Estimating mortality of small passerine birds colliding with wind turbines

Abstract

As demand for renewable energy is rising, wind power development is rapidly growing worldwide. In its wake, conflicts arise over land use changes converting pristine nature into industrial power plants and its associated adverse biodiversity effects, crowned by one of the most obvious and deadly consequences: bird collisions. Most post-construction studies report low levels of avian mortality, but the majority of these studies are conducted primarily on larger birds. However, the diversity and abundance of small passerine birds are rarely reflected in the carcass surveys, although they in numeric proportion to their abundances should be the most numerous. The assumption that surveys find all carcasses seems thus rarely fulfilled and passerine mortality is likely to be grossly underestimated. We therefore designed an experiment with dummy birds to estimate mortality of small-bodied passerines and other small-bodied birds during post-construction surveys, tested in a medium-sized wind farm in western Norway. The wind farm was surveyed weekly during the migration periods by carcass survey teams using trained dogs to find killed birds. The dogs in the carcass surveys were more successful in locating the large than the small dummy birds (60–200 g), where they found 74% of the large dummy birds. Detecting the smaller category (5–24 g) was more demanding and the dogs only found 17% of the small dummy birds. Correcting the post-construction carcass survey outcome with the results from the experiment leads to an almost fourfold increase in estimated mortality rates, largely due to the low detection rate of the smallest category. The detection rates will naturally vary between wind farms, depending on the specific habitat characteristics, the efficiency of the carcass surveys and the search intervals. Thus, implementing a simple experiment with dummy birds to future post-construction surveys will produce more accurate estimates of the wind turbine mortality rates, and thus improve our understanding of the biodiversity effects of conforming to a more sustainable future.

Full text

1
Vol.:(0123456789)
Scientic Reports | (2023) 13:21365 | https://doi.org/10.1038/s41598-023-46909-z
www.nature.com/scientificreports
Estimating mortality of small
passerine birds colliding with wind
turbines
A. L. K. Nilsson
*, S. Molværsmyr , A. Breistøl & G. H. R. Systad
As demand for renewable energy is rising, wind power development is rapidly growing worldwide.
In its wake, conicts arise over land use changes converting pristine nature into industrial power
plants and its associated adverse biodiversity eects, crowned by one of the most obvious and deadly
consequences: bird collisions. Most post-construction studies report low levels of avian mortality,
but the majority of these studies are conducted primarily on larger birds. However, the diversity
and abundance of small passerine birds are rarely reected in the carcass surveys, although they
in numeric proportion to their abundances should be the most numerous. The assumption that
surveys nd all carcasses seems thus rarely fullled and passerine mortality is likely to be grossly
underestimated. We therefore designed an experiment with dummy birds to estimate mortality of
small-bodied passerines and other small-bodied birds during post-construction surveys, tested in a
medium-sized wind farm in western Norway. The wind farm was surveyed weekly during the migration
periods by carcass survey teams using trained dogs to nd killed birds. The dogs in the carcass surveys
were more successful in locating the large than the small dummy birds (60–200 g), where they found
74% of the large dummy birds. Detecting the smaller category (5–24 g) was more demanding and
the dogs only found 17% of the small dummy birds. Correcting the post-construction carcass survey
outcome with the results from the experiment leads to an almost fourfold increase in estimated
mortality rates, largely due to the low detection rate of the smallest category. The detection rates will
naturally vary between wind farms, depending on the specic habitat characteristics, the eciency
of the carcass surveys and the search intervals. Thus, implementing a simple experiment with dummy
birds to future post-construction surveys will produce more accurate estimates of the wind turbine
mortality rates, and thus improve our understanding of the biodiversity eects of conforming to a
more sustainable future.
As the consequences of climate change are becoming more apparent, the global demand for renewable energy is
rapidly rising1. Among the currently available renewable alternatives, wind power is increasingly seen as the most
feasible solution at hand. Compared to hydropower with its extensive damming preventing the natural river ow
severely impacting movements of aquatic organisms2,3, wind power has seemingly a lower overall environmental
impact in the public eye. However, the development of onshore wind farms represents yet another claim on ter-
restrial land use and there is growing concern about the impact on terrestrial biodiversity, particularly airborne
animals such as birds, bats and insects4. While wind farms can induce displacement due to disturbance or act as
barriers to migrating birds, which might ultimately result in habitat loss and increased costs of migration, the
most obvious consequence of wind power development is bird collisions5.
Bird collisions represent a source of direct mortality to birds; birds can collide with rotor blades, towers, the
nacelle or other associated structures. Collisions can be a signicant cause of mortality with potential population
consequences6. Raptor populations are particularly susceptible, but also seabird populations have been aected
by collision mortality7–9. Strategic planning and careful turbine placements might be one of the most important
mitigation measures to reduce collision mortality. However, most studies report relatively low levels of collision
mortality5,6,10–14, but the primary focus of studies of collision mortality, especially in Norway, has been on large
birds, perhaps because they oen are large-bodied and easier to nd in carcass surveys10,15,16. e collision rates
by raptors are certainly alarming in light of their comparatively low reproductive rate and small population sizes17.
Passerine birds, on the other hand, comprise the clear majority of the birds on Earth, both in species numbers
and abundances, and as such they are likely also colliding more oen with wind turbines than other groups of
OPEN
Norwegian Institute for Nature Research, Thormøhlensgate 55, 5006 Bergen, Norway. *email: anna.nilsson@nina.
no
Content courtesy of Springer Nature, terms of use apply. Rights reserved

2
Vol:.(1234567890)
Scientic Reports | (2023) 13:21365 | https://doi.org/10.1038/s41598-023-46909-z
www.nature.com/scientificreports/
birds18,19. Small passerines as well as other small-bodied birds are however oen seriously underrepresented
in the collision statistics, as pointed out by Rydell etal.4 and Grünkorn etal.20. However, Ericksen etal.21 shed
light on wind energy mortality in small passerine birds. Since passerines in general are smaller, their carcasses
are more dicult to nd underneath the turbines, and they also decompose faster than larger birds5,22. us,
there remain considerable uncertainty in whether all carcasses are in fact found during the carcass surveys in
post-construction studies5.
Carcass surveys are surprisingly oen made by human observers12,18,23 instead of using trained dogs which
are proven to be distinctly more ecient in locating carcasses14,24–26. Trained dogs eciently search an area faster
than humans relying only on visual cues, thus enabling searches to be both time- and cost-eective. erefore,
surveys can therefore be conducted at shorter intervals. Irrespective of survey method, most studies assume all
carcasses are found, which might not be true, especially not for surveys conducted by humans or where scaven-
gers are ecient in removing carcasses. A few studies estimate the rate of carcass removal by scavengers14,27,28,
but few studies estimate the search eciency of the carcass surveys. Estimating search eciency is imperative,
because it allows estimation of the true wind turbine mortality and therefore a more accurate assessment of the
biodiversity eects of wind power development23. Such an approach is thus not about controlling the search
parties’ eciency, irrespective of whether they are composed by humans or dogs. Estimating search eciency
in order to assess turbine mortality is particularly important at wind power developments situated at migratory
pathways with millions of passing small-bodied migrants.
Most studies nd few6,14,29 or no dead small-bodied birds but see21,23. Because we had access to a post-con-
struction survey scheme with intense searches by skilled dog equipages able to nd the bodies of the small-bodied
migrants colliding with the turbines in the wind farm, we were able to design a simple experiment to estimate
small passerine search eciency, which easily can be implemented in any post-construction study. We conducted
the experiment at an average-sized wind power development in a coastal mountainous region in western Norway
where there is intense seasonal migration of primarily passerine birds30.
Results
Dummy birds are dened as dead, small passerine, and in a few cases small non-passerine birds that were own
out to under the turbines by a drone and recovered by the carcass survey teams in 2021. Small dummy birds
comprised 5–24 g birds, whereas large dummy birds comprised 60–200 g birds. Among the 47 own-out dummy
birds, only a total of 21 (45%), were found by the carcass surveys. However, the search parties found as many
as 74% of the larger dummy birds (17 of 23), but only 17% (four of 24) of the smallest dummy birds (Χ2 = 4.9,
df = 1, P = 0.027).
In the post-construction survey the dummy experiment was conducted under, 54 and 51 turbine-killed
small-bodied birds were found under the turbines in 2021 and 2022, respectively (Table1). In 2021, the killed
birds of the same size as the large dummy birds were 22, and 29 birds of the same size as the small dummy birds,
in addition to three larger birds. In 2022, 11 were of the same size as the large dummy birds, 29 of the same size
as the small dummy birds, and 11 larger birds. e majority of carcasses, 87% (91 of 105), were small passerine
birds. Five of the carcasses were partly decomposed or only parts of them were found. Four carcasses of Wil-
low Ptarmigan Lagopus lagopus were found intact, but at such a distance from the turbine base as to indicate a
collision with the tower. At some turbines no killed birds were detected, whereas other turbines killed as many
as four birds per migratory period. e maximum number of killed birds by one single turbine was six, two in
spring and four in autumn.
By applying the results from the dummy study to the post-construction study results in 2021 and 2022,
we estimate the accurate turbine mortality for the large dummy-sized birds where there was found 22 and 11,
respectively, and where the detection rate was 0.74, as 33/0.74 = 44.6. e post-construction study found 29 small
dummy-sized birds in each of the study years, where the detection rate was 0.17, and the real turbine mortality
for this category was thus estimated to 58/0.17 = 341.2. Including the 14 larger birds found in the carcass surveys,
this amounts to 400 dead birds in total during the two years of annual migration, including two spring migrations
and two autumn migrations. us, the turbine mortality of small-bodied birds (large and small dummy-sized
birds) from the carcass surveys alone were grossly underestimated, since based on the dummy results, the real
turbine mortality was almost four times as high. e real turbine mortality equals 0.77 dead birds per turbine
per migration month, or 0.62 dead birds per produced GWh.
Among the dummy birds found by the carcass surveys, most were found within the rst week aer placement
(14 of 21; 67%). ere was no statistically signicant dierence in the number of days until dummy birds were
found between large and small dummy birds (df = 3.2, t = 1.9, P = 0.154), although the average for large and small
dummy birds were 5.2 days (SD = 4.7) and 15.7 days (SD = 11.2), respectively.
Among the 26 dummy birds not found by the carcass surveys, seven positions could not be relocated in a
manual search at the end of the autumn season. is might be due to poor drone pictures, snow cover having
altered the appearance of the position, scavenger removal, or few landmarks around the dummy birds on the
drone picture. Among the 19 dummy birds that were relocated on this manual search, 11 were found on the same
position or in the immediate surroundings (one was swept o 1–2 m by a small brook). 10 of the 11 relocated
dummy birds in the same position or in the immediate surroundings were small dummy birds. e only remain-
ing large dummy bird, own out in October was not found until the next summer aer the dummy experiment
was concluded, and not included in the present calculations.
We also estimated the detection rate of dummy birds over time (large dummy birds: df = 14, R2 = 0.53, a = 0.72,
b = − 0.03, t = − 4.0, P = 0.001; small dummy birds: df = 47, R2 = 0.03, a = 0.06, b = − 0.001, t = − 1.2, P = 0.242;
Fig.1) and simulated how rapidly the proportion of dead birds found falls with longer intervals between carcass
surveys (Fig.2). For the small dummy-sized birds the proportion falls very rapidly, and high search frequencies
Content courtesy of Springer Nature, terms of use apply. Rights reserved

3
Vol.:(0123456789)
Scientic Reports | (2023) 13:21365 | https://doi.org/10.1038/s41598-023-46909-z
www.nature.com/scientificreports/
are required to locate the carcasses before they decompose or are removed by scavengers (Fig.2). For the large
dummy-sized birds the proportion does not fall as rapidly with longer search intervals, although it has declined
to 25% detected at 1 month search intervals. For small dummy-sized birds, the detected proportion was even
lower (Fig.2). us, to locate passerine migrants, it is necessary with short intervals between the carcass surveys,
such that each turbine is searched with high frequency.
Discussion
Here, we demonstrate how a simple addition to a post-construction survey design can increase the accuracy of
estimates of biodiversity impacts of wind power developments. ere was a clear dierence in how many of the
large versus the small dummy-sized birds were found by the carcass surveys; search eciency was considerably
higher for the large dummy birds than for the small ones. Based on search eciency, we could calculate more
accurate estimates of the biodiversity impacts, here measured as turbine mortality. Despite very skilled carcass
survey teams, real turbine mortality for small-bodied birds amounted to almost four times as high as what it
would have been if the biodiversity impact had solely been based on what was found by the carcass surveys. e
lower detection of the small dummy birds contributed the most to the considerable increase in the estimates;
even the large dummy birds which consisted of small birds the size of CommonStarlings Sturnus vulgaris and
thrushes Turdus sp. were easily detected by the carcass surveys in the rough terrain. us, all birds larger than
the dummy birds were most likely found. However, the detection probabilities found here are site-specic,
thus they only apply to Guleslettene Wind Farm, and also the rst years aer the wind farm was built. e high
Table 1. e number and species of the turbine-killed birds found under the turbines at Guleslettene Wind
Farm during spring migration (March 15th–May 20th) and autumn migration 2021 (July 15th–October
31st), respectively. Red List status is marked in bold, based on the Norwegian Red List of threatened species42.
*Indicates killed birds of the size of the small dummy birds. **Indicates killed birds of the size of the large
dummy birds.
Species
2021 2022
Spring Autumn Spring Autumn
Eurasian Teal (Anas crecca) 1
Willow Ptarmigan (Lagopus lagopus) 1 3 1
Grouse sp. (Lagopus sp.) 1 1
Black Grouse (Lyrurus tetrix) 1
Great Cormorant (Phalacrocorax carbo) 1
White-tailed Eagle (Haliaeetus albicilla) 1
European Golden Plover (Pluvialis apricaria) NT** 2 2
Jack Snipe (Lymnocryptes minimus)** 3
Common Snipe (Gallinago gallinago)** 1 1 1
Great Black-backed Gull (Larus marinus) 2
Common Wood Pigeon (Columba palumbus) 1
Meadow Pipit (Anthus pratensis)* 6 2
White Wagtail (Motacilla alba)* 2
Eurasian Wren (Troglodytes troglodytes)* 2
European Robin (Erithacus rubecula)* 2 1 1
Redwing (Turdus iliacus)** 12 3
Song rush (Turdus philomelos)** 1
Fieldfare (Turdus pilaris)** 1
Common Blackbird (Turdus merula)** 2 1
Ring Ouzel (Turdus torquatus)** 1
Willow Warbler (Phylloscopus trochilus)* 2 4 2
Common Chicha (Phylloscopus collybita)* 1
Goldcrest (Regulus regulus)* 9 1 10
Spotted Flycatcher (Muscicapa striata)* 1
EuropeanPied Flycatcher (Ficedula hypoleuca)* 1
Eurasian Treecreeper (Certhia familiaris)* 1
Common Starling (Sturnus vulgaris) NT** 1
Eurasian Chanch (Fringilla coelebs)* 1 1
Brambling (Fringilla montifringilla)* 2 1
Red Crossbill (Loxia curvirostra)* 1
Passerine sp. (Passeriformes sp.) 1 1 3
Sum 11 43 20
Content courtesy of Springer Nature, terms of use apply. Rights reserved

4
Vol:.(1234567890)
Scientic Reports | (2023) 13:21365 | https://doi.org/10.1038/s41598-023-46909-z
www.nature.com/scientificreports/
frequency of the carcass surveys contributed to the surprisingly high detection rate of all small-bodied birds in
this post-construction study.
Increasing the accuracy of biodiversity impact estimates by conducting a simple experiment during the
post-construction study is particularly relevant at wind power developments in habitats that are challenging to
conduct carcass surveys in. Few Scandinavian wind power developments are found in habitats so highly suitable
for carcass surveys such as the habitat for instance Aschwanden etal.23 conducted their detailed study in, with
closely mowed grass lands ideal for carcass survey experiments. Many Norwegian wind power developments and
indeed many wind power sites around the world, are situated at exposed mountain plateaus with bare mountain
and rocks, with clis and crevices, and sparse vegetation, which are extremely challenging to conduct carcass
0.00
0.25
0.50
0.75
1.00
01020304
05
0
Days after placement
Probability of finding a dummy
Figure1. Observed probability of locating a dummy bird on the carcass surveys N days aer they were placed
under the turbines at Guleslettene Wind Farm. e datapoints are aggregated for the whole season. Black = large
dummy birds, grey = small dummy birds.
Figure2. e proportion of all dead birds found, based on data from Guleslettene Wind Farm and a simulation
with dierent search intervals. e proportion of birds found drops as time progress, as the likelihood of nding
each individual approach zero. Black = large dummy birds, grey = small dummy birds.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

5
Vol.:(0123456789)
Scientic Reports | (2023) 13:21365 | https://doi.org/10.1038/s41598-023-46909-z
www.nature.com/scientificreports/
surveys in, even in good weather. Such terrain likely also inuences the detection rates because carcasses might
fall down into cracks and crevices which might be almost impossible to access, thus emphasizing the diculty
of not only performing carcass surveys but also the likelihood of nding and locating carcasses. is further
stress the need for trained dogs in the carcass surveys teams.
In challenging survey conditions, it is inherently dicult to conduct carcass surveys especially aer small-
bodied birds. Most post-construction surveys nd few, if any, small-bodied birds14,20,22,28. is makes post-
construction studies of wind power developments at migratory passage sites particularly awkward, because the
bulk of migrants are small-bodied passerine birds31 (but see32). At the post-construction survey at Guleslettene
Wind Farm, despite the particularly dicult terrain and exposed survey conditions, the majority of carcasses
were in fact small passerine birds, and as was evident from this, the search eciency was therefore surprisingly
high. Although most carcasses were of migratory birds, some of the carcasses were likely locally breeding birds
or their edged ospring. e simplest remedy for amending the exclusion of passerine birds is, as mentioned
before, the use of trained dogs in carcass surveys, but, as is also evident from this study, it is also important with
frequent surveys. Carcasses of small-bodied birds have a high detection probability the rst days aer they are
killed but this declines rapidly over time. is is likely due to a combination of decomposition of the carcasses
and scavengers removing carcasses. As small carcasses decompose rapidly, the scent traces that could be picked
up by the dogs in the carcass surveys, especially in wet weather conditions, are reduced. us, anything less
than carcass surveys every second week cannot be expected to truly measure turbine mortality of small-bodied
birds. e rates of decomposition and scavenger removal are highly relevant for estimating turbine mortality,
but estimates of rates are necessarily site-specic, meaning that the rapid decline in detection probability found
here is not directly transferable to other wind parks.
ere was a distinct dierence in detection probability between large and small dummy birds. e dogs had
seemingly little diculty in locating most of the large dummy birds, while the small dummy birds turned out
to be more challenging to detect. However, enough of the small dummy birds were located to enable a separate
estimation of search eciency and thus a more accurate estimation of turbine mortality rates during migration.
e overall search eciency of this study (45%, including all dummy birds) was actually larger than the overall
search eciency of other studies cf 38.7%32. Aschwanden etal.23 found that small birds equal in size to our small
dummy birds had a detection probability of less than 50% when the vegetation height was more than 10 cm,
whereas medium-sized birds equaling our large dummy birds just as in our study had a higher detection prob-
ability. An American study33 estimated search eciency of the smallest birds in their study, which was comparable
to our large dummy birds, to 42–55%. Neither of the comparable studies used trained dogs.
Comparisons of mortality rates between wind power developments are not straightforward, because a multi-
tude of factors can aect the estimates. Some environments have higher bird diversity and abundances, such as
wetlands and coastal regions, and are therefore also associated with intrinsically higher mortality rates34. Some
habitats are also easier to survey and conduct experiments in than others, where grasslands are considerably
less challenging than high mountainous, rocky terrain with outcrops and crevices23,26,35,36. Also, post-construc-
tion surveys are currently not standardized, despite the obvious need for repeatable and adequate monitor-
ing methods5; there is therefore a diversity of approaches used that render comparisons of turbine mortality
extremely complex. e length of the post-construction survey and its design might strongly dier, depending
on conditions set by authorities or the developer’s own ethical standards, as well as the frequency and extent of
carcass surveys, whether scavenger activity has been controlled for or not, whether carcass surveys are conducted
by humans or trained dogs, etc.14,18,23,34. For instance, the post-construction survey at Guleslettene Wind Farm
only included the migration periods, whereas other surveys range from comprising the full annual cycle12,15,33,37
to much shorter periods comprising only a few weeks38,39. Also, the fact that most carcass surveys fail to nd the
small passerines and other small-bodied birds that were the main target of the present study, and, moreover, fail
to correct for search eciency in impact measures, implies that direct comparisons are barely possible. While
Rydell etal.34 assumed that most wind turbines kill between ve and 10 birds annually, they also stated that most
turbines kill very few birds, whereas a few, oen ill-placed turbines kill considerably more. To mitigate biodi-
versity impacts, it is particularly important to detect those ill-placed turbines. Despite the listed shortcomings,
an attempt to compare our more accurate estimate of turbine mortality with these rough estimates likely places
Guleslettene Wind Farm below what would have been expected, despite its location at a migratory passage site
and the presence of ridges providing thermals for raptors and other large soaring birds.
Implementing the study design presented here in more onshore post-construction studies will increase the
present status of knowledge and moreover enable comparisons between dierent projects. is will ultimately
lead to even better siting of onshore wind power projects in the future as society continue to work towards a
more sustainable future.
Methods
Guleslettene is a high mountain plateau close to the coast in the north of Vestland county in Norway (Fig.3). It is
characterized by bare mountain and rocks, and low vegetation of lichens, mosses and grasses. Heather (primarily
Europeanblueberry Vaccinium myrtillus and crowberry Empetrum nigrum) occurs in lower or more sheltered
parts of the mountain plateau. e wind farm at Guleslettene is spread across almost the entire plateau and con-
sists of 47 wind turbines with an annual estimated total production of 700GWh. e turbine bases rise 90m, and
the rotor blade radius is 68m. e post-construction study at Guleslettene was conducted during 2021 and 202240.
We carried out a dummy experiment to estimate the mortality of small-bodied passerines and other small-
bodied birds during the rst year of the post-construction survey. Carcass surveys for the dummy experiment
was carried out at the same time as the carcass surveys for the rst year of the post-construction study (2021).
e post-construction study was a requirement for the wind development concession at Guleslettene. e carcass
Content courtesy of Springer Nature, terms of use apply. Rights reserved

6
Vol:.(1234567890)
Scientic Reports | (2023) 13:21365 | https://doi.org/10.1038/s41598-023-46909-z
www.nature.com/scientificreports/
surveys were continued during the second year of the post-construction study in 2022. e carcass surveys were
conducted weekly under all the turbines with the aid of trained dogs during spring (March 15th–May 15th) and
autumn migration (July 15th–31st October). e carcass survey teams used one dog in each search around a
turbine, although some carcass survey teams had access to more than one dog. A 100-m radius around each
turbine was surveyed41, where half of the wind farm was covered during day 1 and the rest during day 2. In some
instances, surveys were also conducted on day 3. Because of the location of the wind farm at a high, exposed
mountain plateau, days with blizzards, heavy rain or storms occasionally prevented surveys to be completed as
planned or had to be aborted aer initiation. Each turbine was surveyed on average every 9 days (8.8 ± 3.4 days).
ere were 906 individual carcass surveys in 2021, measured as individual turbine searches, conducted for 65
days during 25 weeks. Each turbine survey lasted on average 20.3 min (SD = 8.2), but several of the turbines had
so steep surroundings as to render it unsafe to search in the entire 100-m radius.
To avoid that the dogs would follow our tracks in the snow in early spring and any odor traces in the snow
as well as on the bare ground as the season progressed, we always ew out all dummy birds using a drone. To
avoid confusion with real collision fatalities, as the dummy experiment was carried out simultaneously with the
post-construction carcass surveys, a thin white rope was tied to the tibia of all dummy birds. e loop of the
thin rope was also used as an attachment point to the drone. All dummy birds were placed within the 100 m
radius from the turbine base the dogs were searching. Most dummy birds were birds previously found by the
dogs at the wind farm. In the cases where there were not enough birds available some birds were acquired from
Flesland airport in Bergen or birds that had hit the windows of one of the researchers during the spring (two
EurasianBullnches Pyrrhula pyrrhula). All dummy birds were kept in a freezer before and between placement
under the turbines. Some of the dummy birds were in good condition aer use (meaning they were found shortly
aer placement) and were reused in the experiment.
In total 47 dummy birds were placed around the turbines, whereof a total of 14 were own out successively
during spring and a total of 33 likewise during the autumn (Supplementary Data 1). All dummy birds were
frozen immediately aer collection, both before experiments and aer retrieval by the carcass survey teams. To
avoid attracting raptors and other avian scavengers such as Hooded CrowsCorvus cornix and NorthernRavens
Corvus coraxwhich might increase their collision rates, we only used small-bodied dummy birds. e dummy
bird species were Eurasian Chanch Fringilla coelebs, Eurasian Bullnch Pyrrhula pyrrhula, Fieldfare Turdus
pilaris, EuropeanGolden Plover Pluvialis apricaria, Meadow Pipit Anthus pratensis, Willow Warbler Phylloscupus
trochilus, Barn Swallow Hirundo rustica, Song thrush Turdus philomelos, European robin Erithacus rubecula,
Redwing Turdus iliacus, CommonStarling Sturnus vulgaris and CommonBlackbird Turdus merula. Because
of the obvious size dierences, we also divided the dummy bird species into small (EurasianChanch, Eura-
sianBullnch, Meadow Pipit, Willow Warbler, Barn Swallow, and European Robin; 5–24 g) and large (Fieldfare,
European Ggolden Plover, Song rush, Redwing,Common Starling, and Common Blackbird; 60–200 g) dummy
birds for the analyses. At the close of the autumn season, we set out to locate the 26 dummy birds not found by
the carcass surveys using known GPS positions and drone pictures of the original locations.
Figure3. Map over the Wind farm at Guleslettene, in western Norway.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

7
Vol.:(0123456789)
Scientic Reports | (2023) 13:21365 | https://doi.org/10.1038/s41598-023-46909-z
www.nature.com/scientificreports/
Statistical analyses
e proportion of detected dummy birds was dened as the number of dummy birds found by the carcass survey
teams divided by the total number of own-out dummy birds of each size category. To test whether there was a
dierence in detection rate between large and small dummy birds, we used a Chi-square-test. We tested whether
there was a dierence in the number of days required to retrieve a large versus a small dummy bird with a T-test.
To nd the proportion of detected birds, the time of placement and time of retrieval for each dummy bird
were combined with the search occurrences for each individual turbine. For each search, we then calculated the
number of days each dummy bird had been present. For each day aer placement from 0 to 49 we then calcu-
lated, to avoid too many missing values, the rolling average (± 1 day) proportion of dummy birds found. is
also meant that a dummy bird that was never found would count in multiple searches. To these data there were
tted linear regression models, one for large dummy birds, and one for small dummy birds.
Based on these linear regression models a simulation was performed to visualize what a real-life scenario
would look like, and which proportion of the killed birds would be found with dierent search intervals in the
carcass surveys. As actual bird deaths are unlikely to be a uniform distribution throughout the year, the turbine
killed birds found at Guleslettene was used to create an estimate of how many birds was killed in the windfarm
for each day. is estimate was created by taking the average of number of birds found by the real carcass surveys
from 10 days before to 5 days aer each date. If there were no birds found in that period (either because of no
searches or no ndings) the kill rate was set to 1 bird per day, both to ensure a minimum threshold for daily
mortality estimates and to make the simulation less likely to miss bursts of dead birds at long search intervals.
e estimated death rate per day varied between 1 and 5, and could be a decimal number. ereaer, two basic
population models of killed birds were created, one for large and one for small birds. For each population
model, for each day the number of killed birds from the estimated death rates was added to the population. For
simplicity we assumed a similar death rate for large and small birds. Based on this, we simulated the proportion
of killed birds detected, at dierent search intervals, using the linear regression models of detection probability
over time created before (see Supplementary Fig.S1 for an example). is included repeated searches for each
dead bird, meaning that a bird not detected on the rst search aer its death, would still be in the population
for the next search.
Data availability
Data are available in the Supplementary Data le, but also upon request from the authors (anna.nilsson@nina.no).
Received: 3 July 2023; Accepted: 7 November 2023
References
1. Goodale, M. W. & Milman, A. Cumulative adverse eects of oshore wind energy development on wildlife. J. Environ. Plan. Manag.
59, 1–21 (2016).
2. Po, N. L. et al. e natural ow regime. Bioscience 47, 769–784 (1997).
3. Lawrence, D. Klimaendring og framtidige ommer i Norge. Report No. NVE-Report 81 (2016).
4. Rydell, J. et al. e eect of wind power on birds and bats. In A synthesis. Report 6511 (2012).
5. Drewitt, A. L. & Langston, R. H. Assessing the impacts of wind farms on birds. Ibis 148, 29–42 (2006).
6. Drewitt, A. L. & Langston, R. H. Collision eects of wind-power generators and other obstacles on birds. Ann. N. Y. Acad. Sci.
1134, 233–266 (2008).
7. Everaert, J. & Kuijken, E. Wind Turbines and Birds in Flanders (Belgium) (Research Institute for Nature and Forest, 2007).
8. Hunt, G. & Hunt, T. e trend of golden eagle territory occupancy in the vicinity of the Altamont pass wind resource area: 2005
survey. In Unpublished Report of the California Energy Commission, PIeR Energy-Related Environmental RESEARCH, CeC-500-
2006-056 (2006).
9. Lekuona, J. & Ursúa, C. Avian mortality in wind power plants of Navarra (Northern Spain). In Birds and Wind Farms: Risk Assess-
ment and Mitigation. Quercus. Madrid, Spain 177–192 (2007).
10. Duriez, O., Pilard, P., Saulnier, N., Boudarel, P. & Besnard, A. Windfarm collisions in medium-sized raptors: Even increasing
populations can suer strong demographic impacts. Anim. Conserv. 26, 264–275 (2023).
11. Everaert, J. Collision risk and micro-avoidance rates of birds with wind turbines in Flanders. Bird Study 61, 220–230 (2014).
12. Hull, C., Stark, E., Peruzzo, S. & Sims, C. Avian collisions at two wind farms in Tasmania, Australia: Taxonomic and ecological
characteristics of colliders versus non-colliders. N. Z. J. Zool. 40, 47–62 (2013).
13. Krijgsveld, K. L., Akershoek, K., Schenk, F., Dijk, F. & Dirksen, S. Collision risk of birds with modern large wind turbines. Ardea
97, 357–366 (2009).
14. Falkdalen, U., Lindahl, L. F. & Nygård, T. Fågelundersökningar vid Storruns vindkraanläggning, Jämtland. Report No. 9162065742
(Naturvårdsverket, 2013).
15. Dahl, E. L., Bevanger, K., Nygård, T., Røska, E. & Stokke, B. G. Reduced breeding success in white-tailed eagles at Smøla windfarm,
western Norway, is caused by mortality and displacement. Biol. Conserv. 145, 79–85 (2012).
16. May, R. F. et al. Collision risk in white-tailed eagles. Modelling collision risk using vantage point observations in Smøla wind-power
plant. In NINA Rapport (2010).
17. Marques, A. T. et al. Understanding bird collisions at wind farms: An updated review on the causes and possible mitigation strate-
gies. Biol. Conserv. 179, 40–52 (2014).
18. Jain, A., Kerlinger, P., Curry, R., Slobodnik, L. & Lehman, M. Annual Report for the Maple Ridge Wind Power Project Postconstruc-
tion Bird and Bat Fatality Study–2008 (Report by Curry & Kerlinger LLC, Report for PPM Atlantic Renewable, 2009).
19. Johnson, G. D. et al. Final Report: Avian Monitoring Studies at the Bualo Ridge, Minnesota Wind Resource Area: Results of a 4-Year
Study (Western Ecosystems Technology Inc, 2000).
20. Gr ünkorn, T. et al. (2016).
21. Erickson, W. P., Wolfe, M. M., Bay, K. J., Johnson, D. H. & Gehring, J. L. A comprehensive analysis of small-passerine fatalities
from collision with turbines at wind energy facilities. PLoS ONE 9, e107491 (2014).
22. Rydell, J. et al. Vindkraens påverkan på fåglar och addermöss—Syntesrapport (2011).
23. Aschwanden, J. et al. Bird collisions at wind turbines in a mountainous area related to bird movement intensities measured by
radar. Biol. Conserv. 220, 228–236 (2018).
Content courtesy of Springer Nature, terms of use apply. Rights reserved

8
Vol:.(1234567890)
Scientic Reports | (2023) 13:21365 | https://doi.org/10.1038/s41598-023-46909-z
www.nature.com/scientificreports/
24. Bevanger, K. M. et al. Pre-and Post-construction Studies of Conicts Between Birds and Wind Turbines in Coastal Norway (BirdWind).
Report on Findings 2007–2010. Report No. 620 (2010).
25. R eitan, O. Søk etter døde fugler i Smøla vindpark 2011–2013. Report No. 8242626200 (NINA Rapport 1010, 2014).
26. Stokke, B. G., Nygård, T., Falkdalen, U., Pedersen, H. C. & May, R. Eect of tower base painting on willow ptarmigan collision
rates with wind turbines. Ecol. Evol. 10, 5670–5679 (2020).
27. Bevanger, K., Bakke, Ø. & Engen, S. Corpse removal experiments with willow ptarmigan (Lagopus lagopus) in power-line corridors.
Ökol. der Vögel (Ecol. Birds) 16, 597–607 (1994).
28. Hjernquist, M. B. Eekter på fågellivet vid ett generationsskie av vindkraverk—Kontrollprogram, Näsudden, Gotland 2009–2013
(Havdhem, 2014).
29. Hötker, H., omsen, K.-M. & Köster, H. Impacts on Biodiversity of Exploitation of Renewable Energy Sources: e 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, 2006).
30. Systad, G. H. R. et al. Undersøkelser av trekkaktiviteten for fugl på Guleslettene 2018–2019. Observasjoner, radarkartlegging og lyt-
tedata med supplerende material. Report No. 1693 1–80 (Norwegian Institute for Nature Research, 2019).
31. A lerstam, T. Bird Migration (Cambridge University Press, 1990).
32. Johnson, G. D. et al. Collision mortality of local and migrant birds at a large-scale wind-power development on Bualo Ridge,
Minnesota. Wildl. Soc. Bull. 1, 879–887 (2002).
33. Gritsky, R., Downes, S. & Kronner, K. Klondike III (Phase 1) Wind Power Project Wildlife Monitoring Year One Summary October
2007–October 2008 (Northwest Wildlife Consultants, 2009).
34. Rydell, J., Ottvall, R., Pettersson, S. & Green, M. Vindkraens påverkan på fåglar och addermöss—Uppdaterad syntesrapport 2017
(Naturvårdsverket, 2017).
35. Stokke, B. et al. Langtidseekter av Smøla vindpark på den lokale bestanden av havørn (Haliaeetus albicilla). Report No. NINA Rap-
port 2053 15 (2021).
36. May, R. et al. Paint it black: Ecacy of increased wind turbine rotor blade visibility to reduce avian fatalities. Ecol. Evol. 10,
8927–8935 (2020).
37. Stokke, B. G. et al. Langtidseekter av Smøla vindpark på den lokale bestanden av havørn (Haliaeetus albicilla). Report No. NINA
Rapport 1899 (2020).
38. Svahn, K. & Dahlén, J. Kunskapsunderlag angående fågelkollisioner vid Rögle/Västraby (PM från Enetjärn Natur, 2017).
39. E kelund, S. Sammanställning av undersökningar av fåglar, addermöss samt insekter (Räpplinge vindkrapark, 2015).
40. Nilsson, A. L. K., Molværsmyr, S., Breistøl, A., Hamre, Ø. & Systad, G. H. R. Eekter på fugletrekket over Guleslettene. Etter-
undersøkelser ved Guleslettene vindkraverk. Report No. NINA Rapport 2228 1–46 (Norsk institutt for naturforskning, 2023).
41. Hull, C. & Muir, S. Search areas for monitoring bird and bat carcasses at wind farms using a Monte–Carlo model. Austral. J. Environ.
Manag. 17, 77–87 (2010).
42. Stokke, B. G. et al. Artsgruppeomtale Fugler (Aves). Norsk Rødliste for Arter 2021. https:// artsd ataba nken. no/ rodli stefo rarte r2021/
Artsg ruppe ne/ fugler (2021).
Acknowledgements
e authors are especially grateful for the hard work put in by the carcass survey teams at Guleslettene Wind
Farm: Henning Bortne Bøen, Jan Eirik Håkonsen, Rune Olsbø, and Jan Bjarte Skrøppa. anks also to Andre
Hatleset at Guleslettene Wind Farm for facilitating the o-script experiment presented here.
Author contributions
A.L.K.N., S.M. and A.B. conceived and designed the experiment. S.M., A.B. and A.L.K.N. collected the data.
A.L.K.N., S.M. and A.B. draed the manuscript. A.L.K.N. and S.M. analysed the data, whereas all authors con-
tributed by interpreting the data and the results and gave nal approval for publication.
Funding
e funding was provided by the Research Council of Norway (160022/F40) and also by Guleslettene Wind
Farm AS.
Competing interests
e post-construction study was nanced by Guleslettene Wind Farm AS. e carcass survey teams partook
by registering the dummy birds at the same time as they conducted the carcass surveys, but the experiment
reported here was not required as part of the post-construction study by Norwegian authorities. Data handling
and manuscript preparation was nanced by the Norwegian Institute for Nature Research through the Research
council of Norway (nr. 160022/F40).
Additional information
Supplementary Information e online version contains supplementary material available at https:// doi. org/
10. 1038/ s41598- 023- 46909-z.
Correspondence and requests for materials should be addressed to A.L.K.N.
Reprints and permissions information is available at www.nature.com/reprints.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional aliations.
Content courtesy of Springer Nature, terms of use apply. Rights reserved

9
Vol.:(0123456789)
Scientic Reports | (2023) 13:21365 | https://doi.org/10.1038/s41598-023-46909-z
www.nature.com/scientificreports/
Open Access is article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the
Creative Commons licence, and indicate if changes were made. e images or other third party material in this
article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.
© e Author(s) 2023
Content courtesy of Springer Nature, terms of use apply. Rights reserved

1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com