Assessment of bat and avian mortality at a pilot wind turbine at Coega, Port Elizabeth, Eastern Cape, South Africa

Abstract

A recent push for wind energy development in South Africa has led to bat and avifaunal monitoring programs throughout the country. During the course of one year, three inspections per week (total 154 inspections) were made for evidence of bat and bird casualties at a pilot wind turbine in the Coega Industrial Development Zone, Port Elizabeth, Eastern Cape, South Africa. We recorded 17 bat fatalities and one live but injured bat was collected. Two bat species were involved, Cape serotine (Neoromicia capensis) and Egyptian free-tailed bat (Tadarida aegyptiaca). Of the 18 casualties, 15 were recorded mid-December to mid-March. One bird, a little swift (Apus affinis), was hit by a rotor blade. This is the first study to document bat and bird mortalities over the period of a year at a wind turbine in sub-Saharan Africa.

Keywords:

• turbine
• fatality
• wind farm
• bat
• avian
• South Africa

View correction statement:

Erratum

Introduction

South Africa developed a White Paper on renewable energy in 2003 in which it set itself a renewable energy target of 10,000 GWh of energy to be produced from renewable energy sources. Consequently many independent power producers are submitting proposals to establish wind farms, particularly in the coastal areas of the Western and Eastern Cape Provinces and the first of these are expected to become operational during 2013. However, the toll wind turbines take on wildlife, particularly birds and bats in relation to onshore wind turbines, is well documented (Howe et al. 2002; Jain et al. 2011). Bats, particularly migratory species and those that rely on tree roosts, are susceptible to injury or mortality by barotrauma or direct collision with rotor blades (Erickson et al. 2002; Cryan & Barclay 2009) and studies have shown that bat casualties often outnumber avian mortality rates (Barclay et al. 2007). Both local and migratory bird species are susceptible to direct collision with rotor blades (Krijgsveld et al. 2009) and population declines due to habitat loss during construction have been recorded (Pearce-Higgins et al. 2012). Best practice guidelines for avian and bat monitoring at wind farms in South Africa have been published (Jenkins et al. 2011; Sowler & Stoffberg 2011). However, to date, no published studies have documented fatalities due to wind turbines involving birds or bats in South Africa. A pilot wind turbine was erected in June 2010 in Zone 9 of the Coega Industrial Development Zone (IDZ), Nelson Mandela Bay Municipality, South Africa. This is the only large wind turbine to be erected in the Eastern Cape to date. The turbine is a Vestas V90 turbine coloured white with a hub height of 95 m and a generation capacity of 1.8 MW. Three rotor blades are present with a radius of 45 m giving a total height of 140 m. This study increases the available international knowledge on bird and bat mortality due to wind energy turbines.

Methodology

Mortality observations at the pilot wind turbine

The wind turbine is located on the east side of the Coega River, set back from a ridge (33^o45′157′′S 25^o40′597′′E). A concrete base approximately 20 m in diameter surrounds the turbine, and there is a 60 m×30 m gravel laydown area to the northwest and a gravel road on the northwest leads to and past the turbine. All other areas are surrounded by vegetation consisting of a matrix of Sundays valley thicket, comprising dense evergreen bush 2–3 m in height with a high diversity of plant species and Grassridge bontveld comprising grass, low shrubs and succulents (Stewart & Reeves 2009). The Coega River and nearby clay quarry are possible sources of water for both birds and bats, and bats may further be attracted to the array of insects (Long et al. 2011) at the river.

The area around the pilot turbine was carefully searched for evidence of bat and bird casualties over the course of an hour on a Monday, Wednesday and Friday of each week. The area was searched in a spiral pattern, starting at the base of the turbine in approximate incremental steps of 3 m between spirals. However, no measuring devices were used to denote the accuracy or ensure repetition of searches. A total of 154 inspections were undertaken between 4 March 2011 and 12 March 2012. Most of the search time was spent within a radius of 60–70 m of the turbine. Substantial growth of vegetation in winter 2011 made finding carcasses amongst the vegetation more difficult as the project progressed. When a casualty was found, the description of position, GPS co-ordinate, distance from turbine base, direction from turbine door, state of the carcass and approximate time of death were recorded. Species, sex, injuries and predicted cause of death were determined by postmortem.

Carcass experiments

A total of 34 bird carcasses were placed in the Coega IDZ and monitored to see how quickly they were scavenged by predators and how long evidence of the carcass remained visible to an observer. Fresh carcasses were frozen and thawed overnight prior to placing. During the winter period of 2011 (May–September) one laughing dove (Streptopelia senegalensis), 20 feral pigeons (Columba livea), three speckled pigeons (Columba guinea) and one white-necked raven (Corvus albicollis), were placed in various positions typical of the environment of the area (in the open, in long grass, next to or under bushes). The number of carcasses used per season was determined by availability. The fresh wing and feathers of a Hadeda ibis, (Bostrychia hagedash) that had been scavenged below an 11 kV powerline was also monitored. During the summer period (December 2011–January 2012) carcasses of five S. senegalensis, a female Cape sparrow (Passer melanurus), a Cape crow, (Corvus capensis) that likely had been electrocuted and a little swift (Apus affinis) casualty at the turbine were monitored. Carcasses were monitored every few days. Scavenging by mammalian predators was indicated by bitten off, rather than plucked, feathers (Hardey et al. 2006; Thirgood et al. 1998).

Results

Mortality observations at the pilot wind turbine

Bats

We recorded 18 bat casualties at the pilot turbine on 11 occasions (Fig. 1). Casualties involved two insectivorous bat species, 10 Cape serotine (Neoromicia capensis) (five male, five female) and eight Egyptian free-tailed bats (Tadarida aegyptiaca) (four male, three female, one unknown). Due to the difficulty of analysing bat carcasses to conclusively determine cause of death, we grouped all bats into two casualty categories, presence or absence of visible injury, and measured the distance of the carcass from the turbine base (Table 1). We observed nine instances of visible injury and six instances of no visible injury. Two instances of death could not be assessed for injury because carcasses had been run over by vehicles on the ground prior to discovery. We predict that barotrauma may be the cause of mortality in those bats with no visible injury. The injured bat was cared for and released after one month.

Figure 1  Number of bat casualties found at the pilot turbine, Coega Industrial Development Zone between March 2011 and March 2012.

Table 1  Summary of bat mortality findings from the pilot turbine at the Coega Industrial Development Zone between March 2011 and March 2012.

	Egyptian free-tailed	Cape serotine
	Tadarida aegyptiaca	Neoromicia capensis
Number of casualties	8	10
Median distance from tower (m)	13.6	18
Min distance from tower (m)	4.2	4
Max distance from tower (m)	39	40

Birds and insects

One bird casualty, an adult little swift (Apus affinis) of undetermined sex, was found 30 m away from the pilot turbine. The bird had suffered major trauma, with its right wing torn off. Flocks of Apus spp. were sometimes observed feeding above the turbine, very close to the blade tips. Single feathers were found near the pilot turbine on nine occasions. It is not thought that these represent bird casualties.

Lying on the ground below the pilot turbine there were often several insects, either dead or with some limbs missing, indicating contact with a rotor. Lepidoptera, Coleoptera mostly of the family Scarabaeidae, and Pyrgomorphidae were frequently encountered.

Carcass experiments

During winter, carcasses lasted a median of 7 days before being found by a scavenger (range 1–19 days, n=26) and during summer a median of 4 days (range 3–15 days, n=8). Of those carcasses placed in the vicinity of the turbine (n=6), carcasses were scavenged between 7–9 days (n=3) during winter and during summer between 3–4 days (n=3). Three of the 34 carcasses (8.8%) were scavenged within 2 days and a further five, total eight (23.5%) within 3 days of being placed. Of the 34 bird carcasses monitored, 19 (56%) were removed without trace and bitten off feather remains indicated scavenging by a mammal in the remaining 14 cases and one carcass decomposed. In two cases a diurnal mammal was known to be responsible as the carcass was checked morning and evening.

Discussion

This study provides the first published account of bat and bird casualties over the period of at least a year at a wind turbine in sub-Saharan Africa. The observed casualties of 18 bats over the course of a year indicates that wind turbines in the Eastern Cape could pose a serious risk to open-air and clutter-edge bat species, such as T. aegyptiaca and N. capensis. Neither of the species found at the turbine are listed in the threatened species categories of the IUCN or South African Red List of Threatened Species (Friedmann & Daly 2004; Jacobs et al. 2008; Mickleburgh et al. 2008).

Little is known of the migratory habits or movements of bats in South Africa, and although not ideal, bat fatalities at wind turbines may provide information on the migration or physiological habits and adaptations of at-risk species. It appears that both N. capensis and T. aegyptiaca are not present in the colder months of the year, mid-April to October, indicating either seasonal movement in both species or deep torpor or hibernation in response to a decrease in insect abundance near turbines. Most bat casualties (83.3%) were found mid-December to mid-March. The bat mortality rate of 18 is within the worldwide range of mortality rates previously researched. In Australia, 22 white-striped mastiff bats (Tadarida australis) were found over a period of four years (Hall & Richards 1972). In the US, bat mortality rates range from corrected rates of 0 bats/turbine/year (Erickson et al. 2004) to 42.7 bats/turbine/year (Kerns & Kerlinger 2004). In northwest Europe, the estimated rate of bat fatalities ranged from 0 to 20 bats/turbine/year (Rydell et al. 2010) although studies that utilise estimated rates should be carefully considered due to the potential for a large discrepancy between estimated and realised mortality rates (Ferrer et al. 2012).

The fatality search methodologies described are similar to those of Johnson et al. (2003) and Baerwald and Barclay (2011) in that the discovered individual's species, age, sex, degree of composition and distance from turbine base were recorded and searches were conducted using walking transects. Johnson et al. (2003) used a square plot search methodology, while we walked around the based of the turbine at approximate 3 m intervals. We searched at the turbine three times per week, while Jain et al. (2011) searched turbines either every 2 or 3 days, and Baerwald and Barclay (2011) searched 10 turbines every day and the remaining turbines once a week. We similarly performed a scavenging experiment, but did not perform a searcher efficiency trial and thus were unable to indicate a corrected fatality rate using the proposed equations by Baerwald and Barclay (2011), Johnson et al. (2003) or Smallwood (2007).

Bat fatality at this particular turbine is of concern, as reproductive capacity of the species in question is restricted to one pup a year in T. aegyptiaca (Monadjem et al. 2010) and two pups per year in N. capensis (van der Merwe 1994). There are plans for wind farms in many areas of the Eastern and Western Cape, South Africa and this study indicates that large numbers of wind turbines could have a detrimental effect on bat populations in these areas.

The single bird mortality recorded at this particular turbine is within the range of 0.36 to 7.70 birds/turbine/year recorded in Canada and the US, where the mean casualty rate is 2.16 birds/turbine/year (Brown & Hamilton 2004; National Research Council 2007).

The carcass scavenging experiments indicated that 8.8% of carcasses were scavenged within 2 days and 23.5% within 3 days. By searching for turbine mortalities on Mondays, Wednesdays and Fridays, even if a bat became a casualty the following evening, it is likely that less than 10% of casualties were scavenged prior to the searches on Wednesdays and Fridays and less than 25% prior to the search on Mondays. The accelerated rate of decomposition in summer presumably allowed mammalian predators to locate the carcasses by smell much sooner. There are several mammalian predators in the area. It is likely that small grey mongoose (Galerella pulverulenta) and yellow mongoose (Cynictis penicillata) were responsible for scavenging most carcasses during this study. Mongoose scats were found around the base of the pilot turbine and a yellow mongoose was seen close to a partially eaten carcass in another part of the study area. If mortality at a wind turbine becomes frequent, scavengers are likely to patrol the area around the turbine on a frequent basis. The quick scavenging rate (median 7 days in winter, 4 days in summer) and possible habituation of scavengers to a potential food source has consequences for the frequency of monitoring required to obtain reasonable estimates of mortality rates at wind farms in the Eastern Cape.

Similar carcass scavenging rates to those observed.during this study have been found in some other studies. When one to five birds were placed at 52 different turbines in a 247.9 ha study area in California, 50 of 63 caracasses were scavenged and on average first removal occurred after 4.45 days (Smallwood et al. 2010). In Spain, 522 bird carcasses were placed under power lines wherein 32% were removed within the first 2 days (Ponce et al. 2010). In a study at Buffalo Ridge Wind Resource Area in Minnesota, rates ranged from 4.4 to 6.3 days before first scavenging event (Osborn et al. 2000). Strickland et al. (2001) found larger birds lasted on average 26.7 days while small birds lasted 23.4 days. Unlike this study, bird carcasses remained longer in the summer. Smallwood (2007) found scavenging rates greatly differ based on the state of the carcass, species, location of and size of carcass prior to placement.

Acknowledgements

Electrawinds Africa and Indian Ocean Islands (Pty) Ltd funded the project. We would like to thank Paul Lochner of the Council for Scientific and Industrial Research for his input to this study.

Notes

This version has been amended. Please see Erratum (http://dx.doi.org/10.1080/03014223.2012.758579).