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

7 Examination of challenges and recommendations for filling data gaps to assess the impacts of offshore windfarms

The suggestions below are based on the literature review and workshop discussions with respect to evidence gaps and approaches to filling these for Manx Shearwater, European Storm-petrel and Leach's Storm-petrel. Note OWSMRF is currently conducting a detailed review of the knowledge gaps and research recommendations relating to parameters required for PVA for Manx Shearwater and European Storm-petrel.

7.1 Detectability and diel variation in marine distributions

There is an important need for experimental validation of potential biases in aerial survey methods, including detectability, identification and diel variation. Detectability could be tested by carrying out targeted digital aerial surveys (DAS) or vessel-based surveys with an experimental approach. For example, decoy models of birds could be used to assess detectability under different conditions, but only for birds drifting on the sea surface. Surveys could also be performed alongside large scale high-resolution tracking of birds, or decoys, but achieving large enough sample sizes of tagged birds is likely to be difficult. Radar or thermal imaging could be used to conduct nocturnal surveys of leased areas, but identification to species level may not be possible. Evidence needs with respect to detectability are greatest for the two storm-petrel species and for diel activity the highest priority is Leach's Storm-petrel (Table 6).

7.2 Basic morphometric data

Body length and wingspan measurements would be relatively easily collected by ringers or fieldworkers working with Procellariiformes in Scotland. However, while these data could readily be collected, the evidence need is low for all three species (Table 6).

7.3 Flight data

The evidence needs for flight speeds and heights are high for all three species, with the exception of medium scoring for Manx Shearwater flight height (Table 6). Estimates of flight parameters such as speed and height can be gained from tracking data, but acquiring accurate estimates is difficult, even with high resolution data. Where possible, "instantaneous" flight speeds from GPS tags, based on Doppler-shift information derived from the movement of the tag relative to the movement of the satellites (Safi et al., 2013), will be more accurate than that derived from distance covered between successive fixes. Tags providing high resolution tracking data are available for Manx Shearwaters but the accuracy of flight height data from high resolution GPS tracking of this species is still low. Small (< 1.5 g) barometric pressure loggers can be used to estimate flight heights when deployed alongside GPS devices , but because of the need to calibrate to local environmental pressure, accuracy may sometimes be low. Since tags deployed on storm-petrels must be much smaller than those used on Manx Shearwaters, the limitations on battery life and data storage capacity mean that GPS data collected for storm-petrel foraging trips tends to be of lower resolution. However, it would be possible to collect higher resolution data for short periods of storm-petrel foraging trips. Tracking of fledglings is challenging due to the difficulty of retrieving tags for data download.

Flight height can be measured with radar, although these measurements have biases. Distinguishing between similar species (e.g. Manx vs Balearic Shearwater) is difficult, but mobile radar units could be deployed in areas where only a single species is expected (e.g. Rum for Manx Shearwaters). It may be possible to estimate flight heights from vessel-, or turbine-mounted cameras or from aerial or thermal imagery, but again, accuracy is likely to be low. Flight height can also be accurately measured using laser rangefinders (Largey et al., 2021).

It is important to note that flight within wind farm developments may differ from that elsewhere so work within wind farms is important, but currently limited for these species by the lack of overlap between their marine distributions and operational wind farms.

7.4 Avoidance/attraction behaviour

Assessment of macro-avoidance of windfarm development is best achieved by comparing marine distributions of seabird pre- and -post construction. In light of the limited tracking of the three focal species to date in Scotland, we recommend further tracking studies from key colonies to better understand the pre-construction movements and distribution of these species. Such tracking studies should continue as construction occurs and after it is completed, to inform understanding of meso- and micro-avoidance behaviour.

Currently there is little known overlap between operational offshore wind farms and shearwater and storm-petrel marine distributions, so there is limited scope for collecting data on the species' micro, meso and macro avoidance behaviour within and around wind farms, although the evidence need is high (Table 6). There is some overlap between Manx Shearwater distributions and wind farms in the Irish Sea/Solway Firth, and this could be an area in which to focus initial studies, although the number of birds moving close to/within wind farms may be small.

If a suitable site was available, GPS tracking birds could reveal macro- and meso-scale avoidance of wind farms. VHF receivers could be placed on turbines or other infrastructure, as has been done on oil and gas structures elsewhere, but obtaining sufficient sample sizes of VHF-tracked birds would be challenging. Radar can be used to quantify flight lines without the need for tagging birds, and changes to flight lines would provide evidence of avoidance or attraction. Portable radar devices are available, with a detection distance of 72 nautical miles.

Tracking or visual observations could be used to assess avoidance or attraction behaviour in relation to sound. Experiments with sound could be conducted using similar methods to those suggested for light attraction experiments, below.

Whilst no studies have been conducted to date on the energy requirements of chicks of European Storm-petrels or Fulmars, from which to inform assessment of the consequences on productivity of displacement of breeding adults from feeding areas, well—established methods are available and such studies would be feasible.

7.5 Light attraction/disorientation

The evidence need around light attraction/disorientation is high for all three species but there are challenges to addressing the knowledge gaps (Table 6). To address the current knowledge gaps regarding the spatial scale, age classes affected, environmental drivers, and influence of light characteristics on light attraction of nocturnal Procellariiformes, we recommend a series of experiments are conducted, for both Manx Shearwaters and storm-petrels, to examine the behaviour of both adults and fledglings at varying ranges from experimentally manipulated light sources. Such experiments will be logistically challenging to perform, but given magnitude of the current knowledge void, and the impact of potential light attraction on the estimates of collision rate, these studies could be considered a high priority.

Possible approaches could include use of thermal video equipment to record flight paths of adults attending the colony, and fledglings leaving the colony, in response to lights of differing wavelength, intensity and distance from the colony, under differing levels of ambient light, and visibility (i.e. foggy/clear). See Gauthreaux and Belser (2006) for an example of tracking flight paths of nocturnal migrants in relation to illuminated communications towers in USA. Monitoring for flight paths could be supplemented by targeted tracking of adults using GPS tags, and fledglings using coded VHF (MOTUS) tags. Tracking fledglings as they leave the burrow is challenging (see papers by Rodriguez et al. (2015b, 2022) for an account of the difficulties), but use of VHF tags, and a suitable array of detection stations, would overcome the difficulties of tag life and detection frequency.

Such fieldwork would require the erection of lights in view of a breeding colony, but ideally as close to the sea as possible. Lunga, Treshnish Isles might offer a suitable location where several small, low-lying skerries are situated between 700 m and 1500 m from nesting areas of European Storm-petrels and Manx Shearwaters. Such skerries would provide a suitable platform for installing lights, which could be varied in an experimental manner to systematically assess the attraction of light of differing wavelength, intensity, splay, pulse frequency etc. St Kilda may provide a suitable location for these studies on Leach's Storm-petrel and Manx Shearwater, locating test lights on the coast opposite the breeding colony on Dùn, which currently hosts about 6,000 pairs of Leach's Storm-petrel. Mousa, Shetland would provide a logistically favourable site for studies on European Storm-petrel.

Since behaviour in relation to lights near the colony may be different from behaviour at sea, experiments using lights on vessels or marine structures would also be beneficial. On-board observers and thermal imaging could be used to record the behaviour and number of birds in the vicinity.

Similar experiments, both on land and at sea, have been carried out in New Zealand by the Northern New Zealand Seabird Trust, University of Auckland and Saint Martin's University, and Lukles et al. (2021) provide useful recommendations for future work. Studies on light attraction of Leach's Storm-petrels are also being performed by researchers at Memorial University Newfoundland, using a portable radar system deployed at colonies. Coordination and discussion between research groups working on light attraction of Procellariiformes would be extremely beneficial.

7.6 Diet

While diet data do not explicitly feed into assessment methods and are therefore not considered a priority, an understanding of diet and the distribution of food resources is useful for predicting the level of impact caused by displacement or the likelihood of attraction to wind farms. Diet samples can be relatively easily collected by ringers or fieldworkers. Storm-petrels often produce regurgitates upon capture in mist nets, and faecal samples can be collected from nest sites (especially nest boxes). Obtaining diet samples from tracked birds would be particularly useful. Regurgitate and faecal samples from tracked (and untracked) European Storm-petrels have been collected on Mousa (Cardiff University/RSPB) and Treshnish Isles (RSPB), and for Leach's Storm-petrels on St Kilda (RSPB), but most have not been analysed. Molecular diet work (i.e. metabarcoding) is likely to provide more detailed information than traditional visual analysis, but visual analysis is also valuable, and much cheaper than molecular methods.

7.7 Apportioning impacts to protected colonies

Evidence needs with respect to foraging ranges and connectivity to Plan Options are high for all three species and medium with respect to colony locations (Table 6). Current knowledge of foraging ranges, locations and sizes of SPA colonies suggests that the following features of Scottish SPAs may be impacted by developments within Plan Options:

1) Manx Shearwater at St Kilda, Rum and Copeland

2) European Storm-petrel at Mousa, Auskerry, Sule Skerry, North Rona, Priest Island, Treshnish and St Kilda

3) Leach's Storm-petrel at the Flannan Isles and North Rona

Tracking of European Storm-petrels has been carried out on Mousa, Shetland over five years (2014-2018) between mid-July and mid-August, and at Lunga, Treshnish (19 individuals) for a single year in late July and August. Leach's Storm-petrels (14 individuals) have been tracked from St Kilda in a single year in July. Manx Shearwaters have been tracked from Rum, with GPS data for 20 trips from nine chick-rearing birds in 2010 and 58 trips from 15 chick-rearing birds in 2011 included in Dean et al. (2015). To establish ecological connectivity between Plan Options and these protected features, it would be beneficial to carry out tracking at the remaining breeding sites and to increase the sample of birds tracked, and the seasonal coverage of tracking, at Rum, St Kilda, Lunga and Mousa. The logistics of tracking on the Flannan Isles or North Rona would be extremely expensive and challenging, and success could not be guaranteed. Tracking at the remaining sites would be somewhat more straightforward, though not easy.

Storm-petrels breeding in the Northern Isles may be vulnerable to impacts of wind farms in the east and north-east of Scotland when departing south on migration, if they migrate southwards through the North Sea. Migration routes are currently poorly known, particularly for juveniles, which may be at particular risk of light attraction in the days and weeks after fledging. Storm-petrels can be tracked using geolocator (GLS) tags to identify migration routes and nocturnal illumination events. GLS tags require recapture of the bird, which is difficult (though possible) for adults but considerably more challenging for juveniles, which do not return to UK waters for two years. Visual inspection of light curves from eight birds tracked by RSPB (unpublished data) indicates nocturnal "light spikes" in wintering feeding areas, possibly as birds approach fishing vessels. While VHF/MOTUS tags are too large to be deployed on leg rings (for long-term studies) on storm-petrels, they could be used to collect multi-annual data on the movements of juvenile or non-breeding Manx Shearwaters. A network of receivers would need to be established to use VHF/MOTUS tags, but it may be possible to place these on turbines or other offshore structures, if incorporated at the planning stage.

It is important to note that, given the long-distance movements of these Procellariiformes, there may also be connectivity between ScotWind Plan Options and colonies outside of Scotland, and as far away as Canada in the case of Leach's Storm-petrel (Bicknell et al., 2012, Bicknell et al., 2014). Tracking of the species from colonies elsewhere would be required to determine the extent of overlap.