Aarhus University Seal / Aarhus Universitets segl

No. 232: Second year post-construction monitoring of bats and birds at Wind Turbine Test Centre Østerild

Therkildsen, O.R. & Elmeros, M. (Eds.). 2017. Second year post-construction monitoring of bats and birds at Wind Turbine Test Centre Østerild. Aarhus University, DCE – Danish Centre for Environment and Energy, 142 pp. Scientific Report from DCE – Danish Centre for Environment and Energy No. 232. http://dce2.au.dk/pub/SR232.pdf



In June 2010, the Danish Parliament passed a Public Works Act to establish a national test centre for wind turbines near Østerild in Thy, Denmark. This legislation requires that a bird, bat and vegetation monitoring programme should be implemented. 

Since 2011, the technical facilities at the test centre have gradually been de-veloped and various structures erected on site. The test centre comprises a total of seven test sites for wind turbines of up to a maximum height of 250 m. Each test site consists of a single wind turbine, each with at least one mast for meteorological measuring equipment (up to 150 m in height) located immediately to the west of each turbine. These masts are secured with guy-wires. The test centre also comprises two masts at heights up to 250 m secured with guy-wires. These masts also support aviation safety lighting. 

The Department of Bioscience at Aarhus University was commissioned by the Danish Nature Agency to undertake a monitoring programme of bats and birds in the test area. 


The development of wind energy facilities is a major cause for concern for the conservation of bats. If wind turbines are constructed at locations in or near important bat habitats, e.g. forests and wetlands, or on bat migration routes, the turbines may cause substantial fatalities. If the density of wind turbines is sufficiently high, the cumulative effects of even very low fatality frequencies per turbine per year may affect the status of bat populations. The detrimental impact from wind turbines conflicts with national and international obligations to maintain viable bat populations. Factual knowledge on the conservation conflict is needed to develop ecologically sustainable wind energy facilities. We studied bat activity and behaviour at wind turbines at the national test centre for large wind turbines in Østerild in northwestern Denmark. The test centre was developed in an area dominated by coniferous forest and arable farmland.  

Bats were monitored at the test centre area and on ponds within 2.5 km of the wind turbines to record the presence of bats, to assess the potential effects on populations, and to examine selected aspects of bat interactions with wind turbines and the aggregation of insects to turbines. A pre-construction survey was performed during July-October 2011 when forest clearing in the project area also commenced. Post-construction monitoring and studies were carried out during August-October in 2013 and 2014. Two turbines located in coniferous forest were made the focus of studies of variations in bat activity in relation to distance from wind turbines and insect aggregations. At one of these sites the turbine was operational in 2013, but in 2014 this was only a turbine tower without an active rotor. Bat activity at nacelle height was recorded at these ‘forest turbines’ during September-October 2013 and 2014.


Ten species was recorded in the test centre area. Pond bat (Myotis dascyneme) and Daubenton’s bat (Myotis daubentonii), and Nathusius’ pipistrelle (Pipistrellus nathusii) were the most common species in all years. Soprano pipistrelle (Pipistrellus pygmaeus), common pipistrelle (Pipistrellus pipistrellus), serotine (Eptesicus serotinus), noctule (Nyctalus noctula), Leisler’s bat (Nyctalus leisleri), parti-coloured bat (Vespertilio murinus) and brown long-eared bat (Plecotus auritus) were recorded irregularly.  

Bat activity (number of call sequences per hour) was high at ponds within 2.5 km of turbines and in the test centre, while activity close to turbine sites was relatively low. However, bat activity was significantly higher at the two turbine sites situated in forest compared to activity at three turbine sites in open habitats. Highest bat activity was recorded at both turbine sites and ponds in the first post-construction year (2013). Activity in 2014 dropped to similar levels recorded during the pre-construction survey. The temporal variation in bat activity during the three survey months differed between years. 

Bat activity was significantly greater near turbine towers than along forest edges nearby (50m and 150m) and at a meteorological mast, and bats were observed foraging around the turbine towers. The bat activity at the turbine towers was correlated with insect aggregations around the towers. Insect aggregations and bat activity at the two neighbouring turbines differed despite the turbines being only 600m apart.  

Bats were recorded at nacelle height on four nights during 55 survey nights in September-October 2014. The wind speed was 1.6-7.3 m/s and the temperature was 10-14 °C on those nights.  

No dead bats were recorded during systematic carcass searches with dogs 2015 at selected wind turbines in late summer and autumn during 2013-. However, two dead Nathusius’ pipistrelles were coincidentally found in 2014 at a wind turbine located in forest during the more intensive studies at that site. Similar studies were not carried out at the turbines outside of the forest.

Conclusion and perspectives 

Overall, the habitat changes and the operation of turbines seem not to have altered species presence in the test centre area. Whether the increased bat activity during the first post-construction year was caused by an improved accessibility to the survey area along the large roads, and the subsequent decline was caused by increased disturbance or mortality, or natural variations in bat activity in an area, cannot be determined as no supplementary data on population sizes in the region or comparable data were available from unaffected control areas.  

No fatalities were detected during the systematic carcass searches, so it was not possible to estimate fatality rates associated with turbines. For this reason, the overall number of bat fatalities will probably be low even when the test centre is fully developed. However, we cannot conclude that the wind turbines at the test centre have no significant adverse effects on local or national bat populations. The two coincidentally recorded fatalities of Nathusius’ pipistrelles at the site exceed the critical average level for bat fatalities per turbine per year that has been suggested from population models when considering the cumulative effects of all the wind turbines on a national level. While this threshold is based on German mortality rates for Swedish bat populations, it cannot be excluded that the mortality of even a few individuals from small local populations in Østerild may have detrimental effects on their status. 

Presence-absence surveys are unsuitable for monitoring changes in population size and the effects of wind turbines on bats at local, national and international scales. Declines in species presence may not be detected before substantial declines in populations have occurred. 

The studies in Østerild on bats behaviour around wind turbines demonstrated that: 

  • Bat activity at wind turbines is elevated in coniferous plantations in areas with relatively low bat occurrence. Wind turbines situated in forests are likely to represent more of a threat to the conservation of bat populations than wind turbines constructed in open, barren habitats.
  • Fatality rates may reach critical levels at wind turbines in coniferous plantations in areas with generally low bat occurrence. Construction of turbines in or near other forest types, in areas with larger bat populations and greater bat diversity will represent an even more significant threat to bat populations.
  • Several survey sessions from spring to autumn are needed in both pre- and post-construction periods to record species presence, activity levels and fatality rates at a wind turbine site and assess the risk to the status of bat populations. As bat activity levels may differ between years at a location, pre- and post-construction surveys to assess impacts from wind turbines should span several years, e.g. min. 3 years.
  • Bats activity is elevated around wind turbines compared to the activity detected at distances up to 150m in bat flight routes and potential foraging habitats. This shows that bats are attracted to the turbines and actively explore the structures.
  • The bat activity at the wind turbine towers is correlated with the number of insects settling on the towers, suggesting aggregations of insects in their vicinity appear to attract actively foraging bats, which put the bats at greater risk of collisions.
  • There can be significant differences in bat activity between neighbouring turbine sites. Thus, monitoring of bats must be performed at several sites in a planned wind turbine park and during post-construction to monitor bat activity and fatality risk. 
  • Continuous measurement of bat activity with ultrasound detectors at rotor height and at ground level seems to be a more effective method than carcass surveys to monitor and assess mortality risk. At tall wind turbines towers (>100m) recordings should also be made at medium heights as some bat species are only detectable at <50m.
  • Curtailment of rotor activity during the hours just after dusk and just before dawn on nights with wind speeds <6 m/s at temperatures above 15 °C does not eliminate the risk of wind turbine induced bat fatalities.  

To better ensure ecologically sustainable wind energy facilities, further more intensive studies are needed to develop quantitative tools to assess fatality risk at wind turbines and their effects on bat populations.  

The results from the Østerild studies regarding bat activity levels and mortality rates at wind turbines in forested areas cannot be extrapolated to other bat species or wind turbines in different landscapes, habitats, and forest types in other regions of Denmark.  

Parallel studies are needed on wind turbines in landscapes with different habitat characteristics, higher bat diversity and activity levels to assess fatality risks and impacts on populations. Studies at a wider size range of turbines are also needed to estimate the relationship between bat fatality risk and turbine size.  


The monitoring programme for birds comprises one baseline (2011/12) and two post-construction study periods (2013/14 and 2015/16). 

The test centre is located near several Special Protection Areas (SPAs), which are sites designated for their particular importance for birds. These SPAs have been classified for rare and vulnerable breeding birds (as listed on Annex I of the Directive) as well as for regularly occurring migratory species according to Article 4.2 of the EC Birds Directive and generally following the criteria for designation of wetlands of international importance. As a result of their high conservation interest the monitoring programme has focused on this group of species in both the baseline and post-construction studies. 

In 2012, we presented the results from the baseline monitoring programme. On the basis of a preliminary assessment, we considered the potential impacts of the combined structures on the bird species occurring in the study area unlikely to be significant. 

In 2015, we presented the results from the first year of the post-construction bird studies, which were carried out from August 2013 to October 2014, together with an intermediate assessment of the potential impacts of the test centre on the bird populations occurring in the study area. 

Here we present the results from the second year of the post-construction monitoring programme, which was carried out from August 2015 until August 2016, together with a final assessment of the potential impacts of the test centre on the bird populations occurring in the study area. During the second year post-construction study period the test centre reached full capacity with 7 wind turbines in operation.  

Initially, pink-footed goose, taiga bean goose, whooper swan and common crane, were included in the baseline investigations. However, on the basis of the results obtained during the baseline studies, light-bellied brent goose, white-tailed eagle and nightjar were also subsequently included as focal species in the post-construction programme.   

Apart from minor modifications and special efforts targeted towards light-bellied brent goose and nightjar, the design of the post-construction study was similar to the baseline study, which aimed at generating species-specific data, whenever this was technically possible. For this reason, although data were partly collated from comprehensive automated recording processes, the collection of high quality and high resolution data at the species level was given priority at all times in the investigations. We used visual transect counts and laser range finder data, which was combined to provide the basic information for the assessment. In addition, we conducted carcass searches using trained dogs under turbines and masts to quantify actual fatality rates.  

In general, the second year post-construction study supported the conclusions from the previous study years. We confirmed that the test centre is not situated on a migration corridor, although seasonal migration took place to some extent. During the day, flight activity in the study area was dominated by local birds moving between feeding areas and night roosts in northwest Jutland, some of which has been designated as SPAs for the species included in the study. We demonstrated local movements to take place on a regular basis for a number of species, which was the case in the previous study years. 

From the results of the baseline study, the species for which we estimated that more than one annual collision with wind turbines would take place were cormorant (3 individuals per year), pink-footed goose (21-46), greylag goose (3-6) and golden plover (65).   

Based on the first year post-construction study, we estimated that the annual collision rate with wind turbines that exceeded one would be for cormorant (6-14), pink-footed goose (10-23), greylag goose (23-52), buzzard (0.8-1.6), golden plover (3-7), wood pigeon (0.5-1.2) and passerines (3-5). 

From the results of the second year post-construction study, the species for which we estimated that more than one annual collision with wind turbines would take place were cormorant (7-15), whooper swan (2-5), pink-footed goose (14-31), greylag goose (19-44), kestrel (0.71-1.60), buzzard (1.2-2.7), common crane (0.6-1.3), golden plover (7-15), wood pigeon (7-17) and passerines (7-19). 

For all of these species, a high proportion of individuals passing the study area did so at rotor height.  Nevertheless, this still only resulted in a relatively limited number of predicted collisions even for these species. It is also important to note that in contrast to the baseline study, the post-construction study period covered the whole annual cycle, except for June-July. 

For the remainder of the species that regularly occur in the study area, including the focal species taiga bean goose and white-tailed eagle, we predicted that the annual number of collisions would be less than one. This was typically because, for these species, a high proportion of individuals and flocks migrating occurred at flight altitudes below the rotor height of the wind turbines. The amount of data collected was not sufficient to estimate collision numbers for light-bellied brent goose. 

In summer 2015, we used miniature GPS data loggers to track movements of nightjars to investigate the extent to which they forage in the proximity of wind turbines in the study area. Unfortunately, the GPS data obtained from the single male, which we were able to track for three nights, was insufficient to draw any conclusions with regard to the foraging patterns of nightjars and associated risk of collision between wind turbines and masts. 

Only one bird, a goshawk, was retrieved during the carcass searches. We were unable to determine whether the fatality was caused by a collision. No other carcasses were found during the searches. We consider the almost complete absence of collisions between birds and the structures at the test centre to be highly unlikely. We therefore assume that either some fatalities were not detected because their remains were not available or missed by the dogs or they were removed by scavengers between searches.  Nevertheless, the results from the carcass searches indicate that the number of collisions is probably rather small. The results from the carcass searches therefore support our conclusion that although collisions between turbines and other structures at the test centre are to be expected, they will occur at a low rate. 

Our investigations demonstrated that many species showed vertical and horizontal avoidance in relation to wind turbines and measuring masts. This active avoidance response may also explain some of the discrepancy between collision estimates obtained by the Band method and the lack of fatalities found during the carcass searches. The Band model assumes a uniform distribution of bird flights in the area. This is opposed to our analyses which indicate that several species actively avoid the wind turbines and measuring masts. Hence, our analysis suggests that the Band model has a tendency to overestimate collision risk. 

On the basis of this final assessment, which uses more reliable estimates of collision risk based on two post-construction study years, we still consider that the potential impacts of the combined structures on the bird species occurring in the study area are unlikely to be significant. We stress that our crude estimates of the number of collisions should be interpreted with caution including comparison of collision estimates between the three study periods.  

Since the test centre had only four turbines in operation during the first year post-construction study period, the previous assessment did consider a fully developed test centre. During the second year post-construction study period the test centre became fully developed with seven turbines simultaneously in operation. The presence of more turbines had only limited effect on our estimates of the risk of collisions with turbines. We are therefore relatively confident to conclude that the overall impact of the test centre on bird species is considered unlikely to be significant.  

For three of our focal species, which are rare breeders in the study area, i.e. white-tailed eagle, common crane and nightjar, a single fatality will inevitably have negative impact on the local and regional populations. However, we consider this potential impact on the population to be short-term. At least for common crane and white-tailed eagle the continued growth of the populations will make them more resilient to added mortality from wind turbines and other human pressures in the future.  

We therefore recommend that the mortality related to human developments on the white-tailed eagle, common crane and nightjar populations, particularly the impact of the continued development of wind energy in the region, is closely monitored in the future. 

It is important to keep in mind that the data collected during the baseline and the post-construction programmes only covers less than three years. We are therefore cautious when we assess the extent to which there may be year-to-year variation in the occurrence of birds both during night and day. In particular, different weather conditions can affect flight behaviour and migration pathways, which may affect the risk of collisions.