Offshore wind developments - collision and displacement in petrels and shearwaters: literature review

Literature review of the risk of collision and displacement in petrels and shearwaters from offshore wind developments in Scotland.


5 Risks from collision, displacement and lighting attraction

Assessment of the risk of bird collisions at wind farms principally focuses on risks associated with a bird being struck by a rotating blade when passing through the rotor-swept area. The probability of collision, for a bird on a collision course with a turbine, depends on (i) the flight height of the bird, (ii) the likelihood of the bird altering its flight path to avoid the rotor swept area (i.e. avoidance), and (iii) if the bird passes through the rotor-swept area, whether it is struck by a rotating blade. Before considering these components in turn it should be noted that other collision risks may be associated with wind farms and their operations, such as collision with masts and aerials on the support vessels, or with moorings associated with floating wind platforms.

Whilst some components of the overall assessment of the collision risk posed by wind farms, and their population-level consequences, can be computed with estimable precision and accuracy, other components, such as the avoidance rate, or in the case of nocturnal procellariform seabirds, the attraction rate, are subject to considerably greater uncertainty, which render estimates of collision rate and population consequences highly speculative.

In this section we review the available published information to parameterise the collision risk models, and information which may assist the estimation of avoidance rates. Critical to the latter is the extent to which nocturnally active seabirds such as shearwaters and storm-petrels may be attracted to the illuminations required for turbines, support vessels and the construction or expansion of ports. We firstly consider factors other than illumination which may contribute to attraction of nocturnal Procellariiformes for offshore windfarms. In the final section, we explicitly consider the evidence for light attraction.

Flight height estimates presented below are obtained from aerial and vessel-based surveys, necessarily conducted under adequate weather and lighting conditions and usually including ship-following birds. These values may change under different weather and lighting conditions. Many sources providing assessments of the time a species spends at collision risk height do not specify the assumed turbine dimensions, and since turbine technology is rapidly evolving, collision risk levels may also change. Data on flight speeds have been obtained from tracking studies and refer to ground speeds, taking no account of non-linear flight paths and measured at the interval of the tracking device. They will therefore underestimate instantaneous flight speed to an unknown degree. Further, most tracking studies have been conducted on breeding adult birds, and parameter values may differ for immatures or juveniles or for different times of year.

5.1 Attraction of shearwaters and storm-petrels to offshore structures

A number of studies in Canada have found clear evidence that shearwaters and storm-petrels may be attracted to offshore structures such as drilling platforms, likely due to local prey enhancement as the structure acts as an artificial reef (Baird, 1990, Montevecchi, 2006, Burke et al., 2012). The foundations associated with offshore turbines may similarly act as artificial reefs, and cause changes in patterns of sediment transport and accumulation that could provide spawning grounds for benthic species. Whilst there is limited evidence for attraction of shearwaters and storm-petrels to oil and gas platform in the UK (Bourne, 1979, Sage, 1979), likely due to low densities of these species in the northern North Sea where seabird interactions with oil platforms have been studied, other authors report attraction of a variety of diurnal seabird species to oil platforms, likely as a result of local prey enhancement (Tasker et al., 1986). If fishery activity is reduced within windfarms, then local increases in fish density may result in these areas attracting seabirds, such as Manx Shearwaters, storm-petrels, and their avian predators such as large gulls and skuas. Aguado-Giménez et al. (2016) found that European Storm-petrels were attracted to fish farm cages 5 km from the coast during daylight, likely due to local prey enhancement. Procellariiform species are highly pelagic and are extremely unlikely to be attracted to offshore structures for the purposes of roosting, as is seen in species such as cormorants and shags (Dierschke et al., 2016).

5.2 Collision risk

5.2.1 Manx Shearwater

5.2.1.1 Flight style

Manx Shearwaters are classed as glide-flappers (Spear and Ainley, 1997b), using both flapping and gliding flight and engaging in slope-soaring behaviour (Thompson, 1987, Spivey et al., 2014). Gliding and soaring flight may increase with increasing wind speed (Gibb et al., 2017). Flight speed (see below), wing shape, relatively high wing loading, and tail shape (rounded, not forked) suggest that Manx Shearwaters have only moderate flight manoeuvrability (Warham, 1977, Furness and Wade, 2012).

5.2.1.2 Flight height

The species is generally considered to have low collision risk as it apparently spends limited time flying at rotor blade height (i.e. usually flies less than 20 m above sea level; Garthe and Hüppop, 2004, King et al., 2009, Cook et al., 2012, Furness and Wade, 2012, Furness et al., 2013, Bradbury et al., 2014, Certain et al., 2015). However, current flight height data for this species is based on aerial or vessel-based at-sea surveys, which can only take place during daylight and in relatively calm weather and may not be representative of the behaviour of Manx Shearwaters under all conditions. The species rarely uses level, flapping flight, but usually engages in slope-soaring, which leads to constant variation in flight height, although generally birds will remain low to the sea surface where the shear is strongest (Spivey et al., 2014). Flight heights may increase in stronger winds (Spear and Ainley, 1997b, Ainley et al., 2015) and modelling by Johnston and Cook (2016) indicated an increase in mean flight height between April and September.

Of 6,957 Manx Shearwater recorded during vessel-based surveys at 10 offshore wind farm sites, 0.04% (95% confidence interval <0.01–10.1%) were flying at heights that would put them within the rotor-swept zone (assumed to be 20–150 m above sea level), and models suggested their flight height distribution was unlikely to vary with distance to the coast (Cook et al., 2012). Models by Johnston and Cook (2016) estimated the proportion of flight time within the rotor-swept zone was 0.0 (95% confidence interval 0.0–0.0), based on boat survey data, and 0.0 (95% credible interval 0.0–0.02) based on digital aerial survey data.

5.2.1.3 Flight speed

Breeding Manx Shearwater GPS-tracked from Skomer, Wales, by Guilford et al. (2008) had a mean ground speed of 11.13 ± 9.55 m/s during flight. Behavioural models of GPS data for birds breeding on Skomer and Lighthouse Island, Northern Ireland, indicate median ground speeds of 8.9 m/s during direct or transiting flight and 2.01 m/s during foraging, when flight is more tortuous (Dean et al., 2013). Breeding Manx Shearwaters tracked from Great Blasket and High Island, Ireland in 2014 and 2015 had a mean ground speed across whole trips of 1.58 m/s (SD = ± 0.79 m/s, range 0.36–5.88 m/s), although ground speeds within trips would have shown greater variation (Wischnewski et al., 2019). Tracking from Lundy Island indicated mean ground speeds of 10.89 ± 3.31 m/s during flight, with clusters around 11 and 15 m/s in low wind speeds and greater variation in higher wind speeds, when birds were more likely to engage in soaring flight (Gibb et al., 2017).

Mean ground speeds differed between adults and immatures GPS-tracked from Skomer, with mean (± SE) speeds of 7.0 m/s ± 0.32 m/s for adults and 4.97 ± 0.25 m/s for immatures on short trips and 5.83 ± 0.17 m/s for adults and 5.14 ± 0.22 m/s for immatures on long trips (Fayet et al., 2015).

5.2.1.4 Temporal activity patterns

For breeding Manx Shearwaters tracked from Skomer and Lighthouse Island (Copeland) in July and August of 2009–2011, the percentage of time spent in different behaviours varied between breeding stages and colonies, with birds spending an average of 10% of their time in direct flight (i.e. transiting/commuting) and 63% foraging during incubation, and 15% in direct flight and 57% foraging during chick-rearing (Dean et al., 2013). Direct flight and foraging increased in the hour before sunrise, peaked just after sunrise and were lowest around midday when birds spent more time resting on the water (Dean et al., 2013). There was then a second peak in flight before sunset and a rapid decline at the onset of darkness. Foraging occurred almost entirely within daylight and twilight and birds roosted on the water in the evening and at night. Other GPS tracking studies from Skomer show similar activity patterns during incubation and chick-rearing (Guilford et al., 2008, Fayet et al., 2015). However, dietary analysis of Manx Shearwaters on Rum indicates that birds may have been foraging at night during the pre-laying period (Thompson, 1987).

5.2.1.5 Avoidance behaviour

Limited data are available on wind turbine avoidance behaviour of Manx Shearwaters given that there is little overlap between the species' distribution and currently operational wind farms, but Dierschke et al. (2016) preliminarily classified the species as weakly avoiding wind farms. Surveys of the Robin Rigg offshore wind farm in the Solway Firth detected a decline in the number of Manx Shearwaters in the area during construction and operation, compared with pre-construction (Canning et al., 2013b, Canning et al., 2013a), suggesting some macro-avoidance, but birds were observed close to turbines (Dierschke et al., 2016). An obvious gap in Manx Shearwater distribution was observed at North Hoyle wind farm in Liverpool Bay (Dierschke et al., 2016).

Flight speed, wing and tail morphology suggest that Manx Shearwaters may have limited manoeuvrability for micro-avoidance of turbine blades and associated structures. Flight agility is likely to be influenced by wind speed. Warham (1977) noted that in low winds shearwaters often come in fast and crash land at the colony but on windy evenings can stall and land lightly. In the context of collisions with turbine, shearwaters are likely to have lowered manoeuvrability under conditions when blades are turning more slowly. Adults, sub-adults and fledgling shearwaters of various species are known to collide with human-made structures on land, and this can sometimes result in high numbers of fatalities (Podolsky et al., 1998, e.g. Albores-Barajas et al., 2016), further indicating low levels of micro-avoidance.

5.2.2 European Storm-petrel

5.2.2.1 Flight style

European Storm-petrels fly with a combination of flapping and short glides, often moving in zig-zags and sometimes shearing in strong winds (Flood and Thomas, 2007). When feeding they hover or patter on the surface of the water, dipping to seize food items (Flood and Thomas, 2007). Smaller-bodied Procellariiformes have greater manoeuvrability in flight due to lower wing loading (Warham, 1977) and storm-petrels are highly manoeuvrable in snatching prey for the sea surface.

5.2.2.2 Flight height

Vessel-based observations suggest European Storm-petrels generally fly within 2 m of the sea surface, but occasionally up to 5 m (Flood and Thomas, 2007). They may fly lower in strong winds to shelter in wave troughs, as observed in the oceanitid and Oceanodroma storm-petrels (Ainley et al., 2015). Largely as a result of its low flight height, the European Storm-petrel is generally considered to be at low risk of collision (King et al., 2009, Cook et al., 2012, Furness and Wade, 2012, Furness et al., 2013, Bradbury et al., 2014, Certain et al., 2015), but data on flight heights for this species are limited. Observations of 52 European Storm-petrels on surveys of two offshore wind farm sites included a mean of 2% (range 0–2.5%) flying at heights that would put them at risk of collision with wind turbine blades (Cook et al., 2012).

5.2.2.3 Flight speed

European Storm-petrels tracked from Ireland had a mean trip speed of 4.05 (range 2.62–4.93) m/s and the maximum ground speed of any bird between two consecutive GPS locations was 11.18 m/s (Wilkinson, 2021). Mediterranean Storm-petrels tracked from Sardinia during incubation in 2020 had a mean speed of 4.0 ± 0.9 (range 2.1–5.2) m/s and a maximum speed of 9.8 ± 2.0 (6.7–12.5) m/s, while those tracked during chick-rearing in 2019 had a mean speed of 2.63 ± 0.9 (1.1–4.1) m/s and maximum speed of 7.38 ± 1.7 (4.5–9.8) m/s (De Pascalis et al., 2021). For Mediterranean storm-petrels tracked from Benidorm Island, the mean (± SD) speed was 4.18 ± 0.68 m/s (range 3.46–4.82 m/s) and the maximum travel speed was 10.17 ± 3.33 m/s (range 6.41–22.46 m/s) (Rotger et al., 2021). The mean speed for birds engaging in area-restricted search behaviour (i.e. foraging) was 2.03 ± 0.86 m/s (range 0.63–3.95 m/s) (Rotger et al., 2021).

5.2.2.4 Temporal activity patterns

European Storm-petrels depart from and return to the colony at night and while on foraging trips will forage both diurnally and nocturnally (D'Elbee and Hemery, 1997, Bolton, 2021). A two-state hidden Markov model for European Storm-petrels tracked from west Ireland assigned 60.6% of locations from High Island birds as foraging behaviour and 39.4% as transiting, while for Illauntannig foraging and transiting were assigned to 59.2% and 40.8% of locations, respectively (Wilkinson, 2021). Note that resting behaviour was not considered by Wilkinson (2021), but Mediterranean storm-petrels tracked from Benidorm Island spent a mean (± SD) of 35.23% ± 9.77 (range 19.00–54.00%) of the time resting on the water (Rotger et al., 2021).

5.2.2.5 Avoidance behaviour

We found no information in the literature regarding the extent of macro-, meso- or micro-scale avoidance by European Storm-petrels.

5.2.3 Leach's Storm-petrel

5.2.3.1 Flight style

Leach's Storm-petrel is classed as a glide-flapper, using a combination of flapping and long, shearing glides and hovering or pattering on the surface of the water to seize food items (Spear and Ainley, 1997b, Flood and Thomas, 2007). It has a very low wing loading (Warham, 1977) and its flight path can be irregular and unpredictable, with rapid changes of speed and direction, and becoming highly erratic in strong winds (Spear and Ainley, 1997b, Flood and Thomas, 2007).

5.2.3.2 Flight height

Vessel-based observations suggest that Leach's Storm-petrels generally stay within 5 m of the sea surface (Flood and Thomas, 2007) and they may fly lower during strong winds to shelter in wave troughs (Ainley et al., 2015). The species is usually assumed to have a low risk of collision, but data are limited and information for the European Storm-petrel is often used as a proxy (King et al., 2009, Langston, 2010, Furness and Wade, 2012, Furness et al., 2013, Bradbury et al., 2014).

5.2.3.3 Flight speed

Our literature search did not identify any estimates of flight speed for Leach's Storm-petrel, but Pollet et al. (2019) suggest it is relatively slow, similar to the 4 m/s given by Withers (1979) for Wilson's Storm-petrel.

5.2.3.4 Temporal activity patterns

Leach's Storm-petrels depart from and return to the colony at night (Ainslie and Atkinson, 1937) and are believed to forage both diurnally and nocturnally (Pitman and Ballance, 1990, Hedd and Montevecchi, 2006). More detailed information on their at-sea activity is lacking.

5.2.3.5 Avoidance behaviour

We found no information in the literature regarding the extent of macro-, meso- or micro-scale avoidance by Leach's Storm-petrels

5.2.4 Northern Fulmar

5.2.4.1 Flight style

The Norther Fulmar is a flap-glider, uses gliding flight extensively during foraging (Pennycuick, 1987) and increases gliding behaviour with higher wind speeds (Ainley et al., 2015). It has intermediate flight manoeuvrability (Warham, 1977, Garthe and Hüppop, 2004).

5.2.4.2 Flight height

Fulmars are generally considered to be at low risk of collision as they apparently spend limited time at collision risk height (Garthe and Hüppop, 2004, Cook et al., 2012, Fijn et al., 2012, Krijgsveld, 2014, Leopold and al., 2014, Harwood et al., 2018). Modelling based on 29,168 vessel-based observations estimates that the proportion of Fulmars flying at collision risk height (where the lower limit of the rotor-swept area is 20 m above sea level) is 0.002 (95% CI 0.000–0.061; Johnston et al., 2014). However, the species may fly higher in stronger winds (Spear and Ainley, 1997b, Ainley et al., 2015) and this behaviour is unlikely to be captured in vessel-based surveys, which are conducted only in relatively calm conditions.

5.2.4.3 Flight speed

The mean air speed of Fulmars measured off Foula, Shetland, using an ornithodolite was 13.0 m/s (Pennycuick, 1987). A male Fulmar GPS-tracked from Eynhallow, Orkney, during incubation, had an overall ground speed of 7.9 m/s and a maximum hourly ground speed of 17.6 m/s during its outward journey to the Mid-Atlantic Ridge, and an overall ground speed of 7.7 m/s and maximum hourly ground speed of 13.5 m/s on its return journey, although the return leg was largely against a headwind (Edwards et al., 2013). Hourly transit ground speeds were faster during the day (median = 9.4, range = 0.9–17.6 m/s) than at night (median = 4.6, range = 0.2–9.5 m/s), but hourly ground speeds during area restricted search (median = 1.2, range = 0.1–6.9 m/s) did not differ between day and night (Edwards et al., 2013). However, ground speeds from tracking data tend to be underestimates, and Weimerskirch et al. (2001) suggest the species regularly attains ground speeds of 19.4 m/s, aided by wind. Elliott and Gaston (2005) found that ground speeds of Fulmars in Nunavut, Canada, were lower during incubation (9.2 m/s) than chick-rearing (10.8 m/s) and, in contrast to Edwards et al. (2013), found that ground speeds were significantly lower for outgoing birds (8.8 m/s) than incoming birds (10.2 m/s).

5.2.4.4 Temporal activity patterns

The diurnal pattern of colony attendance by breeding Fulmars is very variable, suggesting different levels of nocturnal foraging at different breeding sites (Dott, 1975, Furness and Todd, 1984, Ojowski et al., 2001, Danielsen, 2011). Analysis of tracking data also suggests a combination of diurnal and nocturnal foraging (Edwards et al., 2013). Observations at sea near Shetland in the breeding seasons of 1992-94 found that Fulmars spent 81% of time resting or swimming and only 19% of time flying (Ojowski et al., 2001), but tracking by Edwards et al. (2013) suggested that foraging bouts involve short searching flights and only brief periods on the water, when prey is captured and consumed. Given the wide range of prey and varied foraging ecology of the Fulmar, it is likely that the time it spends in different behaviours is also highly variable.

5.2.4.5 Avoidance behaviour

Dierschke et al. (2016) classified Fulmars as weakly avoiding offshore wind farms, based on post-construction studies at 20 sites, but the authors note that data for this species are limited and it may actually display strong avoidance behaviour. It is possible that the lack of fishing vessels within wind farm areas makes them unattractive to Fulmars (Neumann et al., 2013, Braasch et al., 2015), but there is conflicting evidence regarding the influence of fishing vessels on Fulmar distributions (see section 4.5.4).

5.2.5 Sooty Shearwater

5.2.5.1 Flight style

Like Manx Shearwaters, Sooty Shearwaters are glide-flappers (Spear and Ainley, 1997b) with intermediate flight manoeuvrability (Warham, 1977, Garthe and Hüppop, 2004).

5.2.5.2 Flight height

Sooty Shearwaters are considered to have low collision risk as they generally fly very close to the sea surface and therefore below blade height (usually assumed to be 20–150 m above sea level), but this is based on very small sample sizes (Paton et al., 2010, Cook et al., 2012) and an assumption that Sooty and Manx Shearwaters fly at similar heights (Furness and Wade, 2012). Like Manx Shearwaters, Sooty Shearwaters may fly higher in stronger winds (Spear and Ainley, 1997b, Ainley et al., 2015).

5.2.5.3 Flight speed

Our literature search did not identify any estimates of flight speed specifically for Sooty Shearwater, but Spear and Ainley (1997b) estimated average ground speeds for diving shearwaters, a group which includes Sooty Shearwater, as 10.7 ± 2.3 m/s with a headwind, 14.0 ± 3.5 m/s with a tailwind, and 13.2 ± 4.6 m/s with a crosswind. Flying with a cross wind is by far the most common method used by Procellariiformes, including Sooty Shearwater (Spear and Ainley, 1997a).

5.2.5.4 Temporal activity patterns

While in the northern hemisphere, Sooty Shearwaters spend a large proportion of their time on the water and just 23.9 ± 15.2% of their time in flight, although this increases to 67 ± 24.1% once they begin their return migration to their breeding ground (Hedd et al., 2012, Bonnet-Lebrun et al., 2021). When on the water, they are resting, feeding, digesting (Bonnet-Lebrun et al., 2021) or moulting (Keijl, 2011). In July 2007, Keijl (2011) photographed 76 individuals in a flock gathered off Rockall, to the west of the Scottish mainland, 46% of which were in active primary moult. On their wintering grounds Sooty Shearwaters are particularly stationary at night, when they are on the water for 89% of the time (Hedd et al., 2012), although they are more active on nights with increased moonlight (Bonnet-Lebrun et al., 2021).

5.2.5.5 Avoidance behaviour

We found no information in the literature regarding the extent of macro-, meso- or micro-scale avoidance by Sooty Shearwaters.  

5.3 Displacement and barrier effects

There is a lack of empirical evidence relating to displacement, disturbance and barrier effects for these procellariiform seabirds, and therefore high levels of uncertainty regarding their vulnerability (Wade et al., 2016, Kelsey et al., 2018). These species are all generally considered to have a low vulnerability to displacement and disturbance from offshore wind farms and associated activities such as ship and helicopter traffic, and often rank lower than all other Scottish seabird species in terms of population impacts (Furness and Wade, 2012, Furness et al., 2013, Bradbury et al., 2014, MMO, 2018, Rogerson et al., 2021). They will associate with vessels at sea and display limited escape behaviour and short flight distances when approached by boats (Furness et al., 2013). However, there is some evidence of Manx Shearwaters and Fulmars avoiding offshore wind farm developments during the construction and operational phases (see sections 5.2.1.5 & 5.2.4.5; Dierschke et al., 2016), and the deficiency of data for the other species does not indicate a lack of impact. A higher level of disturbance may occur during the construction phase, when activity, noise and light levels may be greatest. The impacts of artificial light on nocturnally active species may also result in increased levels of displacement (see section 5.4).

Habitat specialisation is a key consideration when assessing vulnerability to displacement, with the negative impacts likely to be greater for specialists than generalists. Manx Shearwaters, European Storm-petrels and Leach's Storm-petrels cover large distances when foraging during the breeding season and appear to forage on a broad range of taxa (see section 4), which could suggest a lack of specialisation. However, all three species apparently travel long distances to target specific oceanographic features (see section 4; Scott et al., 2013, Dean et al., 2015, Hedd et al., 2018, Wilkinson, 2021), and displacement from these important foraging areas would likely have negative consequences. Displacement of Manx Shearwaters from key rafting sites may also result in population-level impacts, if displacement requires them to spend energy on flight, and thereby consume resources that would otherwise have been devoted to their chick on arrival at the colony.

Older chicks can be left unattended and unfed for several days at a time due to their accumulation of large lipid reserves (Ricklefs and Schew, 1994, Bolton, 1995b, Hamer et al., 1998), which could help to buffer them against a reduction in provisioning frequency due to increases in parental foraging trip durations caused by barrier effects. However, during the first week after hatching, chicks are unable to thermoregulate adequately and need to be brooded by adults. In this period they are particularly vulnerable to starvation and inclement weather as adults must divide their time between nest attendance to brood the chick and foraging at sea. Most breeding failures occur at this stage, as the energetic demands on adults, in relation to time available for foraging, are greatest (Bolton, 1995a). Increased energy demands on adults, for example due to displacement from profitable feeding areas, or increased flight paths due to barrier effects, would likely lower chick survival rates.

Several studies have shown that the flight paths of petrels and shearwaters are orientated to maximise the energetic benefits of crosswinds (Spear and Ainley, 1997b), often resulting in circular (rather than direct "out and back") patterns to foraging trips (Ventura et al., 2020). Displacement and barrier effects may prevent the optimisation of foraging tracks to maximise the energetic benefits of cross winds. During the pre-laying exodus, female Manx Shearwaters undertake long foraging trips to oceanic waters (Dean, 2012) to acquire the nutrients required for egg formation. The single large egg represents a considerable resource investment, and the inward flight to the colony for egg laying is likely to be energetically expensive, at a critical time in the breeding cycle. Increased flight costs imposed by barrier or displacement effects during this period may have particularly high costs on breeding success. The foraging behaviour of female storm-petrels during the pre-laying period is unknown, but since they lay one of the largest eggs in relation to body size of any bird (approximately 30% of female body weight; Davis, 1957a), increased foraging costs imposed by displacement or barrier effects are likely to be particularly severe.

5.4 Lighting attraction and disorientation

The nocturnal attraction of birds to light, often with fatal consequences, has been known for several centuries. Early settlers of the Azores archipelago lit bonfires on the cliffs at night to attract seabirds, which they beat from the air with sticks, to be used as feed for their pigs (Fructuoso, 1561). Studies in the USA estimated that in the mid-1960s more than one million nocturnal migrant birds died annually by collision with illuminated communications towers (Gauthreaux and Belser, 2006) and observers in both the UK and Canada have reported that hundreds, or even thousands, of seabirds, predominantly species of storm-petrel, are killed by attraction to the gas flares of hydrocarbon platforms (Sage, 1979, Wiese et al., 2001, Baillie et al., 2005, Montevecchi, 2006, Burke et al., 2012), although Bourne (1979) disputed the identification of birds killed at platforms in the North Sea. Tasker et al. (1986) did not report shearwaters and storm-petrels attracted to platforms in the central North Sea, but observations were from a region and season associated with low densities of these species (Waggitt et al., 2020). Attraction distances and the possible influence of light position relative to flight paths are considered further in following sections.

There have been several reviews of the attraction of seabirds to artificial light (Montevecchi, 2006, Laguna et al., 2014, Rodríguez et al., 2017) and we do not repeat that information here. Rather, we review the literature in the specific context of the issues surrounding the assessment of the impacts on seabirds of wind farm development and operation. Wind farms are required to be illuminated in accordance with marine navigation regulations (DECC, 2011, IALA, 2013, MCA, 2021) and the Air Navigation Order (CAA, 2016). In addition, a large programme of port expansion is underway in Scotland to support the construction and maintenance of new offshore wind farms, and this will result in increased illumination in coastal areas. The central issue is the extent to which illumination of wind farm structures, associated infrastructure (such as wet storage), construction activities and the vessels and ports associated with wind farm operations will: (i) attract seabirds, and (ii) modify seabird behaviour in their proximity.

The literature on light attraction in birds does not always make a clear distinction between: (i) attraction per se (i.e. "phototaxis"), which could potentially operate over ranges of tens of km, and (ii) the alteration of flight paths of birds when in close proximity (i.e. within tens of metres) of illuminated structures (i.e. "disorientation"). Long-range light attraction may result in birds being displaced from foraging areas and activities. Light-induced disorientation may cause birds to circle light sources for many hours (Gauthreaux and Belser, 2006) with obvious implications for collision risk. Existing models of collision risk assessment (Band et al., 2007, Band, 2012, Masden, 2015, McGregor et al., 2018) do not explicitly model the scenario of birds circling a turbine, but rather consider a straight flight path only. In the context of the assessment of impacts of wind turbines on seabirds, it is helpful to make a clear distinction between these two effects (attraction and disorientation), and the spatial scales at which they operate. The first will affect the number of birds brought into the vicinity of wind turbines and associated structures, vessels and shore facilities ("macro" and "meso" scales sensu Cook et al., 2018), and the second will affect the length of time birds remain within the proximity of potential collision surfaces and the number of occasions an individual bird may pass through the rotor-swept area ("micro" scale sensu Cook et al., 2018). These two effects of artificial light may have different drivers, and impact juveniles and adults differently, as discussed below. We do not consider light attraction to be a separate impact pathway, but it may exacerbate one or more of the recognised impact pathways (e.g. collision, displacement).

5.4.1 Evidence for light-induced disorientation

There is abundant evidence of light-induced disorientation for a wide range of avian groups, including shearwaters and petrels. Such evidence includes: the grounding of fledgling Manx Shearwaters, Leach's and European Storm-petrels in lit areas of the village on Hirta, St Kilda (Miles et al., 2010); collision of Manx Shearwaters with lighthouses and other illuminated structures (Archer et al., 2015, Guilford et al., 2019); grounding of European Storm-petrels onto rocks lit by researchers' head torches (Albores‐Barajas et al., 2011); grounding of European and Leach's Storm-petrels on hydrocarbon platforms (Sage, 1979, Wiese et al., 2001, Baillie et al., 2005, Montevecchi, 2006, Burke et al., 2012, Gjerdrum et al., 2021), and the grounding of Leach's Storm-petrels on vessels (Wynn, 2005, Wakefield, 2018, Wilhelm et al., 2021) and industrial developments (Wilhelm et al., 2021).

While the distance from which birds have been attracted to such light sources is usually unknown, observers report that, once attracted to the vicinity (i.e. within several tens of metres) of a powerful light source, birds seem unable to escape and become vulnerable to collision. Rodríguez et al. (2022) showed formally that flight tortuosity of fledgling Cory's Shearwaters Calonectris borealis heading from inland breeding sites to the sea increases with the level of light radiance over which they are flying. Tracks of tagged individuals reveal that they remain in flight within the lit areas for several hours before grounding.

Many studies describe procellariform seabirds being drawn downwards towards bright light shining from below (e.g. Rodríguez et al., 2015a, Rodríguez et al., 2017, Rodríguez et al., 2022). In such cases the birds' natural flight height is lowered by light attraction/disorientation. It is not clear to what extent light attraction/disorientation may result in birds that are flying close to the sea (below rotor swept height), being drawn upwards to heights within the rotor swept area, although this is likely to be the case for storm-petrels stranded on oil platforms. The impact of light attraction on flight height must be considered.

Seabird species that rear their young underground seem particularly, if not exclusively, sensitive to light-induced attraction/disorientation. In the case of fledglings this is perhaps because young fledge with somewhat under-developed visual acuity due to a lack of visual stimulation in the darkness of the nest chamber (Atchoi et al., 2020). It is notable that measurements from eyes of two Manx Shearwaters captured on the point of fledging indicated that their optical structure was slightly myopic (i.e. would not produce a focussed image on the retina; Martin and Brooke, 1991). Hence, the young of burrow-nesting shearwaters, storm-petrels, and puffins appear particularly vulnerable to grounding in well-lit areas on their fledging flights from the colony (Atchoi et al., 2020), whereas the young of closely related surface-nesting species, such as Fulmars, are not vulnerable to light-induced grounding. While numerous studies have shown that light-induced grounding is much more prevalent among recently-fledged juveniles, the timing of some grounding events of Leach's Storm-petrels on offshore oil platforms (in April–August before any young of the year have fledged; Gjerdrum et al., 2021), and the stranding of likely breeding Leach's Storm-petrels on a docked seismic vessel (Wilhelm et al., 2021), show that adults may be light-attracted on occasion too. Collins et al. (2022) found no impact of oil platforms on the behaviour of breeding Leach's Storm-petrels GPS-tracked in Newfoundland, but only 1.1% of trips involved exposure to oil platforms at night and around 30% of tracked birds were not recaptured, so their fate is unknown. While juveniles are clearly more susceptible than adults to light-induced grounding, it is not clear for how long post-fledging such susceptibility persists, and whether birds grounded weeks or months after fledging were forced to land by severe weather (e.g. Teixeira, 1987) rather than light attraction.

Petrels and shearwaters are more likely to be disorientated by artificial light under conditions of low ambient light (i.e. a new moon), and during conditions of fog, mist or light rain. Guilford et al. (2019) showed experimentally that, during foggy conditions but not clear nights, light emanating from windows resulted in disorientation of adult Manx Shearwaters, causing them to collide with the building. They suggested that when the birds were suddenly close to a relatively bright light, the light-scatter caused by fog compromised the birds' dark-adapted visual guidance. Alternatively, they suggested that Manx Shearwaters may use a light-dependent magneto-receptor, located in the eyes, for navigation (Hore and Mouritsen, 2016), which could become temporarily disrupted by saturation in the presence of bright light. However, several experimental studies have failed to find evidence for the existence of such a magnetic compass in either adult (Padget, 2017) or fledgling (Syposz, 2020) shearwaters, and the sensory basis of navigation in Procellariiformes remains unclear.

Experimental reduction in artificial lighting (outside lights turned off and the majority of windows shielded with blinds) in the village of Hirta, St Kilda resulted in fewer grounded fledgling Leach's Storm-petrels, but the number of grounded Manx Shearwaters remained high (Miles et al., 2010). The authors concluded that Manx Shearwaters may be more vulnerable than storm-petrels to disorientation, or that they also navigate towards low-frequency sounds, since many grounded individuals were located close to generators or extractor fans, one being found impaled in the outlet duct of an extractor. Potential attraction of shearwaters to low-frequency noise, and implications for attraction to wind turbines and associated structures and vessels, requires further consideration.

5.4.2 Evidence for light attraction

While there is clear evidence for the disorientation of burrow-nesting Procellariiformes by artificial light sources, the extent of long-range attraction is more difficult to quantify. There are reports of European Storm-petrels being attracted to garden fireworks and moth traps (Miles et al., 2010), which they are unlikely to have been overflying, and suggests they were attracted by the artificial illumination. The number of individuals recovered in campaigns to rescue grounded fledglings are typically very low in relation to the local population size (e.g. Miles et al., 2010, but see Le Corre et al., 2002, Rodríguez et al., 2015b, Rodríguez et al., 2022), suggesting that birds are not attracted over large distances, or if so, only a small proportion of individuals are affected, or recovered. For example, the number of fledgling Manx Shearwaters recovered in the town of Mallaig, Scotland (Syposz et al., 2018), broadly corresponds, given the size and distance of the colony that is the likely source of the majority of individuals (Rum, 27 km away), with the number predicted if birds disperse randomly in all directions and the small proportion that orientate towards Mallaig are then attracted from very short range.

Two cases where large numbers of fledglings, representing large proportions (up to 40%) of the local population, are encountered grounded in brightly illuminated urban areas are Barau's Petrels Pterodroma baraui on Reunion Island, Indian Ocean (Le Corre et al., 2002) and Cory's Shearwaters on Tenerife (Rodríguez et al., 2015b, Rodríguez et al., 2022). The grounding of large proportions of the cohort of fledglings may imply that birds are attracted from large distances. In both cases, nesting sites are mainly located in high altitude areas in the island interior, and fledglings fly over brightly lit coastal areas (some more than 10km distant from the nearest colonies) to reach the sea. When flying over these areas birds become vulnerable to disorientation from powerful light sources below them. The sensitivity of birds to disorientation when overflying powerful light sources projected upwards is evidenced from the disorientation of very large numbers of nocturnal migrants by ceilometers (bright lights shone vertically to measure the height of the cloud base; Rich and Longcore, 2006), and the effectiveness of spotlights directed upwards to ground and capture storm-petrels returning to the colony at night (Ishmar et al., 2015). Whilst the minimum distance between colony locations and some grounding sites is more than 10km in these studies, the distance from which birds are attracted by light may be considerably less. Since birds may overfly these coastal areas on route to the sea, the high disorientation sensitivity of procellariiform seabirds to light sources from below could potentially account for the high rate of grounding in the case of Reunion Island and Tenerife, without birds being attracted from large range.

Several recent, and highly innovative, studies have started to assess the behaviour of fledgling Procellariifomes in response to artificial light. The first (Troy et al., 2013) modelled the numbers of Newell's Shearwaters Puffinus newelli recovered in different sectors of Kauai Island, Hawaii, in relation to location and size of colonies, light radiance levels across the island, and models of fledgling movement. They concluded that the observed spatial pattern of groundings indicated that fledglings were attracted back to the island by coastal illumination after they had reached the sea, and from distances of up to 10 km from the coastline. These modelled estimates of attraction range receive empirical support from two studies (Rodríguez et al., 2015b, Rodríguez et al., 2022) that tracked fledgling Cory's Shearwaters as they overflew brightly lit coastal areas in Tenerife on their flights to the sea. Both studies were conducted over multiple years, and each found that c. 14% of fledglings were later recovered grounded. Although neither study attempted to estimate the distance from which fledglings may become attracted towards artificial light, inspection of the tracks suggests that abrupt course deviations towards lit areas could occur from a range of several kilometres. All birds recovered by Rodríguez et al. (2015b) were grounded within 16 km of their breeding colonies, and 50% were found within 3 km of their nest site. Once above brightly lit areas many birds showed highly tortuous flight paths, circling to remain within the lit areas, before descending to ground level, as illustrated here. On multiple occasions birds that had reached the sea, and were up to 2.5 km from land, returned to brightly lit areas on the coast.

On St Kilda, considerable numbers of Leach's and European Storm-petrels breed within 2 km and in direct line of sight of the village illuminations, but the number of grounded fledglings is very small in relation to the size of the breeding populations, representing <<1% of the number of young likely to fledge annually (Miles et al., 2010). If the number of fledglings encountered grounded is an accurate reflection of the numbers attracted and disorientated, these findings suggest that fledglings are not susceptible to attraction to these light sources from long range, although the level of illumination in the village was relatively low (32 outside lights and 11 buildings with indoor lighting; Miles et al., 2010). In contrast, the vast majority of Manx Shearwaters breeding on St Kilda do not fledge in sight of the village and would not pass within sight on a direct route to the sea, raising the likelihood that they are attracted to illumination after having reached the sea, and may be attracted from a considerable range (>2 km) to illuminated areas. Similar differences in the numbers of storm-petrels and shearwaters encountered grounded in Hawaii and the Canary Islands have led other authors to suggest that the larger species of Procellariiformes may be more vulnerable to light attraction (Telfer et al., 1987, Rodríguez and Rodríguez, 2009). Any such conclusions may be premature however, since the smaller size and largely dark plumage of storm-petrels may result in lower detection rates during searches for grounded birds and storm-petrels may be able to take flight after grounding in enclosed situations more readily than shearwaters, which require an open space in which to take a "run up" to become airborne. Due to their smaller size, storm-petrels are also more likely to be depredated (e.g. by cats and dogs) and removed (Wilhelm et al., 2021). The susceptibility of storm-petrels to light-attraction and disorientation may be higher than implied by the numbers of individuals encountered grounded.

5.4.3 Attraction to vessels

In addition to attraction to or disorientation by lights at ports and on turbines, the potential for interaction of Procellariiformes with wind farm service vessels should also be considered. There are many anecdotal accounts of nocturnal seabirds, especially storm-petrel species, alighting on ships at night. For example, Wakefield (2018) reports that on several occasions during a research cruise by RRS Discovery to the mid-Atlantic, Leach's Storm-petrels were found on the ship's decks at night and caught by hand. These groundings usually occurred in misty conditions and were likely caused by birds being attracted to or disorientated by the deck's flood lights. Of 1,823 seabirds (all burrow-nesting Procellariiformes) recorded on board rock lobster fishing vessels around the Tristan da Cunha archipelago and Gough Island between 2013 and 2021, 4% died after being attracted to/disorientated by artificial lights (Ryan et al., 2021). As discussed above, it is not clear to what extent the grounding of storm-petrels on vessels results from macro- or meso-scale light attraction, or whether they are attracted to vessels by other cues (such as olfaction, low frequency sounds, or visual cues associated with a food source). Storm-petrels are known to follow a wide range of vessels, probably in search of food brought to the surface by the wake or vessel lighting or, in the case of fishing vessels, for offal. They can also be attracted to stationary vessels if any oily waste is released. In calm conditions European Storm-petrels may be attracted from distances of >1 km (M. Bolton pers. obs.) and may aggregate in large numbers. In the context of use of vessels for service operations for wind turbines, nocturnally active Procellariiformes (especially storm-petrels) are sensitive to attraction (by phototaxis, olfaction, or visual cues associated with food sources), and may subsequently become disorientated, either by lighting associated with the vessel, or navigation lights on nearby turbines.

5.4.4 Implications of the capabilities and sensitivities of the visual system of petrels and shearwaters for light disorientation/attraction

Petrels and shearwaters have been a particular focus for studies of avian vision for many decades (Lockie, 1952, Hayes and Brooke, 1990, Martin and Brooke, 1991) due to the species' need for visual capabilities to fly and forage under a wide range of light intensities, and in air and water, where the refractive properties of light differ. As a result, a considerable amount of detailed information exists on the microscopic and optical structure, and the visual fields, of the eyes of Manx Shearwaters, Fulmars and storm-petrels (Mitkus et al., 2016), which can inform our understanding of their behaviour in the vicinity of lit structures at sea (Atchoi et al., 2020). In brief, the retinas of Manx Shearwater, Fulmar and Leach's Storm-petrel all possess a central region (variously termed "Area centralis" (Lockie, 1952), "horizontal strip" (Hayes and Brooke, 1990), "visual streak" (Mitkus et al., 2016)), which receives light input from the horizon when the bird's head is normally orientated. The central part of this region is equipped entirely with cones—photoreceptors that operate under high light intensities (i.e. daylight) that are capable of colour vision and are responsible for high spatial acuity. This horizontal central structure is found in a range of seabirds and other species inhabiting open landscapes and provides high acuity to detect objects at, or close to, the horizon in well-lit environments. The outer margin of the central horizontal strip is equipped with rods—photoreceptors that operate under low light conditions—which have low spatial acuity. The density of rods increases from the central strip to the periphery of the retina (Lockie, 1952). The density of rods in the peripheral retina (which receives light from above and below the horizon when the head is normally orientated) is two-fold greater in the Manx Shearwater than Fulmar, and four times greater than the House Sparrow Passer domesticus, which is not active at night.

Martin and Brooke (1991) measured the visual field of the eyes of the Manx Shearwater and found that the eyes are directed slightly forwards and downwards when the head is normally orientated, with a blind spot above and behind the crown. In normal flight the eyes will therefore receive greater light input from in front and below the bird than from above and behind. During daylight, when the pupil is contracted to restrict the amount of light entering the eye, light falls on the centre of the retina, and objects on or close to the horizon are rendered with high spatial acuity, while objects further from the horizon are rendered with lower acuity. In low light levels at night the pupil opens to allow more light to enter and this is detected by the high density of rods located towards the periphery of the retina. Thus, the optic system of shearwaters and petrels provides high acuity for objects close to the horizon during daylight, and high sensitivity (though low acuity) to low light levels at night. Disorientation of shearwater fledglings when overflying brightly lit areas may result from saturation of the visual pigments of the rods (Verheijen, 1985), which cannot be adequately rectified by contraction of the pupil to limit entry of light to the eye. Birds are in effect blinded and can no longer see visual details that they could detect when dark-adapted. Alternatively, bright light may cause contraction of the pupil, so little light falls on the peripheral rods, and the birds are unable to discern poorly lit objects beyond the brightly lit areas, and so circle to remain within the illuminated field.

5.4.5 Influence of light wavelength on visual perception of shearwaters and storm-petrels

Manx Shearwaters are known to forage at depths of up 55 m (Shoji et al., 2016). Since light of shorter wavelengths (blue) penetrates water to greater depths than that of longer wavelengths (red), to maximise acuity when foraging at depth it is likely that the cones of Manx Shearwaters have greater sensitivity to blue than red light. Since storm-petrels dive to a very limited degree (max 5 m; Albores‐Barajas et al., 2011), they have less need for enhanced sensitivity to blue light.

Experiments to examine the response of adult Manx Shearwaters in flight over the colony to different intensities and wavelengths of light showed that birds were more responsive to (avoided) bright white than dim white light and showed greater avoidance of blue and green light than red light (Syposz et al., 2021a). There was no difference in the birds' behaviour when exposed to red light compared to no light. These results indicate that Manx Shearwaters have greater sensitivity to light of shorter wavelengths (blue and green) than long (red).

These findings appear to contrast with a number of largely observational (not experimental) studies that have examined the effect of light wavelength and pattern of illumination (constant vs flashing) on the collision rate of nocturnal migrants (principally passerines) with communication masts and onshore wind turbines in North America. These studies have compared the flight paths and/or number of birds found dead under structures with different types of illumination and may suffer from uncontrolled bias. However, they broadly indicate that flashing red lights causes less attraction and collisions than steady constant red light (Gehring et al., 2009, Kerlinger et al., 2010), and whilst constant red light caused greater attraction than flashing white light (Gauthreaux and Belser, 2006), Gehring et al. (2009) found no difference in the number of collisions at masts with flashing red or flashing white light. It has been suggested that red light may interfere with magnetoreception in migrating passerine birds: three passerine species showed normal orientation under dim monochromatic light from the blue-green range of the spectrum, while they were disoriented under yellow and red light (Wiltschko and Wiltschko, 2002). Gauthreaux and Belser (2006) recommend the use of flashing white lights in place of steady red lights to reduce the risk of collision of nocturnal (mainly passerine) migrants with communication masts in USA.

Several studies have failed to find evidence of magneto-reception in shearwaters (Padget, 2017, Syposz et al., 2021b) and it is possible that differences in the sensory systems used for navigation in nocturnal Procellariiformes and passerines may result in important differences in their sensitivities to attraction/disorientation by light of particular wavelengths. Several hundred million migrant birds cross the North Sea annually, at risk of collision with wind turbines (Hüppop et al., 2006), and the benefits of a particular lighting regime to reduce collisions of nocturnal Procellariiformes, such as the use of red navigation lights, must be weighed against likely impacts on other species.

5.4.6 Non-collision consequences of light attraction of seabirds that may affect their survival and productivity

If light-induced disorientation leads to individual birds circling the navigation lights on the nacelle or tower of turbines for protracted periods (as has been reported for birds disorientated by lighthouses or gas flares) the probability of collision with turbine blades or other surfaces is vastly increased, and may approach unity. However, individuals that are attracted to and disorientated by light associated with wind farms may become vulnerable to other lethal and sub-lethal impacts. If wind farms provide roosting opportunities for large gulls, or other predatory species (skuas, falcons), storm-petrels and Manx Shearwaters are likely to be vulnerable to predation (Hey at al. 2020), particularly if wind farm illuminations provide sufficient ambient light for effective hunting by these predators (Watanuki, 1986). Sub-lethal affects that may influence survival in the longer term, or the ability to rear young, could accrue from the wasteful expenditure of energy in circling flight for protracted periods. This may lead to loss of body condition resulting in birds becoming more vulnerable to starvation or predation. Flight costs of European Storm-petrels have been estimated at 3.9 times basal metabolic rate (Bolton, 1995a), close to the maximum sustainable work rate (Drent and Daan, 1980). Prolonged periods of flight, without opportunity to feed or rest, may lead to dehydration or exhaustion of birds that escape collision. Conversely, many fisheries use artificial light to attract prey and there is a possibility that birds could benefit from increased foraging opportunities if artificial lighting around wind farm developments increases prey availability by attracting it close to the sea surface. The evidence base around Procellariiformes exploiting prey resources concentrated near the surface by artificial light is limited, but European Storm-petrels have been observed foraging around illuminated fish farms at night in the Faroe Islands (B. Porter, pers. comm.).

5.5 Options for mitigation

The second of the two expert workshops held as part of this project focussed on mitigation options to reduce the impacts on Procellariiformes of offshore wind farm developments and associated activities and infrastructure. Table 4 summarises the mitigation options discussed at the workshop and in the published literature. Full reports of both workshops are provided in Appendix 1. NatureScot (2020) have suggested several potential mitigation options for reducing the impacts on birds of lights placed on wind farms for the purposes of aviation safety. These mitigation options do not relate specifically to offshore wind farms or Procellariiformes but we include them in Table 4.

Table 4. Suggested mitigation options for reducing the impacts on Procellariiformes of offshore wind farm developments and associated activities and infrastructure, collated from the published literature and discussions during the two expert workshops held as part of this project.

Option

Evidence base

Comments

Technical/legislative feasibility

Alter pattern of illumination (flashing rather than steady lights)

Good evidence from numerous studies in USA that flashing lights cause less attraction/collision of migrant nocturnal passerines.

Not systematically tested for Procellariiformes.

Bardsey lighthouse changed to a red flashing light in 2014 and this resulted in a huge reduction in collisions of Manx Shearwaters.

Need consistency in lighting across wind farms to avoid confusion to mariners and to comply with international standards, which precludes modification.

Even apparently simple changes in lighting require intervention at early stage of turbine design/construction

Alter wavelength of lights

Studies conducted primarily on passerines provide little empirical evidence that white light causes less attraction/collision than red light (white light contains red). Green may be much better than white.

Experiments conducted on Manx Shearwater showed greater avoidance of white, blue and green than of red light.

Not clear what the attraction properties of red vs white light are for Procellariiformes.

Most vertebrate rods are maximally sensitive to green wavelengths and whether particular species are attracted to or repelled by green light would require specific behavioural studies. Green light should only be used if it is highly directed.

Need consistency in lighting across wind farms to avoid confusion to mariners and to comply with international standards, which precludes modification.

Even apparently simple changes in lighting require intervention at early stage of turbine design/construction

Search and rescue (SAR) lights need to be red to avoid reducing the night vision of crew.

Directional intensity / shielding of lights

Some suggestion in the literature that birds are most sensitive to attraction of light from below. Fitting of shields to prevent upwards light radiation at a coastal resort in Hawaii reduced the number of grounded Newell's shearwaters by 40% over 2 seasons (Reed et al., 1985)

Birds may also be attracted upwards towards light, as is likely the case for storm-petrels stranded on offshore oil and gas platforms, which tend to be several tens of metres above the sea surface.

Already set out in ICAO requirements and EASA CS-ADR-DSN Chapter

Q. This focusses the 2000 cd lighting in the horizontal plane

and reduces the intensity of the light from above and below. Both regulations stipulate minimum requirements as well as additional recommended vertical angles, which cannot be ignored without justification. Most lights will incorporate this as standard.

Marine lighting is also focused on the horizontal plane but needs to remain visible to all sizes of vessels both close to turbines and at the extreme range of the light.

Reduce intensity of lights

The effectiveness for reducing bird collisions is unknown, but likely to reduce the range from which any "attraction" might occur.

Not enough evidence on the impact this would have on different seabird species.

Intensity more important than colour in bird night vision.

Impact of different intensities depends on atmospheric conditions. Any conditions creating large, diffuse pools of light likely to be a problem.

Already set out in CAA guidance CAP 764. Lights can be dimmed to 200 cd in good visibility (greater than 5km). 200 cd lights can still be visible to the human eye > 20 km in good visibility conditions.

Reduce number of turbines illuminated

Dependent on the range at which any "attraction" of birds to light might occur, the reduction in the number of turbines illuminated is likely to reduce the number of individual birds brought into the proximity of turbines

If the number of turbines lit is reduced, the intensity of lighting may have to increase to compensate.

Reduce or cover lighting associated with maintenance vessels and associated activities and infrastructure (e.g. ports, wet storage)

Reduction of vessel lighting and the use of blinds has successfully reduced the number of collisions of burrow-nesting Procellariiformes with fishing boats (Ryan et al., 2021).

Blinds for vessels should be easy to implement, but changes to safety lighting are likely to be more difficult.

No lighting, or turning off lighting at key times (e.g. fledging period)

There is good evidence for light-induced disorientation (i.e. circling) of Procellariiformes (especially storm-petrels), so elimination of lighting is likely to reduce the number of occasions an individual passes through the rotor-swept area, on a flight past a turbine.

Lack of lighting may result in collisions by birds that cannot see the turbines on nights with particularly low ambient light.

Not possible for offshore wind farms due to safety concerns. Should not be considered as a mitigation option.

Radar-activated lighting

Reduction in collisions will depend on the proportion of time turbines are left unilluminated, during periods when light-induced collisions would otherwise occur.

CAA support this in principle and are considering the parameters in

detail. In the meantime, CAA are happy to discuss the approach on a case-by-case basis. In use in other countries, to differing extents, but it is acknowledged that the costs are high.

Detection systems are not currently possible for all marine vessels, especially ill-equipped recreational vessels, and lighting provision must cater for all users.

Additional lighting to guide birds away from wind farms

Currently unclear whether this would be effective. May result in further attraction / disorientation / displacement of target birds.

Would need to consider wider impacts on species other than Procellariiformes.

Additional lighting may be more feasible than reduced lighting.

Shut down turbines during meteorological conditions likely to result in high collision rate

Collision risk is reduced if turbines are not rotating.

Since conditions that generate high collision rate are usually associated with lower wind speed, little economic impact on electricity generation?

Unlikely to be acceptable given the importance of offshore wind for future UK energy production.

Increase minimum blade height

May help to reduce collisions at times/in conditions when birds are flying higher (e.g. Manx Shearwaters fly higher in stronger winds).

Has benefits outside of mitigation for birds.

Requires feasibility assessment on a case-by-case basis.

Increase detectability by marking blades / towers

Maximising visibility of blades is likely to reduce the number of collisions as birds would be better able to avoid them.

There needs to be consistency across wind farms to avoid confusion to mariners and to comply with international standards.

Deter birds (seabirds and / or avian predators) using sound

Currently unclear whether deterrence using sound would be effective.

If birds could be deterred by sounds outside of human hearing range this would avoid interference with regulation sounds used for maritime safety.

There needs to be consistency across wind farms in their use of fog horns.

Train crew in safe handling / release of stranded birds

Would not prevent collisions but may reduce mortality of grounded / stranded birds.

Posters at harbours in Pembrokeshire, Wales, provide guidance for mariners in case of Manx Shearwaters stranding on their vessels. Similar schemes have been implemented in other countries for other seabird species.

Has been done elsewhere and could be relatively cheap to implement. Could be built into relevant consenting conditions.

Contact

Email: ScotMER@gov.scot

Back to top