1. Introduction

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.
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 proposed Berwick Bank development will, if consented, provide an estimated 4.1 GW of renewable energy, making it one of the largest offshore wind farms 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. 
4. Turbine capacity will range from 14 – 24 MW per machine, with a maximum number of turbines on site to be 179 - 307. As part of ensuring minimised impacts to wildlife, such as potential displacement of seabirds, the Berwick Bank Development array area was reduced by approximately 20% in May 2022, from 1,314 km2 to 1,010 km2.

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

2.             Purpose of the report

2. Purpose of the report

5. This Technical Report assesses the potential effects of displacement on seabirds during the operational phase of the proposed Berwick Bank OWF primarily based on the interim advice of the joint Statutory Nature Conservation Bodies (SNCBs, 2017) on a Matrix Approach to assessment. Further analyses are presented using the SeabORD modelling tool (Searle et al., 2018) as requested in the Scoping Opinion (4 February 2022), for context. These approaches are described in Section 3 and full results provided in Annex D.
6. Furness et al. (2013) defines displacement as ‘a reduced number of birds occurring within or immediately adjacent to an offshore wind farm’, involving birds present in the air and on the water (SNCBs, 2017). Birds that do not intend to utilise an offshore wind farm, but would have previously flown through the area, and which either stop short or detour around a development, are subject to barrier effects (SNCBs, 2017). For the purposes of assessment, it is usually not possible to distinguish between displacement and barrier effects (e.g., to determine if individual birds may have intended to travel to, or beyond an offshore wind farm, even when tracking data are available). Vessel and helicopter traffic associated with OWFs also have the potential to cause temporary disturbance to sensitive species, with some species avoiding the area altogether, potentially resulting in a loss of optimal rafting, foraging and moulting habitat. Displacement affects species differently, with the potential to have population level impacts for species which are less adaptive or highly constrained in their foraging range, such as in the breeding season.

3.             Assessment Approaches

3. Assessment Approaches

3.1.        Overview of approaches

3.1. Overview of approaches

7. Consultation Representations and Advice from MSS and NatureScot (4 February 2022) and discussions through the Ornithology Road Map process (Appendix 11.8), led to agreement that displacement assessment was required for five species:

• kittiwake Rissa tridactyla;
• guillemot Uria aalge;
• razorbill Alca torda;
• puffin Fratercula arctica; and
• gannet Morus bassanus.
  8. Species were selected based on their abundance in the proposed Berwick Bank Development Array area, highlighted by the two years of baseline data (Appendix 11.1: Ornithology Baseline Technical Report), and on evidence about their sensitivity to displacement and barrier effects (Furness et al., 2013; Bradbury et al., 2014; SNCB, 2017)).
  9. The Scoping Opinion recommended that estimates of displacement and barrier effects as generated by the publicly available individual-based modelling approach “SeabORD” should be presented for kittiwake, guillemot, razorbill and puffin, where feasible (Searle et al., 2018). 
  10. SeabORD is intended to simulate the behaviour and energetics of individual birds from breeding seabird populations under baseline conditions (i.e. with no offshore wind farm present) and compares the resulting demographic estimates to model runs undertaken in scenarios which have the offshore wind farm(s) of interest present (so that birds undertaking foraging trips from the colony have the potential to incur energetic costs from barrier effects and of increased intra-specific competition for food if they are displaced). These effects are estimated in terms of changes to adult and chick mortality, with the available outputs relating to the individual SPA populations that are of interest to the assessment. The estimated mortality to adult birds relates only to the breeding period.
  11. SeabORD relies upon predictions of the distribution of seabird prey resources and of foraging birds. Both of these aspects are determined by the availability of Global Position System (GPS) tracking data from breeding birds associated with the colonies of interest. In addition, the model is underpinned by a range of other assumptions and predictions (e.g. on the relationships between adult body mass and survival), each of which have associated uncertainties (Vallejo et al., 2022 (volume 3, appendix 11.4, annex H)).
  12. Details of the SeabORD modelling undertaken for the Proposed Development are provided in Annex D. An assessment of the uncertainty and validity of the underlying model parameters and assumptions is presented in Vallejo et al., 2022 ((volume 3, appendix 11.4, annex H)).
  13. Since SeabORD does not include gannet, Marine Scotland Science, in their scoping representation of 16th December 2021, advised that an analysis of the extensive GPS tracking data be undertaken to inform assessment of displacement and barrier effects for this species. Details of the analysis undertaken are given in Annex E, following the approach agreed through the Ornithology Roadmap Process (RM6; Appendix 11.8).  
  14. Given the issues encountered with SeabORD, as outlined in Annexes D and H, and discussed during the Ornithology Roadmap Process (RM4 and RM5; Appendix 11.8), the SNCB matrix method was used as the primary method for assessment of displacement effects for each of the five relevant species (SNCBs, 2017). The matrix provides a table of the displacement rates, from zero per cent to 100 per cent, against mortality rates, again from zero per cent to 100 per cent. For a given population-size and any combined value of displacement rate and mortality rate, the matrix provides a prediction of the number of birds that may die as a result of displacement from an offshore wind farm. Although the estimated effects are derived by applying specified displacement rates, the resulting predicted impacts are assumed to encompass both displacement and barrier effects.
  15. Seasonally specific displacement and mortality rates were recommended by NatureScot and Marine Scotland Science in their scoping representations of 7 and 16 December 2021, respectively (the “Scoping Approach”; Section 3.5). In line with the evidence presented in Annex G, an additional set of displacement and mortality rates have also been taken forward for assessment (the “Developer Approach”; Section 3.5).
  16.  Displacement matrices were produced for each of the five species, using a number of species-specific parameters:
      1. spatial extent – the distance from turbines that displacement impacts are considered likely to affect the species;
      2. mean seasonal peak population – a mean of the estimated number of birds within the impacted area in each appropriate bio-season;
      3. displacement level - the percentage of the population assumed to be displaced from the impacted area; and
      4. mortality level – the percentage of displaced birds assumed to die as a consequence.

3.2.        Spatial Scales

3.2. Spatial Scales

17. Following the joint SNCB interim advice (SNCB, 2017), and as advised in the Scoping Opinion (4 February 2022), displacement matrices were formulated for two separate spatial scales:

• the proposed Berwick Bank Development Array; and
• the proposed Berwick Bank Development Array plus a 2 km buffer.
  18. The Project Design Envelope (PDE) is based on the following design principles: minimum turbine spacing of four rotor diameters; and maximum turbine spacing of 15 rotor diameters. The Development Array covers 1,010 km2, with between 179 and 307 turbines. As such, there are likely to be large distances between the largest turbines, with each turbine potentially spaced between a minimum of 1km and a maximum of 3.33 km (14MW) or 4.65 km (24 MW) apart.

3.3.        Seasonal Definitions

3.3. Seasonal Definitions

19. The Matrix Approach requires potential displacement to be assessed separately for species in the breeding season and non-breeding season, where appropriate.
20. In previous assessments for consented Forth and Tay OWFs, displacement of guillemot, razorbill, puffin and kittiwake were assessed quantitatively in the breeding season. In the non-breeding season, guillemot and razorbill were also assessed quantitatively, with only a qualitative assessment required for puffin and kittiwake for some projects. This is because displacement is not considered to limit these species in the non-breeding periods when birds are not constrained by having to return to colonies, or, in the case of puffin, because they disperse rapidly and widely after the breeding season. This is the basis of the “Developer Approach” presented.
21. However, following the Scoping representations from MSS and NatureScot (December 2021) and Scoping Opinion (4 February 2022) non-breeding season displacement has been assessed quantitatively for kittiwake and gannet; there is no requirement to assess non-breeding season impacts for puffin (“Scoping Approach”).
22. Seasonal definitions are based on NatureScot guidance (2020); this was agreed through the Ornithology Roadmap process (RM1). Seasonality is complex and periods differ between species based on life history traits, with timings an approximation.
23. Bio-seasons used are:

• Breeding season: birds are strongly associated with a nest site, including nesting, egg-laying and provisioning young.
• Non-breeding season: period of time where no breeding takes place, which may encompass birds over-wintering in an area and migration periods between breeding and wintering sites, dependent on the species.
  24. The bio-seasons based on NatureScot (2020) identified for each species are summarised in Table 3.1. However, the use of NatureScot non-breeding season definitions presents issues for non-breeding season apportioning (Technical Appendix 11.5: Ornithology Apportioning Technical Report). Since non-breeding season apportioning is reliant on information for Biologically Defined Minimum Population Scales (BDMPS) (Furness, 2015), mean seasonal peaks and displacement mortality was also estimated for the non-breeding seasons defined in Furness (2015) for those species where the autumn and spring passage and winter periods are defined within the non-breeding season (gannet, kittiwake and razorbill). This was conducted for the Berwick Bank Development Array plus a 2 km buffer only; as only the 2km assessment informs the apportioning analysis. These outputs are reported in section 4.1 and further used within the Technical Appendix 11.5: Ornithology Apportioning Technical Report. Non-breeding displacement for these three species within the Berwick Bank Development Array plus a 2 km buffer, as defined by NatureScot (2020), are presented in Annex F for reference.

Table 3.1: Species-specific breeding and non-breeding seasons based on NatureScot guidance (2020) and Furness (2015). Start and end months are inclusive unless stated otherwise.

Mean Seasonal Peak Population Estimates

25. As per the joint SNCB interim guidance (SCNBs, 2017), assessment of displacement impacts were conducted on the mean seasonal peak (MSP) population estimates, calculated as the peak count for each species in each appropriate bio-season, and then taken as an average over two years of surveying (March 2019 – March 2021). For example, the MSP population estimate for the breeding season was calculated as the average of the peak count in the breeding season in year one and the peak count in the breeding season in year two.
26. For all estimates, unidentified birds recorded in a category (e.g., large auk) have been apportioned to species based on the relative abundance ratios of identified species within the category (e.g. guillemot and razorbill). For the three auk species (guillemot, razorbill and puffin), the estimates were also adjusted for availability bias to account for birds likely to be diving at the time of survey. A full description of survey methodology and how monthly population estimates were calculated and apportioned for non-ID species groups, can be found in section 3 of Technical Appendix 11:1: Ornithology Baseline Technical Report.
27. For seasons starting or ending halfway through the month, the 15/16 was used as a mid-month cut off. Surveys were assigned to a season based on the day that the survey was flown. This approach avoids duplicative use of a single monthly estimate which could artificially inflate seasonal abundance estimates.
28. To account for missed and later rescheduled flights during the survey programme, some flights were assigned to different months or years to ensure even coverage of seasons in both years (Table 3.3). The Applicant discussed this allocation during the Ornithology Road Map process (RM4, Technical Appendix 11.8) and followed subsequent joint advice from Marine Scotland and NatureScot received through email 14 January 2022. Further information on flight scheduling can be found in section 3.1 of Technical Appendix 11:1: Ornithology Baseline Technical Report.

 Table 3.2: Treatment of rescheduled surveys for calculation of mean-seasonal peaks (MSPs)

29. The SNCB interim guidance (SNCBs, 2017) defines displacement as affecting both birds on the water and in flight, therefore, the mean seasonal peaks were calculated from monthly population estimates for all birds present within the assessment boundaries. The monthly population estimates for each species in the Development Array (apportioned for unidentified birds), from which the mean-peaks have been calculated, can be found in Annex A. The monthly apportioned population estimates for each species in the Development Array plus 2 km buffer, from which the mean-peaks have been calculated, can be found in Annex B.
30. The MSP population estimates for each species, in each appropriate bio-season and for each of the Development array and Development array plus 2 km buffer, are presented in Table 3.3.

Table 3.3: Mean seasonal peak (MSP) population estimates for the Berwick Bank Development Array and Development Array plus a 2 km buffer. Data include “no-identification” birds apportioned to species and auk species estimates are corrected for availability bias. Seasonal peaks are presented for reference.

3.4.        Displacment and Mortality Rates used for assessment

3.4. Displacment and Mortality Rates used for assessment

31. For both displacement and mortality rates, a sub-set of the most likely species-specific rates were highlighted in each species matrix. These include the displacement and mortality rates used in both the Scoping Approach and the Developer Approach (Table 3.5).
32. A detailed justification of the displacement and mortality rates selected for the Developer approach can be found in Annex G. Displacement and mortality rates used for the Scoping Approach were as advised in the Scoping Opinion.
33. Assessment of displacement during the non-breeding season was not required for puffin, in either approach, as advised through the Scoping Opinion (4 February 2022). Lower mortality rates were advised for auk species in the non-breeding season in the Scoping Opinion, based on birds being less constrained to nest sites and no longer central-placed foragers. For the Scoping Approach, both advised mortality rates have been assessed and potential mortalities due to displacement identified in the relevant matrices. A displacement rate of 60% has been applied, as advised in the NatureScot scoping representation (December 2021).
34. For the Developer Approach, a displacement rate of 50% and mortality rate of 1% for auks was considered suitably precautionary for both the breeding and non-breeding season. APEM (2022) undertook a review of auk displacement rates, and the Developer Approach aligns with their recommended maximum rate.  
35. For the Developer Approach, the displacement and mortality rates for puffin (50% and 1% respectively) follow rates applied at the Norfolk Vanguard Offshore Wind Farm (MacArthur Green, 2019a). The displacement rate for gannet (70%) was as advised in the Scoping Opinion (4 February 2022), whilst the mortality rate for gannet (1%) was chosen on the basis of previous recommendations from Natural England at the same development (MacArthur Green, 2019b).
36. The displacement rate for kittiwake (30%) was advised in the Scoping Opinion (4 February 2022) and is consistent with previous advice on Forth and Tay Projects (Marine Scotland, 2017); this has been applied in both the Scoping and Developer Approach. However, the Developer Approach applies a single mortality rate of 2%, which is within the range advised under the Scoping Approach (1-3%). The mortality rate of 2% follows previous advice from the Marine Scotland on previous Forth & Tay projects (Marine Scotland, 2017).
37. In addition, the Applicant has further applied the Matrix method to auks using a more nuanced approach; with lower displacement rates applied to the 2km buffer than the Development Array. The relevant methods and results are presented in Annex C and are discussed within Annex G.  

Table 3.4: Displacement and mortality rates used for the Scoping Approach (Scoping Opinion 4 February 2022) and the Developer Approach.

1 Recommended maximum displacement rate from APEM (2022). Review of evidence to support auk displacement and mortality rates in relation to offshore wind farms. APEM Scientific Report P00007416. Ørsted, January 2022

2 Recommended displacement rates from MacArthur Green (2019a). Norfolk Vanguard Offshore Wind Farm. The Applicant Responses to First Written Questions. Appendix 3.3 – Operational Auk and Gannet Displacement: update and clarification

3 Natural England recommended displacement and mortality rates for Gannet for Norfolk Vanguard Offshore Wind Farm.  MacArthur Green (2019b). Norfolk Vanguard Offshore Wind Farm Offshore Ornithology Assessment Update for Deadline 6.

4 Based on MS Scoping Opinion for Forth & Tay projects (2017)

4.             Results

4. Results

4.1.        Displacement Matrices

4.1. Displacement Matrices

38. The displacement matrices provide, for each species and relevant bio-season, the estimated mortality of birds predicted to occur due to displacement, as determined by the relevant specified rates of displacement and mortality (Table 3.4).
39. Displacement matrices for each species, in each bio-season and in both the Berwick Bank Development Array and the Berwick Bank Development Array plus a 2 km buffer are presented in Table 4.1 to Table 4.22.
40. Each cell presents potential bird mortality following displacement from the proposed Berwick Bank OWF during a bio-season given; i) the seasonal mean peak population within the impacted area; ii) the percentage assumed to be displaced from the impacted area; and iii) the assumed percentage mortality amongst the displaced birds. The outputs highlighted in colour are those based on the displacement and mortality rates in Table 3.4 and are deemed the ‘most realistic’ mortality estimates as advised by the: i) Scoping Opinion (highlighted in dark teal) and ii) the Developer Approach (highlighted in orange). Outputs highlighted in light teal reflect potential uncertainty associated with the selected figures, as advised by the SNCB (2017) guidance with regards to presenting the matrices. No adjustments for age classes of birds have been made.

Kittiwake

Table 4.1: Potential kittiwake mortality following displacement from the Berwick Bank Development Array in the breeding season. Estimates considered, in light of empirical evidence, to represent the most realistic scenarios are colour coded, with dark teal representing rates advised by the Scoping Opinion; orange representing rates defined by the Developer Approach; and light teal representing uncertainty around these figures.

Table 4.2: Potential kittiwake mortality following displacement from the Berwick Bank Development Array plus 2 km buffer in the breeding season. Estimates considered, in light of empirical evidence, to represent the most realistic scenarios are colour coded, with dark teal representing rates advised by the Scoping Opinion; orange representing rates defined by the Developer Approach; and light teal representing uncertainty around these figures.

Table 4.3: Potential kittiwake mortality following displacement from the Berwick Bank Development Array in the non-breeding season. Estimates considered, in light of empirical evidence, to represent the most realistic scenarios are colour coded, with dark teal representing rates advised by the Scoping Opinion and light teal representing uncertainty around these figures. A quantitative assessment is not being made for kittiwake in the non-breeding season under the Developer Approach (see Annex G for justification).

Table 4.4: Potential kittiwake mortality following displacement from the Berwick Bank Development Array plus 2km buffer in the spring migration (non-breeding) period. Estimates considered, in light of empirical evidence, to represent the most realistic scenarios are colour coded, with dark teal representing rates advised by the Scoping Opinion and light teal representing uncertainty around these figures. A quantitative assessment is not being made for kittiwake in the non-breeding season under the Developer Approach (see Annex G for justification).

Table 4.5: Potential kittiwake mortality following displacement from the Berwick Bank Development Array plus 2km buffer in the autumn migration (non-breeding) period. Estimates considered, in light of empirical evidence, to represent the most realistic scenarios are colour coded, with dark teal representing rates advised by the Scoping Opinion and light teal representing uncertainty around these figures. A quantitative assessment is not being made for kittiwake in the non-breeding season under the Developer Approach (see Annex G for justification).

Guillemot

Table 4.6: Potential guillemot mortality following displacement from the Berwick Bank Development Array in the breeding season. Estimates considered, in light of empirical evidence, to represent the most realistic scenarios are colour coded, with dark teal representing rates advised by the Scoping Opinion; orange representing rates defined by the Developer Approach; and light teal representing uncertainty around these figures.

Table 4.7: Potential guillemot mortality following displacement from the Berwick Bank Development Array plus 2 km buffer in the breeding season. Estimates considered, in light of empirical evidence, to represent the most realistic scenarios are colour coded, with dark teal representing rates advised by the Scoping Opinion; orange representing rates defined by the Developer Approach; and light teal representing uncertainty around these figures.

Table 4.8: Potential guillemot mortality following displacement from the Berwick Bank Development Array in the non-breeding season. Estimates considered, in light of empirical evidence, to represent the most realistic scenarios are colour coded, with dark teal representing rates advised by the Scoping Opinion; orange representing rates defined by the Developer Approach; and light teal representing uncertainty around these figures.

Table 4.9: Potential guillemot mortality following displacement from the Berwick Bank Development Array plus 2 km buffer in the non-breeding season. Estimates considered, in light of empirical evidence, to represent the most realistic scenarios are colour coded, with dark teal representing rates advised by the Scoping Opinion; orange representing rates defined by the Developer Approach; and light teal representing uncertainty around these figures.

Razorbill

Table 4.10: Potential razorbill mortality following displacement from the Berwick Bank Development Array in the breeding season. Estimates considered, in light of empirical evidence, to represent the most realistic scenarios are colour coded, with dark teal representing rates advised by the Scoping Opinion; orange representing rates defined by the Developer Approach; and light teal representing uncertainty around these figures.

Table 4.11: Potential razorbill mortality following displacement from the Berwick Bank Development Array plus 2 km buffer in the breeding season. Estimates considered, in light of empirical evidence, to represent the most realistic scenarios are colour coded, with dark teal representing rates advised by the Scoping Opinion; orange representing rates defined by the Developer Approach; and light teal representing uncertainty around these figures.