Impact of a Wind Farm on the Avifauna of a Mediterranean Mountainous Environment ^†

Abstract

Possible avian mortality rates due to the presence of a wind farm on Kefalonia Island, Greece, were studied and no bird deaths were recorded. The most commonly occurring species were Buteo buteo and Falco tinnunculus. These species generally have low ecological sensitivity. During the approach of individuals (predators or other large species) close to the turbine area, either no reaction or a slight change in direction by the birds was observed. Similarly, the most frequent type of interaction was flight between pairs of turbines, while less frequent was flight parallel to or above the turbines.

1. Introduction

Since the first years of the implementation of wind energy internationally (1980), it has been observed that the installation and operation of wind farms, along with accompanying works (e.g., road opening, building installations, fences, and above-ground electricity transmission cables), can have an impact on biodiversity, especially on avifauna [1,2]. The impacts of a wind farm on birds are extremely diverse and depend on a wide range of parameters, such as the characteristics of the project, topography of the area, habitats affected, and number and species of birds observed in the area. [3]. Four main impacts have been identified by previous studies [1,2,4,5,6]: (i) disturbance—displacement of birds from the wind farm zone (indirect habitat loss) due to factors such as noise or visual disturbance; (ii) collision—birds are killed or injured as a result of direct contact with rotating propellers or the tower; (iii) barrier effect due to obstacles preventing the movement of birds; and (iv) direct habitat loss or change in habitat structure.

The majority of studies related to the impacts of wind turbines have recorded relatively low mortality levels [7,8,9]. Mortality levels vary greatly between species and regions. The risk of impact depends on several factors that can be grouped into three main categories [10]: species-specific factors, site-specific factors, and wind farm-specific factors. A recent study of 15 wind farms [11] concluded that the important impact parameters were orientation (higher mortality in NW-oriented locations), vegetation type (higher mortality in shrublands), and slope (higher mortality in relatively flat areas). The risk is higher in areas that are regularly used by a large number of birds for feeding or roosting, or in migratory and local flight corridors, especially when they are blocked by wind turbines [1].

The bird activity type increases or decreases the likelihood of their collision with wind farms. Impact incidents are more common for birds during reproduction, wintering, and station-over than for birds passing through during migration, which is probably related to the lower activity levels and higher flight altitudes of migratory birds [10,12]. Small passerines and other land-based species are less affected by wind farms [13] but are more sensitive during migration [10]. In a survey of the morphological characteristics of the victims of a wind farm in Mexico [14], which accounted for the behavior and flight patterns of species, collisions were defined by two distinct steps: some birds are more likely to fly near the turbines and a subset of these birds are more likely to be killed by a turbine. Species that are most likely to fly near a propeller tend to have a larger wingspan and wing load (wing loading). Among these birds, the smallest in size have a higher risk of collision owing to the smaller opening and load of their wings. [14]. This is, to a certain extent, contradictory as birds with these characteristics (small size and wing load, etc.) are species with good flying and maneuvering abilities. However, in this case, the increased sensitivity may be related to species with these characteristics spending a long time in the air in search of food.

In many European countries, extensive research has been conducted on the effects of wind turbines on avifauna. In an extensive analysis of 44 wind farms in Portugal (2005–2015) [11], the most common species that were killed were Delichon urbicum, Apus rubicund, Alauda arvensis, Lullula arborea, Sylvia undata, Alectoris rufa, Buteo buteo, Gyps fulvus, Falco tinnunculus, Ficedula hypoleuca, Phlloscopus collybita, and Circus pygargus. These species mainly feed in the air in flocks (swallows/ashtrays may also be attracted to turbines to feed on insects). Species of the family Alaudidae, which also create flocks (outside of the breeding season), make flights at a great height during the breeding season, and their habitat often coincides with areas of wind farm development. Generally, it seems that species of the Alaudidae family show increased susceptibility to collisions [15,16,17]. In a particularly extensive four-year survey of 46 wind farms at low altitudes in Germany [17], 291 bird killings were identified. The most common species observed in this survey were Columba palumbus, Anas platyrhynchos, Buteo buteo, Vanellus vanellus, Pluvialis apricaria, Milvus milvus, and F. tinnunculus. By extrapolating the results to a wider region, 0.4% of the reproduced population for C. palumbus, 4.5% of A. platyrhynchos, and 7% of B. buteo, while population effects may occur for the species B. buteo and M. milvus. According to another database, the most common victims of wind farms in Germany are B. buteo, M. milvus, and Haliaeetus albicilla [18].

The wind farm observed in the current study produces 2.7 MW of power and is located on Mount Evmorfia, Kefalonia Island, Greece. The wind farm consists of three Enercon turbines (height = 55 m) built on a plateau measuring 45 × 45 m. A road 2 km in length and 5 m in width was constructed, as well as other facilities. A three-year monitoring program (2018–2020) was implemented to record changes in bird density and use of space, seasonal abundance, rate and spatial patterns of flight through the turbines, and any other element of bird behavior and distribution which may be due to the operation of the wind farm.

2. Materials and Methods

The monitoring and recording of changes in the density and use of space, seasonal abundance, rates and spatial patterns of passages, and any other element of bird behavior and distribution which may be due to the operation of a wind farm was the subject of this research. This evaluation considered the general conservation objectives that have been set, as well as the existing reference values concerning the protected area GR 2220006. During research implementation, the specifications of relevant works conducted by the WWF in the Thrace region were taken into account to a significant extent [19,20].

2.1. Carcass Search Studies

Systematic research was carried out on dead birds (or bats) at the three wind turbines in the study area. The research was conducted in a circular area with a radius of 60 m around each turbine, which was at the center of the circle [21]. The aim of search activities was to find corpses or injured birds of any species that may have been killed or injured owing to interactions with the wind farm. During every visit, a special protocol was implemented for the carcass search.

To minimize errors due to the removal of dead birds by carnivorous animals, which is common during mortality calculation studies, particular attention was paid to search frequency. A visit/sample rate of 10 d ± 3 was used to investigate possible mortality (carcass search). This sampling interval/frequency was considered sufficient to examine the impact of possible killings on predatory species [19], which was the focus of this study (i.e., species more sensitive to the effects of wind farms). According to the experiments and tests carried out by the WWF in the Thrace region, the average time during which a dead bird remained in the area before being removed by a predator was 23 days. However, 50% of small corpses, 22% of medium-sized corpses, and 25% of large corpses had been removed after 14 days. Thus, it was considered that the search frequency of this study detected most of the mortality of predatory bird species (and passerines) in the study area. Monitoring visits during the first year (3/2018–3/2019) revealed that for passerines, but mainly for predators, the abundance/activity of bird species was markedly lower during the winter. Therefore, it was assumed that the monitoring program could adequately and effectively cover the period from March to early November.

2.2. Monitoring the Use of Space by Birds from View Points

Vantage point counts were applied to study bird flight patterns, space use, dangerous flights, and behavior of large predators in the study area during the operational phase of the wind farm. One point of view was used in this study. For all bird flights in the study area, the movement of a bird and its characteristics (type and height of flight, location on a map) were studied. In cases where the movement occurred in Zone A (within 250 m from turbine positions), additional data were recorded on the behavior/interaction of the birds in relation to the turbines. The activity of predators and other large species from the point of view was investigated once for every month of the monitoring year (eight times in total), and each count was six hours in duration (48 h in total for every monitoring year)

2.3. Bird Abundance Monitoring

The implementation of the line transect method was based on the technique described by Bibby et al. [22]. Line transect method counts all birds that are heard or found around 25 m from the observer position. Indications of reproduction were also recorded (e.g., nest, food movement or nest materials, male singing) as well as the distance from a turbine. Once a month, a linear transect was carried out around the wind farm to obtain data on the abundance of passerine birds.

2.4. Data Analysis and Impact Assessment

With regard to mortality data (search for victims), an N-estimated mortality assessment can be calculated in accordance with the methodology followed by Carcamo, B., et al. [19], which is based on Equation (1) [23]:

N-estimated = Na × Cz × Cp × Ce

(1)

where Na is the number of deaths (corpses) due to collision; Cz is the correction factor for the study area (Cz = 100/z, where z is the percentage of the total area investigated); Cp is the correction factor for the removal of corpses by cadaveric animals (Cp = 100/p, where p is the percentage of birds that were not removed by animals during the tests); and Ce is the correction factor for the ability of the observers to locate corpses (Ce = 100/e, where e is the percentage of birds found by observers).

Finally, the collected activity and abundance data were used to create additional data (tables), for example: number of observations and bird species per month, summary of interactions, crossings per pair of wind farms, and nearest turbine. The evaluation also considered the ecological sensitivity of each species (protection status, rarity, threat) and susceptibility to the effects of wind farms according to the international literature [1].

3. Results

During the investigation of possible mortality due to collisions, no findings were observed in the study area (bird/bat carcasses or injured birds). This indicates that, in the context of systematic and intensive investigation in accordance with international standards, there was zero mortality for the study period. Concerning the international literature on wind farms with low effects, some mortality of passerine species has often been recorded. In our study area, bird species were present in proximity to the wind turbines, including species with sensitivity to impact effects based on the literature: B. buteo, F. tinnunculus, A. arvensis, and A. apus. However, it appears that the levels of bird abundance and activity were low; therefore, the likelihood of the occurrence of a collision was relatively low.

During the field monitoring for abundance and activity conducted during 2013/2014, a total of 35 bird species were observed in the project research area. In the 2020 monitoring year, a total of 12 species were observed compared to the 14 observed in 2019 monitoring year and the 22 observed in the 2018/2019 monitoring period (the difference may be because monitoring in the period 2018/2019 lasted 12 months, whereas it lasted 8 months in the years 2019 and 2020). Species with a permanent presence (local) and reproducing visitors were observed during the study period (

Table 1. Bird species of the study area, identified during the monitoring period.

Scientific Name	2009/147/ΕΕ	Presence Status	GR2220006	March	April	May	June	July	August	September	October
Alectoris chukar	-	R		++	+	++	++	++	+	++	++
Anthus trivialis		P									+
Apus apus		BP	50–100p				+
Aquila chrysaetos	Ι	R	1i		+					+
Buteo buteo		RPW	9–11i	+++	+	+++	++	++	+	+++	+
Carduelis cannabina		BP				+	+
Circaetus gallicus	Ι	BP	3p	+		++	++
Corvus corax		R			++	+		+	+	+	+
Corvus corone	II2	R		+	+	+	+	+			+
Erithacus rubecula		W	+	+	+
Falco tinnunculus		R						+	+	+
Fringilla coelebs		RW	+	+
Galerida cristata		R						+
Hirundo rustica		BP	+	+
Linaria cannabina		PB		+	+	++	++	+	+
Lullula arborea	I	RW	+	++				+			+
Oenanthe hispanica		BP	200–300p		+	++	++	+	+	+
Oenanthe oenanthe		BP	+	+	+++	+++	+++	+++	+++	+++	+
Pernis apivorus	Ι	PB	1–2p						+
Phoenicurus ochruros		WP	+	+							+
Saxicola torquatus		BP				+
Tachymarptis melba		BP	30–50p		++	+++	+	++		+	+

Presence status: Status of species occurrence in the study area, (R) Resident; (B) Breeding; (W) Wintering; (P) Passage. 2009/147/ΕΕ: Species of Annexes Ι & ΙΙ of Directive 2009/147/ΕΕ for bird protection. GR2220006: Birds of Natura 2000 area GR2220006 (presence/population, individuals/I, or couples/p), + corresponds to <10 bird records/per month, ++ corresponds to 10–20 bird records/per month, +++ corresponds to >20 birds records/per month.

). The presence of local species or populations was low (mainly Corvus cornix, Phoenicurus ochruros, and L. arborea), whereas migratory activity was negligible. The rarest and most protected species of the study area, Falco biarmicus, was not observed, while in the other precious species of the protected area, only Circaetus gallicus was observed.

Regarding the large species that were observed during the point of view counts, four species were most frequently present: B. buteo, F. tinnunculus, Corvus corone cornix, and Corvus corax. In contrast, C. gallicus and Aquila chrysaetos were sporadically present with 1–2 observations recorded. During the 2018/2019 monitoring period, a single A. chrysaetos was observed at a high altitude (5/2018, 500 m alt.), as was C. gallicus (one pass at a great height on 3 June 2018 flying to Mt. Ag. Dynati, one pass on 28 May 2018 also at a great height, and an observation on 26 June 2019). In addition, in the 2019–2020 period, A. chrysaetos was observed on 27 September 2020 at an average height, and individual B. buteo and C. gallicus were observed on 9 May 2020 at a great height. These two species did not nest in proximity (they were observed with negligible frequency/activity while there was no suitable nesting habitat), nor were they observed in the study area collecting food. In 2019, C. gallicus was observed in May and June.

The predatory species with the most observations (all three follow-up periods) was B. buteo (non-Annex I species of Directive 2009/147/EU), which was observed during most of the recording months (not observed in winter). During the monitoring period of 3/2020–10/2020, three crossings were observed at the wind farm (all three in Zone A), with two crossings parallel with pairs of turbines and one far from the wind farm. The frequency of observations was greatest in 2013/2014 (when the wind farm was in operation) with 0.24 crossings per hour, compared to 2018/2019 with 0.17, 3/2019–10/2019 with 0.06, and 3/2020–10/2020 (

Table 2. Summary of bird passages (per month) based on point of view records.

Species	Total	Total Passages Per Hour (2018/2019/2020)	Total Passages Per Hour (2013/2014)
Buteo buteo	18	0.167	0.242
Falco tinnunculus	9	0.028	0.267
Corvus corone	13	0.181	0.183
Corvus corax	3	0.042	0.067
Aquila chrysaetos	2	0.014	0
Circaetus gallicus	2	0.014	0

).

Falco tinnunculus (a non-Annex I species of Directive 2009/147/EU) was observed making 6 transits of the study area in the 3/2019–10/2019 period (0.12 crossings per hour), and 2 transits in the 2018/2019 period (0.028 crossings per hour) compared to 0.267 in 2013/2014 (Table 2). Crow species such as C. corax and C. corone cornix were often observed near the wind farm in the 3/2020–10/2020 period, but as they are not considered birds of prey, their presence was not evaluated further.

Regarding the interactions (behavior) of birds entering Zone A (

Table 3. Interactions of bird species with the wind farm (in Zone A) expressed as numbers of individuals per interaction category (for the period 3/2018–10/2020).

	2018/2019					3/2019–10/2019					3/2020–10/2020
	1	2	3	4	5	1	2	3	4	5	1	2	3	4	5
Buteo buteo	1	4	5	2			1	2			1	2
Corvus corone	3		10
Corvus corax		2	1
Falco tinnunculus		1	1			1	1	4			1
Aquila chrysaetos				1								1
Circaetus gallicus				1
Sum	4	7	17	4	0	1	2	6	0	0	2	3	0	0	0

1: The bird flies away from the wind farm, no interaction. 2: The bird flies parallel to the wind farm or approaches the wind farm without passing between turbines. 3: The bird crosses the area between 2 turbines (or passes near the last turbine). 4: The bird crosses the area between 2 turbines, but flies much higher than the height of the turbine. 5: The bird passes between the turbines of the wind farm.

), the most common behavior in 2020 was Category 2 (the bird flies parallel to the wind farm or approaches it without passing between turbines) with 60%, whereas Category 3 (the bird crosses the area between two turbines or passes near the last turbine) was the most common in previous years (53% and 66% in 2018/2019 and 3–10/2019, respectively). The frequencies of Category 1 (bird flies away from the wind farm, no interaction) and Category 4 (flight at high altitude) were lower.

4. Discussion

The results showed that bird species that are sensitive to impact, including B. buteo, F. tinnunculus, A. arvensis, and A. apus, flew near the wind turbines on Mount Evmorfia. However, bird abundance and activity levels were low; therefore, the probability of collision was low. During the three-year monitoring survey, 22 species (mainly permanent or breeding visitors) were observed in the field research area within a radius of 2 km from the wind turbines. The most important species of the Natura 2000 site, F. biarmicus, was not observed. During monitoring, sporadic crossings of three prey species of importance to the area (A. chrysaetos, C. gallicus, and Pernis apivorus) and protected species included in Annex I of the European Union Birds Directive were observed. These are breeding species on the island, which do not breed in the study area nor visit it for food collection (observed either at a great distance from the wind turbines or at a great height above them during movements to/from other locations). During the counts in 2020, C. gallicus and Aquilla chrysaetos were observed in two cases. Buteo buteo and F. tinnunculus were more frequently observed flying in the study area. These are species with generally low ecological sensitivity and are not included in Annex I of the European Union Birds Directive. When birds of prey or other large carnivorous birds flew close to wind turbines, either no reaction or a small change in direction was observed (the second type of reaction occurred only during the first year of monitoring). Similarly, the most common type of interaction was flight between pairs of wind turbines, whereas the least frequent was parallel flight or flight at a higher altitude. During the three monitoring periods, crossings between the pair of the first and second wind turbines were more frequent than those between the pair of the second and third wind turbines, which may be because the space in the first case is larger. Regarding passerine birds or other small/land species active in the area of the wind farm, the presence of the nesting bird Oenanthe oenanthe was observed during all three monitoring periods.

5. Conclusions

This study includes the results of three-year monitoring actions during the operation of a wind farm on Mount Evmorfia, Kefalonia Island, Greece. As discussed above, no findings were detected during the investigation of possible mortality due to impacts of wind farms. In the context of the systematic and intensive investigation in accordance with international standards, zero mortality was observed for the study period. This finding is particularly important as the international literature shows that even wind farms with low effects often record some mortality of passerine species.