbines) also
diminished, probably due to changes in food
availability between the two study seasons.
In particular, a smaller proportion of the
study area was used for sugar beet
cultivation in 2006/07 compared with
2000/2001 (100 ha versus 64 ha). On arrival
in the Netherlands, Bewick’s Swans start
feeding on water plants in other parts of the
country and only start feeding on crop
remains in Polder Wieringermeer later in the
season (Beekman et al. 1991; Dirksen et al.
1991). The timing of availability of harvest
waste is thus important for wintering
Bewick’s Swans in the Netherlands and an
absence or lower availability of crop remains
might cause shifts to other foraging areas.
Our study found that displacement of
Bewick’s Swans from the wind farm area
was most evident at the start of the season,
when there appeared to be an abundant
food supply for the birds. The swans were
more likely to forage in areas without
turbines while food was available in both the
‘control’ and wind farm areas. Only later in
the season, when food sources were limited
to just the wind farm area, swans
increasingly fed in areas closer to the
turbines. This decreasing distance between
foraging swans and the turbines may be due
to a lack of food further afield, to
habituation to the wind farm, or a
combination of these factors. Displacement
by wind turbines has also been reported for
Whooper Swan Cygnus cygnus and for several
species of geese, with the displacement of
birds evident up to 400 m of the turbines
(Winkelman 1989; Kruckenberg & Jaene
1999). Habituation to wind turbines has
also been found for the same species
(Kruckenberg & Jaene 1999; Larsen &
Madsen 2000; Madsen & Boertman 2008).
Devereux et al. (2008) showed that wintering
farmland birds (non–waterbirds) were not
influenced by wind turbines; however, our
results suggest that these results are not
applicable across all species wintering in
farmland areas.
Barrier effects
Although the swans appeared to be
displaced from potential feeding areas, there
was no evidence for the wind farm acting as
a barrier during the evening flight; the birds
navigated between and around the turbines
during their flights to the roost. This ability
to adjust their flight-paths is in line with
studies made of other waterbird species
(Dirksen et al. 1998; Tulp et al. 1999;
Desholm & Kahlert 2005; Masden et al.
2009). The small size of the wind farm in
this study (nine turbines in two rows) and
the large spacing between turbines may have
helped to ensure that these two lines did not
112 Effects of wind farm on Bewick’s Swans
© Wildfowl & Wetlands Trust Wildfowl (2012) 62: 97–116
act as a barrier to flying birds. The use of
modern large wind turbines may help both
to make the structures more obvious to the
birds (thus reducing collision risk) and also
perhaps reduce the chance that birds
perceive the turbines as barriers because the
larger spacing between individual turbines
makes it easier for the birds to pass between
them (Krijgsveld et al. 2009). The same
reasoning can be applied to increasing the
numbers of turbines within a wind farm, as
more turbines will enhance the perceived
barrier effect. The orientation of the turbine
rows will also have an effect, since turbines
constructed in rows parallel onto the
dominant flight direction of birds
commuting between foraging and sleeping
areas will present less of a barrier than when
perpendicular to it. In the extreme, such a
barrier effect could potentially render
roosting or foraging sites inaccessible,
especially where the energetic costs of
avoidance make significant additional
contributions to energy budgets. Due to
crop rotation, flight-paths could potentially
change between years. In this study, tracking
of flight-paths was limited to only one pre-
and one post-construction year; adequate
assessment of barrier effects requires
monitoring in multiple pre- and post-
construction years.
Collision risk
Avian turbine collision risk varies widely
between species and also between habitats;
for instance, raptors are often found to
collide with turbines in mountainous areas
(de Lucas et al. 2008; Smallwood &
Thelander 2008). Swans and geese are rarely
reported as turbine victims, although swan
collisions with power-lines have been
recorded frequently (e.g. Brown et al. 1992;
Rees 2006). This study found no collision
victims among Bewick’s Swans during the
research period, but the assumed one
collision victim per season would equate to
0–0.04% of swans passing the wind farm
turbines. These probabilities are very low
but are similar to results from extensive
research at two other turbine farms
involving geese and swan in other parts of
the Netherlands (Krijgsveld et al. 2009). The
collision risk at this wind farm is lower than
that calculated for Bewick’s Swans in the UK
(from Chamberlain et al. 2006) at a larger
study site (26 versus 9 turbines), located near
Romney Marsh, a proposed Ramsar site
with nationally important numbers of
Bewick’s Swans for the UK. However,
numbers of swans on Romney Marsh were
much lower (mean maximum = 123 swans
per winter during 2005–2009, Calbrade et al.
2010) than in the current study at the Polder
Wieringermeer.
This study covered no evenings and
mornings with fog or mist; on nights with
poor visibility, collision risk for swans could
be higher (Brown et al. 1992). However,
evenings or mornings with poor visibility
(< 300 m) were rare (five out of 114
dusks and dawns, Royal Netherlands
Meteorological Institute, KNMI-station
Berkhout, 21.11.2006–6.01.2007, downloaded
from www.knmi.nl), so effects of fog or
mist probably have negligible effects on the
collision risks found in this study.
Given our various assumptions, we
suggest a mortality rate of 0–3 swan victims
per winter for the whole wind farm, of
similar order of magnitude to 0.06 swans
Effects of wind farm on Bewick’s Swans 113
© Wildfowl & Wetlands Trust Wildfowl (2012) 62: 97–116
per 180 days found for the UK (from
Chamberlain et al. 2006). Collision risk can
be estimated, but where low, actual collision
rates can be difficult to determine. In other
studies, casualties are most frequent in bird-
rich areas and on mountain ridges (Hötker et
al. 2006; Thelander & Smallwood 2007; de
Lucas et al. 2008), but elsewhere, chances of
collision are much lower. To date, no clear
avian population effects from wind turbines
have been demonstrated, although these
effects will be greater for long-lived species
with low reproductive rates, such as seabirds
and raptors (Thelander et al. 2003; Horch &
Keller 2005; Hötker et al. 2006; Stienen et al.
2007). In the case of the Bewick’s Swans in
Polder Wieringermeer the collision risk
calculated in this study is so low that it is not
expected to cause negative effects on the
locally wintering swans. However, as Polder
Wieringermeer now supports large numbers
of individual wind farms, the combined
effects of all these wind turbines, together
with changes in cropping and land use,
could combine to reduce overall wintering
numbers of swans even in the absence of
collision mortality.
Implications for conservation and
future developments
In conclusion, this study shows that
although the collision risk for swans with
turbines was low at the site, wind farms can
result in a diminished use of foraging
habitat. Increasing demand for renewable
energy could result in more and larger
turbines which could reduce the
attractiveness and carrying capacity of
Polder Wieringermeer for wintering
Bewick’s Swans. Polder Wieringermeer is a
key wintering area for > 3% of the
Northwest European Bewick’s Swan
population, whilst the adjacent Lake
IJsselmeer roosts are of international
importance under the EC Birds Directive
and are designated as a Natura 2000
Specially Protected Area. This Birds’
Directive Annex I species has declined in
recent years (Wetlands International 2006;
Rees & Beekman 2010) so changes to the
potential carrying capacity of these
important areas should be considered with
caution. The increasing use of rural land in
Polder Wieringermeer for the construction
of wind turbines may have adverse impacts
on the quality of the habitat for wintering
waterbirds in the future.
Acknowledgements
This study was commissioned by ECN
Wind Energy Facilities and Wim Stam and
Henk Kouwenhoven of Dutch utility Nuon.
All monitoring was carried out by the Bird
Working Group Wierhaven, Bureau
Waardenburg and Alterra. We thank
everyone involved in the fieldwork for their
contribution. Jan van Gils kindly provided
helpful comments and additional data about
numbers of birds in earlier years. Thanks to
NIOO Nieuwersluis for the swans we used
in our carcass experiments. We express
special appreciation to all the farmers for
their cooperation and the possibility to do
research on their land. We would like to
express our gratitude to Mark Collier, Eileen
Rees, Tony Fox and two anonymous
referees for providing helpful comments to
improve the manuscript. All fieldwork
related to this study has been conducted in
accordance with Dutch law.
114 Effects of wind farm on Bewick’s Swans
© Wildfowl & Wetlands Trust Wildfowl (2012) 62: 97–116
References
Band, W., Madders, M. & Whitfield, D.P. 2007.
Developing field and analytical methods to
assess avian collision risks at wind farms. In M.
de Lucas, G.F.E. Janss & M. Ferrer (eds.), Birds
and wind farms. Risk Assessment and Mitigation,
pp. 259–275. Quercus, Madrid, Spain.
Beekman, J.H., Van Eerden, M.R. & Dirksen, S.
1991. Bewick’s Swans Cygnus columbianus
bewickii utilizing the changing resource of
Potamogeton pectinatus during autumn in the
Netherlands. In J. Sears & P.J. Bacon (eds.),
Proceedings Third IWRB International Swan
Symposium, Oxford 1989. Wildfowl Special
Supplement 1: 238–248.
Brown, M., Linton, E. & Rees, E.C. 1992. Causes
of mortality among wild swans in Britain.
Wildfowl 43: 70–79.
Calbrade, N.A., Holt, C.A., Austin, G.E., Mellan,
H.J., Hearn, R.D., Stroud, D.A., Wotton, S.R.
& Musgrove, A.J. 2010. Waterbirds in the UK
2008/09: The Wetland Bird Survey.
BTO/RSPB/JNCC in association with
WWT, Thetford, UK.
Chamberlain, D.E., Rehfisch, M.R., Fox, A.D.,
Desholm, M. & Anthony, S.J. 2006. The
effect of avoidance rates on bird mortality
predictions made by wind turbine collision
risk models. Ibis 148: 198–202.
De Lucas, M., Janss, G.F.E., Whitfield, D.P. &
Ferrer, M. 2008. Collision fatality of raptors
in wind farms does not depend on raptor
abundance. Journal of Applied Ecology 45:
1695–1703.
Desholm, M. & Kahlert, J. 2005. Avian collision
risk at an offshore wind farm. Biology Letters
1: 296–298.
Desholm, M., Fox, A.D., Beasley, P.D.L. &
Kahlert, J. 2006. Remote techniques for
counting and estimating the number of
bird–wind turbine collisions at sea: a review.
Ibis 148 (Suppl. 1): 76–89.
Devereux, C.L., Denny, M.J.H. & Whittingham,
M.J. 2008. Minimal effects of wind turbines
on the distribution of wintering farmland
birds. Journal of Applied Ecolog y 45: 1689–1694.
Dirksen, S., Beekman. J.H. & Slagboom, T.H.
1991. Bewick’s Swans Cygnus columbianus
bewickii in the Netherlands: numbers,
distribution and food choice during the
wintering season. In J. S e ar s & P. J. B ac on
(eds.), Proceedings Third IWRB International
Swan Symposium, Oxford 1989. Wildfowl
Special Supplement 1: 238–248.
Dirksen, S., Spaans, A.L., Van der Winden, J. &
Van den Bergh, L.M.J. 1998. Nocturnal flight
patterns and altitudes of diving ducks in the
IJsselmeer area. Limosa 71: 57–68. [In Dutch
with English summary.]
Dirksen, S., Spaans, A.L. & Van der Winden, J.
2007. Collision risks for diving ducks at semi-
offshore wind farms in freshwater lakes: a
case study. In M. de Lucas, G.F.E. Janss & M.
Ferrer (eds.), Birds and Wind Farms. Risk
Assessment and Mitigation, pp. 201–218.
Quercus, Madrid, Spain.
Drewitt, A.L. & Langston, R.H.W. 2008.
Collision effects of wind-power generators
and other obstacles on birds. Annals of the
New York Academy of Sciences 1134: 233–266.
Everaert, J. & Stienen, E.W.M. 2007. Impact of
wind turbines on birds in Zeebrugge
(Belgium). Significant effect on breeding tern
colony due to collisions. Biodiversity and
Conservation 16: 3345–3359.
Grunkorn, T., Diederichs, A., Pöszig, D.,
Diederichs, B. & Nehls, G. 2009. How many
birds collide with wind turbines? Natur und
Landschaft 84: 309–314. [In German with
English summary.]
Horch, P. & Keller, V. 2005. W indkraftanlagen
und Vogel – ein Konflikt? Schweizerische
Vogelwarte Sempach, Sempach, Switzerland.
Available on http://www.vogelwarte.ch/. [In
German.]
Effects of wind farm on Bewick’s Swans 115
© Wildfowl & Wetlands Trust Wildfowl (2012) 62: 97–116
Hötker, H., Thomsen, K-M. & Köster, H. 2006.
Impacts on biodiversity of exploitation of renewable
energy sources: the example of birds and bats. Facts,
gaps in knowledge, demands for further research, and
ornithological guidelines for the development of
renewable energy exploitation. Michael–Otto–
Institut im NABU, Bergenhusen, Germany.
http://bergenhusen.nabu.de/bericht/
englische%20windkraftstudie.pdf.
Hunt, W.G., Jackman, R.E., Hunt, T.L., Driscoll,
D.E. & Culp, L. 1998. A population study of
golden eagles in the Altamont Pass Wind Resource
Area: population trend analysis 1994–1997.
NREL/SR–500–26092, Subcontract No.
XAT–6–16459–01. Predatory Bird Research
Group University of California, Santa Cruz,
California, USA.
Krijgsveld, K.L., Akershoek, K., Schenk, F., Dijk,
F. & Dirksen, S. 2009. Collision of birds with
modern large wind turbines. Ardea 97:
357–366.
Kruckenberg, H. & Jaene, J. 1999. Zum
Einfluss eines Windparks auf die Verteilung
weidender Blässgänse im Rheinland
(Landkreis Leer, Niedersachsen). Natur und
Landschaft 74: 420–424. [In German with
English summary.]
Langston, R.H.W. & Pullan, J.D. 2003. Windfarms
and birds: An analysis of the effects of windfarms on
birds, and guidance on environmental assessment
criteria and site selection issues. Report RSPB/
BirdLife International for the Convention on
the Conservation of European Wildlife and
Natural Habitats, Standing Committee 23rd
Meeting, Strasbourg, 1–4 December 2003, 58
pp. Council of Europe, Strasbourg, France.
Larsen, J.K. & Madsen, J. 2000. Effects of wind
turbines and other physical elements on field
utilization by Pink–footed Geese (Anser
brachyrhynchus): a landscape perspective.
Landscape Ecology 15: 755–764.
Larsen, J.K. & Clausen, P. 2002. Potential
wind park impacts on whooper swans in
winter: the risk of collision. Waterbirds 25:
327–330.
Madders, M. & Whitfield, D.P. 2006. Upland
raptors and the assessment of wind farm
impacts. Ibis 148 (Suppl. 1): 43–56.
Madsen, J. & Boertmann, D. 2008. Animal
behavioral adaptation to changing landscapes:
spring-staging geese habituate to wind farms.
Landscape Ecology 23: 1007–1011.
Martin, G.R. 2011. Understanding bird collisions
with man-made objects: a sensory ecology
approach. Ibis 153: 239–254.
Masden, E.A., Haydon, D.T., Fox, A.D., Furness,
R.W., Bullman, R. & Desholm, M. 2009.
Barriers to movement: impacts of wind
farms on migrating birds. ICES Journal of
Marine Science 66: 746–753.
Orloff, S. & Flannery, A. 1992. Wind turbine
effects on avian activity, habitat use and mortality
at Altamont Pass and Solano County Wind
Resource Areas, 1989–1991. California Energy
Commission, Sacramento, California (USA).
www.energy.ca.gov/reports.
Painter, A., Little, B. & Lawrence, S. 1999.
Continuation of Bird Studies at Blyth Harbour
Wind Farm and the Implications for Offshore Wind
Farms. Report by Border Wind Limited DTI,
ETSU W/13/00485/00/00, Hexham, UK.
Percival, S.M. 2007. Predicting the effects of
wind farms on birds in the UK: the
development of an objective assessment
method. In M. de Lucas, G.F.E. Janss & M.
Ferrer (eds.), Birds and wind farms. Risk
Assessment and Mitigation, pp. 137–152.
Quercus, Madrid, Spain.
Rees, E.C. 2006. Bewick’s Swan. T. & A.D. Poyser,
London, UK.
Rees, E.C. & Beekman, J.H. 2010. Northwest
European Bewick’s Swans: a population in
decline. British Birds 103: 640–650.
Schreiber, M. 1993. Windkraftanlagen und
Watvogel-Rastplatze, Storungen und
Rastplatzwahl von Brachvogel und
116 Effects of wind farm on Bewick’s Swans
© Wildfowl & Wetlands Trust Wildfowl (2012) 62: 97–116
Goldregenpfeifer. Natur und Landschaft 25:
133–139. [In German.]
Smallwood, K.S. & Thelander, C. 2008. Bird
Mortality in the Altamont Pass Wind
Resource Area, California. Journal of Wildlife
Management 72(1): 215–223.
Stienen, E.W.M., Van Waeyenberge, J., Kuijken,
E. & Seys, J. 2007. Trapped within the
corridor of the Southern North Sea: the
potential impact of offshore wind farms on
seabirds. In M. de Lucas, G.F.E. Janss & M.
Ferrer (eds.), Birds and wind farms. Risk
Assessment and Mitigation, pp. 71–80. Quercus,
Madrid, Spain.
Thelander, C.G., Smallwood, K.S. & Rugge, L.
2003. Bird risk behaviors and fatalities at the
Altamont Pass Wind Resource Area. National
Renewable Energy Laboratory, Golden,
Colorado, USA. http://www.nrel.gov/wind/
pdfs/33829.pdf.
Thelander, C.G. & Smallwood, K.S. 2007. The
Altamont Pass Wind Resource Area’s effects
on birds: a case history. In M. de Lucas,
G.F.E. Janss & M. Ferrer (eds.), Birds and
Wind Farms. Risk Assessment and Mitigation, pp
25–46. Quercus, Madrid, Spain.
Tulp, I., Schekkerman, H., Larsen, J.K., Van der
Winden, J., Van de Haterd, R., Van Horssen,
P., Dirksen, S. & Spaans, A.L. 1999. Nocturnal
flight activity of sea ducks near the wind farm Tunø
Knob in the Kattegat. Bureau Waardenburg
report 99.64, Culemborg, the Netherlands.
van Gils, J.A. & Tijsen, W. 2007. Short–term
foraging costs and long–term fueling rates
in central-place foraging swans revealed
by giving-up exploitation times. American
Naturalist 169: 609–620.
Wetlands International. 2006. Waterbird Population
Estimates – Fourth Edition. Wetlands
International, Wageningen, the Netherlands.
Winkelman, J.E. 1989. Vog el s e n he t w in dpa rk na bij
Urk (NOP): aanvaringsslachtoffers en verstoring van
pleisterende eenden ganzen en zwanen. RIN–report
89/15. RIN, Arnhem, the Netherlands. [In
Dutch with English summary.]
Winkelman, J.E. 1992. De invloed van de
Sep–proefwindcentrale te Oosterbierum (Fr.) op
vogels. 1. Aanvaringsslachtoffers. Rijksinstituut
voor Natuurbeheer (RIN) report 92/2.
Instituut voor Bos- en Natuuronderzoek
(IBN-DLO), Arnhem, the Netherlands. [In
Dutch with English summary.]
Photograph: Bewick’s Swans on pasture in the Netherlands by Jelger Herder (www.digitalnature.org).
