1. Introduction

1.1. Project background

1. Berwick Bank Wind Farm Limited (BBWFL) is a wholly owned subsidiary of SSE Renewables Limited and will hereafter be referred to as ‘the Applicant’. The Applicant is developing the Berwick Bank Wind Farm (hereafter referred to as ‘the Project’) located in the outer Forth and Tay region ( Figure 1.1   Open ▸ ).

2. The Project is located adjacent to the consented Forth and Tay offshore wind farms (OWFs) consisting of Seagreen to the north, Inch Cape to the northwest and Neart na Gaoithe to the west ( Figure 1.1   Open ▸ ).
3. The Project will, if consented, provide an estimated 4.1 GW of renewable energy, making it one of the largest OWFs in the world. Given the anticipated operational life span of 35 years, the development will make a critical contribution to Scotland’s renewable energy target of 11 GW of new offshore wind by 2030. Initially, pre-July 2021, the area was named as two separate sites, Marr Bank and Berwick Bank, but these have now been merged to a single site, Berwick Bank Wind Farm.
4. Turbine capacity is yet to be confirmed but will be between 14 – 24 MW, with a maximum number of turbines on site to be 179 - 307. Importantly, the minimum lower blade tip height is 37 m (above LAT) for all turbine options as an engineering design measure to reduce collision risk to seabirds.
5. The site boundary of the Project balances maximising the potential for renewable energy generation whilst reducing environmental impacts.

Figure 1.1:  Site boundaries for all consented and proposed wind farms currently within the Outer Firth of Forth

1.2. Study area

6. Three study areas are defined in the Berwick Bank Wind Farm Offshore Scoping Report (October 2021):

• Offshore Ornithology Regional Study Area
• Offshore Ornithology Study Area; and
• Intertidal Ornithology Study Area (reported separately in Appendix 11.2, Intertidal Survey report).
  7. The Offshore Ornithology Study Area is the focus of this technical report and is located within the survey area of the commissioned digital video aerial surveys, used to provide robust data to define baseline characteristics. The Offshore Ornithology Study Area encompasses the proposed Berwick Bank Development Array and a 16km buffer.
  8. Since digital aerial video surveys were commissioned, the Berwick Bank Development Array boundary has reduced (see Chapter 4), and the Offshore Ornithology Study Area was adjusted to reflect this. Differences between the two study areas are indicated in Figure 1.2   Open ▸ . For the purposes of this report and all other reports for this submission which pertain to offshore ornithology, the original Development Array boundary and 16km buffer flown during digital aerial surveys will be referred to as ‘the survey area’ and the updated Development Array plus 16 km buffer will be referred to as the ‘Offshore Ornithology Study Area’. Since the entirety of the Offshore Ornithology Study Area falls within the survey area, no additional data collection was necessary; the survey design for digital aerial surveys within the Offshore Ornithology Study Area is indicated in Figure 1.3   Open ▸ . Total areas for the survey area, Offshore Ornithology Study Area and Development Array area were calculated at 4.981 km2, 3,975 km2 and 1010 km2 respectively. Information from digital aerial surveys will also contribute to the characterisation for the proposed export cable corridor, as outlined in Chapter 11.  
  9. The Offshore Regional Study Area is considered in the Offshore Environmental Impact Assessment Report (EIAR). Its area is defined by the mean-maximum breeding season foraging range (plus one standard deviation (SD)) of gannet, which has the largest foraging range of the key species considered in the ornithology assessment (315.2 km ± 194.2 km (Woodward et al., 2019)). Consequently, the maximum extent of the Offshore Ornithology Regional Study Area was calculated as 509.4 km from the Project.
  10. Within the Offshore Ornithology Regional Study Area, there are many designated protected sites for seabirds in both breeding and non-breeding seasons. Those closest to the Project include the Forth Islands SPA; Fowlsheugh SPA; St Abb’s Head to Fast Castle SPA; and Outer Firth of Forth and St Andrew’s Bay Complex SPA.

Figure 1.2:  Original Berwick Bank Development Array plus 16km buffer (‘the survey area’) flown during digital aerial surveys and updated Berwick Bank Development Array and 16 km buffer (the ‘Offshore Ornithology Study Area’)

Figure 1.3:  The ‘Offshore Ornithology Study Area’ comprising the updated Berwick Bank Development Array and 16km buffer. The area was surveyed using digital video aerial surveys of 2 km spaced transects

1.3. Purpose of the report

11. This technical report provides the baseline ornithological characterisation for the Offshore Ornithology Study Area ( Figure 1.2   Open ▸ ).
12. The report provides baseline information on the seasonal distribution, density, and abundance of seabirds based on:

• available reports and literature (“Desktop study” Section 2), and
• analysis of data from a series of digital aerial surveys undertaken in the period 2019 - 2021 (“Berwick Bank digital aerial surveys” Section 3).
  13. Additional Annexes (A – K) provided separately to this report, detail analytical outputs on species monthly density and abundance estimates.
  14. The species density and abundance estimates presented in this report (Section 4) underpin further technical reports regarding the collision risk, displacement, and population-level effects of the Project on the seabird community present within the Offshore Ornithology Regional Study Area. All technical reports provide information to support the production of the Berwick Bank Wind Farm Offshore EIAR and Habitats Regulation Appraisal (HRA).

2. Desktop Study

2.1. Key data Sources

15. A desk-based review was undertaken of existing literature and data sources that were site-specific or relevant to the Offshore Ornithology Study Area. These are summarised in Table 2.1   Open ▸ below.

Table 2.1:   Summary of key desktop reports

2.2. Review of additional bird data relevant to the baseline

2.2.1.    Berwick Bank boat-based Seabird Surveys (2020 – 2021)

16. Site-specific boat-based surveys were undertaken across the Berwick Bank Development array during July-August 2020 and April-June 2021. The surveys were designed to collect flight height data to potentially inform collision risk monitoring for the Project, and to trial measurement techniques. The use of the generic flight height data of Johnston et al. (2014) is currently the advised source for collision risk modelling, but the flight height data on kittiwake collected visually and with the use of a laser rangefinder were analysed and presented to provide additional context in Appendix 11.3: Ornithology Collision Risk Modelling Technical Report.  
17. Sampling blocks were repeatedly sampled along a continuous transect route broadly following methods as described by Embling et al. (2012; Figure 2.1   Open ▸ ). Where possible the surveys recorded detailed information on the species’ age and plumage of each bird encountered, along with flight height, flight direction, behaviour, interaction with other birds, incidental marine mammal sightings, vessels, and association with other features of interest such as tidal fronts, upwellings, current, detritus or flotsam etc.
18. Flight height recording followed the previous boat-based survey method used for Seagreen in 2017 (Harwood et al., 2018), with flight heights visually estimated in 5m bands. Surveyors also used laser rangefinders when not surveying the line transect or when very few birds were present. Rangefinder data formed the basis of the flight height distribution dataset with visual flight heights used for comparison. To increase data collected using rangefinders additional rangefinders were used opportunistically by other observers. The survey methods for the boat-based surveys undertaken in 2020 and 2021 were agreed with ornithological advisors at Marine Scotland, NatureScot and RSPB.
19. A total of 47,777 birds were recorded of 37 identified species. Five key species accounted for ~89% of the observations: guillemot (32.09%), kittiwake (23.67%), gannet (16.12%), razorbill (9.68%) and puffin (7.26%). Flight height data were not able to be recorded for every bird. For example, at periods where birds ‘swamped’ the sampling platform, species which could be easily targeted with rangefinders (e.g. gannet) were preferentially selected. For gannet and kittiwake, flight height data were collected for 88% and 70% of total observations (visual estimation plus laser rangefinder data), respectively. Gannet and kittiwake rangefinder data recorded flight heights between 0-78.8 m and 0.6-68.6 m respectively, collected from 1,229 observations. Further details are given in Appendix 11.7, Boat-based Survey Report.

2.2.2.    Seagreen Boat-Based Transect Surveys (2009 to 2011)

20. Boat-based surveys were carried out for the first phase of baseline data collection for the former Firth of Forth Round 3 Zone across the Seagreen (Alpha – northwest, and Bravo – southeast) site between December 2009 and November 2011. The surveys were designed to characterise baseline ornithology to inform EIA.
21. As this dataset is now over ten years old and the abundance of seabirds has likely changed since the data were collected, it provides contextual information to support the primary baseline data source, which are the digital aerial surveys of the survey area undertaken between March 2019 and April 2021.
22. A total of 23 monthly boat-based transect surveys were undertaken between December 2009 and November 2011, covering the former Firth of Forth Round 3 Zone, which includes the Project and the Seagreen site. Transects were spaced 3km apart and oriented northwest to southeast to intercept the likely predominant flight lines from major breeding colonies in the Firth of Forth.
23. A total of 24,206 birds of 39 species and 20,436 birds of 37 species were recorded at the Alpha and Bravo sites respectively. Within the Alpha site, guillemot (28.10%), kittiwake (24.80%), and gannet (16.10%) were the predominant species recorded. For the Bravo site, guillemot (29.30%), kittiwake (21.60%), and gannet (16.60%) were the most frequently recorded species.

Figure 2.1:  Boat based surveys undertaken by ECON and RPS during July-August 2020 and April-June 2021

2.2.3.    Seagreen Pre-Construction Monitoring surveys

24. Dedicated digital aerial surveys of the Seagreen site were commissioned between March 2019 and September 2020 to provide data for pre-construction monitoring.
25. Twenty-six strip transects were flown on a roughly monthly basis over the site plus a 12 km buffer. Two kilometre transect spacing was used, with transects orientated roughly perpendicular to the main environmental gradients, such as depth. The proximity of the Seagreen site to the Offshore Ornithology Study Area resulted in approximately 25% overlap of the Seagreen pre-construction monitoring area with the Offshore Ornithology Study Area, equating to nearly 2,000 km2. Preliminary evaluation of the data suggested the presence of several seabird species of importance, including:

• Qualifying species for nearby SPAs:

-          Guillemot, Arctic tern, Sandwich tern, gannet, kittiwake, herring gull, lesser black-backed gull, puffin, little gull, red-throated diver, Manx shearwater, and razorbill.

• European Union (EU) Bird Directive Annex 1 species:

-          Guillemot, red-throated diver, Manx shearwater, Sandwich tern, and Arctic tern.

• Species present in regionally important abundance:

-          Guillemot, gannet, kittiwake, razorbill, and puffin.

26. Baseline data collected during the surveys suggest that kittiwakes, gannets, guillemots and puffins use the area during the breeding season, occurring in relatively high abundance. Relatively high numbers of Arctic terns were also observed in the breeding season. These data suggest that the Seagreen site is used frequently during post-breeding migration by lesser black-backed gulls, guillemots and razorbills. During return migration in the spring, kittiwakes were regularly observed. It is likely that a small population of puffins also utilises the area during the non-breeding season.

2.2.4.    JNCC Seabirds at Sea data: Effort related seabird observation data collected from boats using European Seabirds at Sea (ESAS) methodology

27. The European Seabirds at Sea (ESAS) database holds the most comprehensive data on the abundance and distribution of northwest European seabirds, spanning over 30 years. Camphuysen et al. (2004) developed methods to assess bird distribution and abundance at sea to evaluate impacts of OWFs on bird populations. Line transect distance sampling methodology was the recommended census technique for boat-based surveys, using distance bands and detection functions to correct for the uncertain detection at increasing distances from the survey vessel.
28. The summed counts from the 84 boat-based ESAS surveys (1980 and 1996) trimmed to the Offshore Ornithology Study Area are presented in Table 2.2   Open ▸ . Data were downloaded from https://hub.jncc.gov.uk/assets/5c7d5eca-9b5f-4781-809f-f27c94d94661. These data provide contextual information regarding the species and relative abundance of species likely to be present within the Offshore Ornithology Study Area.

Table 2.2:  Total birds identified within the Offshore Ornithology Study Area from ESAS boat-based survey data between 1980 and 1996

2.2.5.    Analysis of European Seabirds at Sea (ESAS) data to identify suitable locations for marine SPA’s within the British Fishery Limit

29. The ESAS database was used to determine the location of seabird assemblages and identify areas of high seabird densities to ultimately inform the designation of marine SPAs within national waters. Analyses conducted by Kober et al. (2010; 2012) used data from boat-based platforms of opportunity, with those relating to the North Sea region of relevance to the Offshore Ornithology Study Area. Data concerning sitting and flying birds between 1980 and 2004 were collected using line transects and the snapshot method respectively, as described by Camphuysen et al. (2004) and Tasker et al. (1984) and analysed using distance sampling.
30. The Outer Firth of Forth/Wee Bankie/Marr Bank region and the Inner Firth of Forth region were highlighted as important areas for multiple seabird species, such as breeding gannets, puffins and guillemots as well as wintering kittiwakes and puffins. Hotspot analyses identifying the presence of high densities of seabird species within the area supported the classification of the Outer Firth of Forth and St Andrews Bay Complex SPA which is in close proximity to the Offshore Ornithology Study Area.

2.2.6.    IMPRESS final report: Interactions between the marine environment, predators and prey, implications for sustainable sandeel fisheries – Camphuysen et al., 2004

31. The IMPRESS project aimed to quantify the relationships between sandeel populations, hydrography and foraging success and breeding performance of gannets, shags Gulosus aristotelis, kittiwakes and guillemots across the Wee Bankie/Marr Bank complex in the Outer Firth of Forth.
32. The IMPRESS study reported high abundance of kittiwakes, guillemots and gannets within the Berwick Bank development zone during the breeding season. These data will not be used during this assessment but are included to indicate which species are likely to be present in important numbers within the Berwick Bank Offshore Ornithology Study Area.
33. Colony-based studies of seabird breeding populations were conducted, one on the Isle of May considering shags, guillemots and kittiwakes annually between 1997 and 2003, and for gannets on Bass Rock in 1994 and 2004. Colony counts suggested breeding success for gannets and shags in contrast to that for guillemots where declines in breeding success were observed. Kittiwake breeding success was more variable.
34. Boat-based line transect surveys studying seabird distribution, abundance and behaviour were conducted throughout the Outer Firth of Forth region between 2001 and 2004. Data were collected using standard ESAS methods (Tasker et al., 1984). Surveys identified 37 seabird species, the most abundant of which were guillemot (49%), puffin (21%), kittiwake (12%), razorbill (11%) and gannet (6%). Distribution patterns indicated gannets ranged far from colonies, with guillemots and kittiwakes concentrated in the west of the region and shags distributed in coastal regions.

2.2.7.    ISLE of May long-term study (IMLOTS)

35. The Isle of May long-term study (IMLOTS) aims to assess the impacts of environmental change on seabirds and their surrounding ecosystems on the Isle of May, in the outer Firth of Forth. As part of the Seabird Monitoring Programme (SMP), hosted by BTO and managed by JNCC, data concerning seabird breeding success has been continuously collected since 1982 for puffin, guillemot and razorbill, and since 1987 for shag, kittiwake and fulmar. Seabird breeding success is monitored by collecting data on the number of active nests and the number of chicks fledged per active nest.
36. Data collected through IMLOTS will not be directly applied to this report but are instead provided for context regarding which species are likely to be encountered.

2.2.8.    Seabird tagging Data

37. GPS loggers are used to track fine scale movements of many seabird species within the marine environment, the results of which are used in a variety of micro, meso and macro scale applications.    
38. Wakefield et al. (2017) tracked 1,313 seabirds from 29 UK colonies using GPS loggers to determine coarse-scale breeding seabird distributions. Foraging ranges were variable between species with shags staying close to nest sites (median 3.4 km, interquartile range (IQR) 1.6–7.5) and kittiwakes (11.9 km, IQR 4.2–30.9), razorbills (13.2 km, IQR 5.1–26.2) and guillemots (10.5 km, IQR 3.2–19.1) travelling further afield. The study also predicted that breeding shags, kittiwakes, guillemots and razorbills primarily foraged within 100km of the coast in Scotland. Tagging data from the Isle of May was included in the analyses, however site-specific data were not presented.
39. Cleasby et al. (2018) built on Wakefield et al. (2017) using seabird telemetry data collected between 2010 and 2015 to create species distribution models and identify seabird hotspots at the UK- and SPA-level. Four species were considered: shag, kittiwake, guillemot and razorbill. SPA-level hotspots for kittiwake identified the east of Scotland as being of importance to the species, in particular those associated with Fowlsheugh SPA to the northeast of the Offshore Ornithology Study Area. Hotspots at the SPA-level of relevance to the Project were also identified for shags at the Forth Islands SPA.  
40. Lane et al. (2020) collected telemetry data from 154 Bass Rock gannets during the breeding season between 2015 and 2019 to assess the risk of consented and proposed offshore wind farms within foraging range. Gannets generally flew northeast and southeast from the colony with males and females spending more time within proposed offshore wind farm sites during chick rearing than during pre-hatching. The potential collision risk was estimated to be eight times higher during this period, and also to be three times higher for females than males due to their higher flight heights, longer trip duration and increased time spent within proposed wind farm sites. The highest recorded densities of gannets from telemetry data overlapped with proposed wind farm sites within the outer Firth of Forth, with 99% of predicted collisions during the breeding season occurring in this region, compared to other wind farm sites which are within the foraging range of Bass Rock gannets but which are considerably more distant. Additional telemetry data have since been gathered from the breeding gannet colony on the Bass Rock since this study was published, which demonstrate a high degree of inter-annual variation in these patterns (Lane and Hamer, 2021).
41. Indeed, Lane and Hamer (2021) undertook additional tracking work from the Bass Rock, funded by developers from proposed and consented wind farms within the Forth of Tay (Neart na Gaoithe, Seagreen, and the Project), to determine the impacts of offshore wind farms on adult gannet foraging, survival and population dynamics. Tracks from both sexes overlapped with planned OWF sites in the outer Firth of Forth, although the proportion of tracks falling within these areas was higher in males (rather than females as found by Lane et al. 2020). Distance travelled and foraging trip duration were shorter compared to Lane et al. (2020). Trip duration, distance travelled and distance of displacement from the colony was consistent between sexes. Sex-specific survival was likely to be variable, although no trends could be determined.
42. Bogdanova et al. (in prep.) presents the results of GPS tracking of adult guillemots, razorbills and puffins breeding on the Isle of May, and of kittiwakes breeding on the Isle of May, St Abb’s Head and Fowlsheugh in June and July 2021. An assessment of connectivity with three OWF sites in the outer Firth of Forth (Neart na Gaoithe, Seagreen and the Project) was conducted. Locational data were obtained from 23 guillemots, 11 razorbills, 24 puffins and 50 kittiwakes on the Isle of May, 40 kittiwakes at Fowlsheugh and 37 kittiwakes at St Abb’s Head.

2.2.9.    WWT waterbird surveys

43. Waterbirds and seabirds in inshore waters were monitored regularly by WWT Consulting between 2004 and 2009, to provide a comprehensive and robust assessment of UK inshore environments (e.g., Calbrade et al. 2009; Austin et al. 2007). Visual aerial surveys were conducted, using transects spaced at 2 km intervals orientated north to south. Survey methodology and distance sampling protocol from Buckland et al. (2001) and Camphuysen et al. (2004) was adopted.
44. Although the study did not cover the region considered for the Project, it gave reliable population estimates for waterbird species in the wider vicinity of the Project and indicates which species are likely to be present.
45. The “NE1” sampling block is closest to the Offshore Ornithology Study Area, located approximately 20 km from the southwest boundary of the buffer. Auks and gannets were identified as the most abundant species, with 6,199 birds recorded in the survey block overall. However, only non-breeding season surveys were targeted, and it is likely other species of interest may be present at other times.

3. Berwick bank digital aerial surveys

3.1. Overview of survey flights

46. Dedicated digital video aerial surveys conducted by HiDef Aerial Surveying Limited (hereafter “HiDef”) were deployed for ornithological monitoring over the Offshore Ornithology Study Area using methods described in Buckland et al. (2012), Weiss et al. (2016) and Webb and Nehls (2019).
47. The aircraft flight altitude, >500 m, and other technical parameters of the adopted method conform with the guidance of Thaxter and Burton (2009) and updated in Thaxter et al. (2016). To complete the survey coverage in a single day and obtain a suitable snapshot, HiDef deployed two to four aircraft on each occasion to survey the area simultaneously with each aircraft surveying unique subsets of survey transects ( Figure 3.1   Open ▸ ).
48. The survey programme ran from March 2019 to April 2021 and aimed to survey the Offshore Ornithology Study Area at monthly intervals or as close to this as possible. A total of 25 surveys were successfully flown, with a summary of survey flights presented in Table 3.1   Open ▸ .
49. 
    
    The COVID-19 pandemic halted all non-essential survey activity during April 2020, however surveys resumed in May 2020 based on updated risk assessments and stringent COVID-19 mitigation measures. Due to this and further complications (e.g., industrial action of airport personnel), complete coverage was not achieved in all months.
50. The April 2019 survey was missed due to bad weather, the January 2020 survey was flown at the beginning of February 2020, and two surveys were undertaken in May 2020 due to the April 2020 survey being cancelled. For the analysis of the data, some flights were assigned to different months where there was no survey to ensure coverage of all months in both seasons for a two-year period (see section 3.2.8). The Applicant discussed this allocation during the Ornithology Road Map process (RM4) and followed subsequent joint advice from Marine Scotland and NatureScot received through email on 14 January 2022.
51. 
    For all surveys between July 2019 and January 2020, varying proportions of the intended transects were not flown. However, the target coverage for the site (~12.5%) was achieved by increasing the number of cameras from which the imagery was processed. Whilst this increased overall coverage, there remained limited “gaps” spatially where transects could not be surveyed (e.g., December 2019; Figure 3.2   Open ▸ ). Despite differences in survey coverage in some months, the survey results are of sufficient quality to provide robust baseline characterisation of the Offshore Ornithology Study Area.
52. Flight times for each survey are presented in
53. Table 3.2   Open ▸ . HiDef’s survey times are targeted at times of day that exclude 1.5 hours of sunrise and sunset in summer and winter. The dawn and dusk periods are excluded because the sun angle and light levels are too low for digital imagery at these times of the day. Dawn and dusk periods are also often out with airport opening hours.
54. Additional environmental information, such as time of sunset and high tide times, is also presented in
55. Table 3.2   Open ▸ . The extent of the survey programme ensured that sampling occurred over a range of tidal states. Tidal state can influence the activity and habitat use of many bird species (Gilbert et al., 1998) but given the offshore location of the Project, the influence of tidal state on seabird densities is also likely lower than in coastal areas.
56. Aircraft flight patterns are displayed in Figure 3.2   Open ▸ , Figure 3.3   Open ▸ and Figure 3.4   Open ▸ . Variation in presentation of track data is due to differing GPS equipment used in some surveys. The same transect lines were intended to be flown in each survey, however, effort differed between surveys due to slight differences in start and stop times and minor deviations of the aircraft from the transect line. It is important to note that aircraft flight patterns follow the original survey design covering the survey area (contractually agreed with the Applicant in September 2019) and extend beyond the Offshore Ornithology Study Area. The survey design was presented to Consultees through at the first meeting of the offshore Ornithology Road Map process (Technical Appendix 11.8).

Figure 3.1:  Transect splits between aircraft during summer months (April – September) A) 2 aircraft B) 3 aircraft and C) 4 aircraft, during autumn and early spring months (October – March) over the survey area. Note: flight paths follow original survey design (agreed in September 2019)

Table 3.1:  Survey effort and flight information for the 25 surveys of the survey area (original Berwick Bank Development Array plus 16 km buffer)

1April 2019 missed due to bad weather

2January 2020 survey undertaken on 05/02/2020.

3April 2020 survey suspended due to Coronavirus. Two surveys undertaken in May 2020 instead.

4Two surveys flown in April 2021 to ensure representative samples of “April” were available.

For all surveys between July 2019 and January 2020, extra cameras were reviewed to achieve required coverage and to compensate for missed transects.

Table 3.2:  Survey summary of flight dates, flight times and tidal states over the survey area (original Berwick Bank Development Array plus 16 km buffer)

Figure 3.2:  Aerial survey transect coverage for each survey over the survey area (original Berwick Bank Development Array plus 16 km buffer) between March 2019 and March 2020

Figure 3.3:   Aerial survey transect coverage for each survey over the survey area (original Berwick Bank Development Array plus 16 km buffer) between May 2020 and April (S01) 2021

Figure 3.4:  Aerial survey transect coverage over the survey area (original Berwick Bank Development Array plus 16 km buffer) during April (S02) 2021

3.2. Methodology

57. A series of strip transects were flown across the survey area, extending beyond the Offshore Ornithology Study Area, between March 2019 and April 2021.
58. The survey design consisted of 37 strip transects extending roughly north-west to south-east, perpendicular to the depth contours along the coast. Such a placement helped to ensure that each transect sampled a range of habitats (primarily relating to water depth), to reduce variation in bird and marine mammal abundance estimates between transects.
59. Transects were placed at 2 km apart across the survey area, resulting in an overall survey area of approximately 4,981 km2. The Offshore Ornithology Study Area covers 3,975 km2 and is entirely within the survey area.
60. Surveys were undertaken using an aircraft equipped with four HiDef Gen II cameras with sensors set to a resolution of 2 cm Ground Sample Distance (GSD). Each camera sampled a strip of 125 m width, separated from the next camera by ~20 m, providing a combined sampled width of 500 m within a 575 m overall strip ( Figure 3.5   Open ▸ ).
61. Surveys were flown along the transect pattern shown in Figure 3.2   Open ▸ for example, and at a height of approximately 550 m above sea level (ASL) (~1800’). Flying at this height ensures that there is no risk of flushing those species easily disturbed by aircraft noise. Thaxter et al. (2016) recommends a minimum flight altitude of 460-500 m ASL.
62. Position data for the aircraft was captured from a Garmin GPSMap 296 receiver with differential GPS enabled to give 1m accuracy for the positions and recording updates in location at one second intervals for later matching to bird and marine mammal observations.

Figure 3.5:  Schematic depicting the digital video aerial survey methodology

63. For most surveys, data from two out of the four cameras were processed to achieve a minimum target of 12.5% site coverage. This ensured sufficient coverage and number of transects, whilst allowing the remaining unprocessed data to be archived.
64. Where survey transects had been compromised, additional cameras were processed to achieve the 12.5% target. This was necessary for surveys from July 2019 to January 2020 inclusive.

3.2.2.    Image processing

65. Data were viewed by trained reviewers who marked any objects in the footage as requiring further analysis by the Identification (ID) Team, separating them into broad categories of birds, marine megafauna or anthropogenic objects.
66. An object was only recorded where it reached a reference line (‘the red line’) which defined the true transect width of 125m for each camera. By excluding objects that do not cross the red line, biases to abundance estimates caused by flux (movement of objects in the video footage relative to the aircraft, such as ’wing wobble’) are eliminated.
67. As part of HiDef’s quality assurance (QA) process, an additional ‘blind’ review of 20% of the raw data was carried out and the results compared with those of the original review. If 90% agreement was not attained during the QA process, then corrective action was initiated: the remaining data set was reviewed and where appropriate, the failed reviewer’s data discarded and all the data re-reviewed. Additional training was then given to the reviewer to improve performance.

3.2.3.    Species identification

68. Images marked as requiring further analysis were reviewed by specialist ornithologists for identification to the lowest taxonomic level possible and for assessment of the approximate age and the sex of each animal, as well as any behaviour traits visible from the imagery.
69. At least 20% of all birds were selected at random and subjected to a separate ‘blind’ QA process. If less than 90% agreement was attained for any individual camera then corrective action was initiated: if appropriate, the failed identifier’s data were discarded, and the data re-identified. Any disputed identifications were passed to a third-party expert ornithologist for a final decision.
70. All birds were assigned to a species group and where possible, each of these were then further identified to species level. Identifications were given a confidence rating of ‘possible’, ‘probable’ or ‘definite’.
71. Any birds that could not be identified to species level were assigned as ‘No ID’. If, on occasion, the unidentified bird is suspected of belonging to two possible genera, then a broader group category may be used. For example, a bird would usually be assigned to the group category ‘Shearwater species’ if identified as a Manx shearwater, or to ‘Auk species’ if identified as a guillemot. However, if the bird has the potential to be either, then it would be assigned to the group category ‘Shearwater / Auk species’ and the species level recorded as ‘No ID’.
72. A list of scientific names and taxonomic groupings used in the study are presented in Annex A.

3.2.4.    Age, Sex and Behaviour 

73. Where possible, birds were assessed for approximate age and sex. Aging of birds was based on moults and was therefore conducted only on flying individuals and species which show seasonal variation in plumage.
74. The behaviour of each bird was recorded. The direction of travel was also noted in the case of flying birds.

3.2.5.    Geo-referencing

75. All data were geo-referenced, taking into account the offset from the transect line of the cameras, and compiled into a single output; Geographical Information System (GIS) files for the Observation and Track data are issued in ArcGIS shapefile format, using UTM30N projection, WGS84 datum.

3.2.6.    Bird abundance and density estimation

76. The abundance of each species observed was estimated separately using a design-based strip transect analysis. During analysis, survey area data were trimmed to the Offshore Ornithology Study Area. All subsequent data within this report refer only to the updated Berwick Bank Development Array and corresponding 16km buffer (the Offshore Ornithology Study Area), although the various annexes also provide estimates for the Development Array, see Table 5.1   Open ▸ for description of Annexes.
77. Each transect is treated as an independent analysis unit, and the assumption is made that transects can be treated as statistically independent random samples from the site. The length of each transect and its breadth (i.e. the width of the field of view of the camera) multiplied together give the transect area; dividing the number of observations on that transect by the transect area gives a point estimate of the density of that species for the transect. The density of animals at the site (and hence the population size), the standard deviation, the 95% Confidence Intervals (CIs) and coefficient of variance (CV) are then estimated using a non-parametric bootstrap method with replacement (Buckland et al., 2001).
78. The upper and lower 95% confidence intervals were performed by way of a blocked bootstrapping technique to ensure equal transect effort was sampled across each iteration. This was done by using transect ID as the sampling unit with replacement, and then randomly sampling until the total length of the sampled transects equalled approximately the same length as the total survey length. A total of 500 bootstrap iterations were performed from which mean and standard deviation of the sampled means were calculated, as well as the relative standard error as defined by the standard deviation divided by the mean (or the “Coefficient of Variation”, CV). Data were processed in the R programming language (version 4.0.4).
79. The density estimate is expressed as the average number of animals per square km surveyed over the whole site, and the population estimate is then calculated as the average density multiplied by the area of the whole site. The standard deviation is a measure of the variance of the population estimate, standardised by the number of samples (transects). The upper and lower CI define the range that the population estimate falls within with 95% certainty. The CV is a measure of the precision of the population and density estimates.
80. For most species these abundance estimates relate to absolute abundance, but for diving species (such as auks) the abundance relates to relative abundance due to a proportion of animals being submerged at the time of survey.
81. Density and abundance can also be derived using model-based methods. These methods tend to result in similar estimates as from design-based methods, and this was verified by completing model-based density and abundance estimates for five focal species to compare with the design-based estimates: kittiwake, guillemot, razorbill, puffin and gannet. HiDef undertook density surface modelling using the Marine Renewables Strategic Environmental Assessment (MRSea Windows Package package) in R (Scott-Hayward et al., 2013). HiDef adapted and customised some of the MRSea code so that the modelling approach could cope with the specific nuances of the Berwick Bank data (code can be made available on request). This work was undertaken through consultation with the package author (Scott-Hayward) and was reported to consultees during the Road Map consultation process. The methodology and in depth-results are reported in Annex L to this report. The outputs that allow comparison to the design-based estimates are presented in Section 4.1.

Apportioning of unidentified birds

82. Apportioning of ‘unidentified’ birds to species level was also undertaken for the purposes of calculating population estimates. The number of unidentified birds in each species group were assigned to species where appropriate, based on their respective abundance ratios. For example, if identified guillemots and razorbills occurred in a 4:1 ratio, then 80% of unidentified birds would be assigned to guillemot and 20% assigned to razorbill. Apportioned estimates are presented in text, with apportioned and unapportioned estimates presented in the Annexes for reference.

Centcount maps

83. Cent-count maps were created to show the distribution of species. To account for varying survey effort for visualization, design-based population estimates, and statistical modelling, observations were run through a HiDef data aggregation tool which organized those data into regularly spaced bins along transect lines. This tool works by interpolating GPS tracks from the aircraft and locating points along those tracks which are approximately 500 m apart (i.e., centroids of 500 m bins). These 500 m bins then use a nearest neighbour technique to aggregate observations into those centroid locations (i.e., observations are aggregated into the nearest centroid point along the line). To calculate the area surveyed for each bin, the length of the bin is multiplied by the strip width of the cameras (125 m x n cameras, where n is the number of cameras used in that transect). The circles represent the number of birds in each 500 m bin. The larger the circle, the greater the number of birds present.
84. Where more than 40% of the surveys had observations, all surveys have been displayed in centcount maps, including those with no observations. This is with the exception of Survey April (S02) 2021 which was only mapped where observations were present. Where less than 40% of surveys had observations, only surveys with observations have been displayed in maps.