Royal Society for the Protection of Birds (RSPB) (Royal Society for the Protection of Birds (RSPB))
"RSPB Relevant Representation Regarding Norfolk Vanguard Offshore Windfarm
The RSPB supports the deployment of renewable energy projects, providing that they are sited in appropriate places and designed to avoid potential adverse impacts on wildlife. We are grateful for the constructive pre-application discussions that have taken place with Vattenfall in respect of this proposal, particularly through the Evidence Plan process. Although some progress was made during this process, the RSPB remains concerned that some methodological procedures used in the assessment are inadequate to ensure a robust assessment and therefore a proper understanding of the likely impacts of the scheme.
2. Offshore ornithology impacts
We have significant concerns regarding the findings of some of the impact assessments. As a result of the methodological concerns (set out below) and our own recalculation of collision risk using the Marine Scotland stochastic model, the RSPB considers that the impacts have not been adequately assessed and, as such consider that an adverse effect on the integrity (AEOI) of the following SPAs/pSPAs and their species cannot be ruled out for the following ornithological interests:
• The impact of collision mortality on the kittiwake population of the Flamborough Head and Bempton Cliffs SPA and the Flamborough and Filey Coast pSPA (FHBC SPA/FFC pSPA) alone and in-combination with other plans and projects;
• The impact of collision mortality on the gannet population of the FHBC SPA/FFC pSPA alone and in-combination with other plans and projects; and
• The impact of collision mortality on the lesser black-backed gull population of the Alde-Ore Estuary SPA alone and in-combination with other projects.
We also consider that potential for likely significant effects on the following site exists, and that this site cannot be screened out from further assessment at this stage:
• Bruine Bank pSPA (Netherlands) – displacement of razorbills and guillemots, particularly during migratory periods.
In addition, we consider that insufficient evidence has been provided to rule out potential significant impacts on the following North Sea populations:
• Cumulative collision mortality to North Sea populations of kittiwake and great black-backed gull; and
• Cumulative operational displacement to North Sea populations of auks (guillemot, razorbill and puffin) and red-throated diver.
3. Concerns regarding the assessment of collision risk
Our major concerns regarding the assessment process and evidence base are outlined below.
(a) Potential Biological Removal
Potential Biological Removal (PBR) is used in justification of conclusions of no AEOI for both gannet and kittiwake. The RSPB disagree with the use of PBR in this context following the publication of the review by O’Brien et al. (2017) and the RSPB Practitioner’s Perspective (Green et al., 2016), and therefore support NE’s position that PBR does not provide an appropriate threshold for this purpose (as outlined in the Secretary of State’s HRA for the Hornsea Project 2 offshore windfarm). PBR was designed to manage whaling quotas by detecting unsustainable mortality in a population leading to risk of its extinction, whereas SPAs are set up to maintain or restore a population of conservation importance. Levels of acceptable mortality derived from PBR are therefore likely to be higher than those acceptable for a population to continue to meet the conservation objectives of a SPA. Where population modelling is required to inform an assessment, this should be based on Population Viability Analysis (PVA). PVA enables comparison of the change in population size with and without the project after several years, thereby presenting an indication of the magnitude of change attributable to the proposal and is therefore more suitable for assessing the effects of a project on a SPA.
(b) Stochastic Collision Risk Model (CRM)
In order to predict the collision risk mortality of an offshore wind farm in the UK, the Band (2012) model has previously been used in assessment. This model uses a number of input parameters, such as bird size, flight speed and turbine blade dimensions, to calculate the probability of a bird that passes through the swept area of a turbine blade colliding with that blade. For this deterministic model the input parameters were defined as single values with no indication of variability around them. In reality, most of the parameters will exhibit a considerable degree of variability and stochastic collision risk modelling has been developed to allow this to be incorporated into the model and thus generate a potential range of output predicted collision mortalities. MacGregor et al., (2018), under commission of Marine Scotland Science and overseen by an expert steering panel, produced a revised and fully tested stochastic model to widespread stakeholder acceptance. By contrast, the Applicant has presented an entirely untested new version that does not follow a recognised methodology, with insufficient detail provided as to how it incorporates variability or how it overcomes the statistical difficulties of non-independence (the degree of interrelation) of some of the variables. The RSPB therefore does not agree that the model presented by the Applicant is fit for purpose and recommend that the Marine Scotland (MacGregor et al., 2018) model version is used in preference.
(c) Use of median bird densities within the CRM
The Applicant has presented deterministic and stochastic versions of the CRM (see above). For the deterministic version (Band 2012) of the CRM the correct value to use for bird density is the mean monthly value. As detailed in Environmental Statement Appendix 13.1 Ornithology Technical Appendix Annex 3 [APP-217], the values used by the applicant appear to be median values, which will result in the model predicting considerably lower collision mortalities.
(d) Nocturnal Activity Factor
We do not agree with the changes in Nocturnal Activity Factor (a parameter used in collision risk modelling) proposed. The value presented for kittiwake is based on unpublished evidence which does not appear to form part of the examination documentation and therefore we are unable to assess the robustness of the study. The current factor is derived from the expert opinion collected by Garthe and Huppop (2004) and this use is endorsed by Band (2012). A review of seabird vulnerability to offshore wind farms (Furness et al., 2013) recommended that no changes be made to the nocturnal activity scores for these species, and an update, including the same authors (Wade et al., 2016) maintained this recommendation.
It is also not clear how these revised rates account for the distinction between the definition of daylight as used in the Band model and with the official concept of ‘twilight’ and ‘night’. This is an issue as the Band (2012) model considers the nocturnal period as between sunset to sunrise and so treats flight activity that occurs at twilight as being within the nocturnal flight period. Evidence from tagging shows that an important number of seabirds actively forage at twilight.
While we welcome the latest published evidence review for gannet (Furness et al., 2018), we are concerned that the mortalities predicted using revised nocturnal activity rates for gannet (and this is also applicable to kittiwake) are potentially underestimated because they do not account for the potential interaction between survey timing and diurnal behavioural patterns. Peaks in foraging activity at first and last light (see for example Fig. 3 in Furness et al. 2018) will not be accounted for in the assessment if these did not coincide with surveys (the timings of which are currently unknown, but likely to be midday if aerial), and the survey may have been carried out at a time of much lower activity. Thereby the application of the revised nocturnal activity factor recommended by Furness et al., (2018) could result in inaccurate underestimates of collision risk.
The Nocturnal Activity Score presented for gannet in the application documents is also not in accordance with this latest review (Furness et al., 2018) which recommends 8% in the breeding season and 3% in the non-breeding season. The values used in the assessment, 4.3% and 2.3% respectively, will result in a prediction of fewer collisions.
(e) Underestimation of collision mortality
Based on a comparison with our own calculations using the MacGregor et al., (2018) model version with a range of nocturnal activity rates, the Applicant’s model underestimates collision mortality for key species, and it produces significantly reduced predictions compared to the deterministic model, particularly when the deterministic model is also re-run using mean (rather than median) monthly bird densities.
(f) Breeding season definitions
We have concerns about the manner in which biological seasons for gannet and kittiwake have been defined by Vattenfall’s consultants. The use of the ‘migration-free breeding season’ means that months where breeding and migration can overlap are excluded from the analysis of breeding season impacts, artificially reducing the duration of the breeding season and hence risks underestimating collision mortality of breeding birds. The definition of ‘breeding season’ as presented in Furness (2015), should therefore be used, except where colony specific evidence clearly suggests otherwise.
(g) Apportioning of mortality to SPAs
We have concerns about some of the figures used for apportioning of collision mortality to SPAs and the evidence used to support this. The estimated proportion of kittiwake from FHBC SPA/FFC pSPA used in the HRA is 16.5% and is based on no site-specific historical estimates. Notwithstanding the applicants unfounded criticisms of the FAME and STAR tracking of kittiwakes, which is dealt with below, the assessment does not take into account more recent tracking of kittiwakes from FHBC SPA/FFC pSPA carried out in 2017, using lighter tags (<3% bodyweight) and following the birds for a longer period due to a novel attachment method, that showed a high degree of overlap of colony breeding birds with the development site. We therefore do not agree with the value used for apportioning kittiwake collision mortalities to the FHBC SPA/FFC pSPA as it will considerably underestimate the actual impact.
We are concerned that the methods used for apportioning collision mortality of lesser black-backed gulls to the Alde-Ore Estuary SPA are inadequately explained, with insufficient reference to current knowledge and with a cavalier attitude toward precaution. Such calculation is difficult because of two competing factors. Throughout the UK, the urban population of lesser black-backed gulls is increasing, while those in “natural” colonies is decreasing (JNCC, 2018). In simplistic terms this could be argued as reducing the impact apportioned to the SPA. In the Applicant’s calculations of the number breeding birds within foraging range of the developments a number of inland, urban colonies are included, such as Ipswich and Norwich as likely sources of birds foraging in the development areas. While we acknowledge that there is a need for more research on the foraging behaviour of urban gulls, it is unlikely that such gulls, especially those from non-coastal urban colonies will forage in the offshore marine environment to the same extent as those breeding at coastal “natural” colonies, such as the Alde-Ore Estuary SPA. The inclusion of birds from such sites dilutes the potential significance of impact on the Alde-Ore Estuary SPA.
Furthermore in calculating the number of non-SPA birds the Applicant gives a rounded up figure of 5400 birds, then simply doubles it (and rounds up further) to 11000, with scant justification other than saying 5400 was a likely underestimate, but presenting no supporting evidence. By overstating the non-SPA population in this way, the potential impact on the Alde-Ore Estuary SPA is again significantly understated.
We therefore do not agree that these calculations provide confidence to support a conclusion of no adverse effects on integrity of the population of lesser black-backed gull from the Alde-Ore Estuary SPA and consider that a full assessment, including PVA should be carried out.
(h) Gannet avoidance rate
Whilst the RSPB accepts the SNCB’s recommended amendment to the gannet avoidance rate (AR) from 98% to 98.9% for non-breeding birds, we do not agree that this figure should be applied to the breeding season due to the lack of available evidence relating to breeding birds. In light of evidence from Cleasby et al., (2015), which demonstrated that foraging birds flew higher, and were therefore at greater risk of collision, than commuting birds, and given that the BTO avoidance rate review was heavily biased to non-breeding gannets, we prefer a more precautionary AR of 98% for the breeding season.
(i) Criticisms of kittiwake tracking data
The Applicant raises a number of issues with regard to the suitability of tracking data obtained as part of the FAME and STAR projects for use in the assessment. However the Applicant’s report contains a number of misinterpretations and erroneous assertions. In particular:
• It is claimed that the longest foraging trips from FAME/STAR kittiwake data were largely from colonies where the breeding success was zero or close to zero. This is incorrect. The longest trips were recorded from Flamborough and Filey, where breeding success was comparatively high over the time of tracking
• The claim that tagged birds were more likely to have failed is also incorrect. For the FAME and STAR data, where remote download tags were used, birds had to be re-caught when on the nest so were required to be successful, at least up until the point of recapture, in order to obtain the data.
• Tagging conducted in 2017 used tags that were less than 2.5 % of the birds’ body weight and observed longer foraging ranges with multiple actively breeding birds visiting the Norfolk Vanguard site.
• The applicant repeatedly asserts that the RSPB data are not available. This is simply not true, they are available upon formal request, and the consultants who authored this section have had data made available to them under such a request.
(j) Potential for mitigation of impacts on the Alde-Ore Estuary SPA
The RSPB are concerned at the Applicant’s interest in mitigating impacts on lesser black-backed gull through predator management at the Alde-Ore Estuary SPA. Mitigation measures need to counter or at least lessen effects arising from the development proposals and must be over and above required management measures for the SPA. Due to the uncertainty around the relative importance of the various factors affecting this population, and therefore the likely effectiveness of the proposed management measures on productivity as well as no evidence that they will be over and above required management for the site, we do not agree that measures of this sort should be considered as mitigation for SPA impacts.
(k) Screening out of Bruine Bank pSPA
The Bruine Bank pSPA in Dutch waters is approximately 20km from Norfolk Vanguard at the closest point and is of importance for wintering razorbill and guillemot. Although the site was screened, we do not consider that sufficient evidence has been provided to screen this pSPA out from further assessment, particularly in relation to displacement of birds during the migration period.
The RSPB reserves the right to add to and/or amend its position in light of changes to or any new information submitted by the Applicant.
Band, W. (2012) Using a collision risk model to assess bird collision risks for offshore windfarms. SOSS02 Project Report to The Crown Estate.
Cleasby, IR, Wakefield, ED, Bearhop, S, Bodey, T W, Votier, SC and Hamer, KC (2015), Three-dimensional tracking of a wide-ranging marine predator: flight heights and vulnerability to offshore wind farms. J Appl Ecol, 52: 1474–1482. [redacted]
Furness, RW (2015) Non-breeding season populations of seabirds in UK waters: Population sizes for Biologically Defined Minimum Population Scales (BDMPS). Natural England Commissioned Reports, Number 164
Furness, R, Wade, H & Masden, E (2013). Assessing vulnerability of seabird populations to offshore wind farms. Journal of environmental management. 119C. 56-66. 10.1016/j.jenvman.2013.01.025.
Furness, R, Garthe, S, Trinder, M, Matthiopoulos, J, Wanless, S, Jeglinski, J (2018) Nocturnal flight activity of northern gannets Morus bassanus and implications for modelling collision risk at offshore wind farms, Environmental Impact Assessment Review, 73,1-6, [redacted]
Garthe, S. & Hüppop, O. (2004) Scaling possible adverse effects of marine wind farms on seabirds: developing and applying a vulnerability index. Journal of Applied Ecology 41: 724-734.
Green, R. E., Langston, R. H. W., McCluskie, A., Sutherland, R. and Wilson, J. D. (2016), Lack of sound science in assessing wind farm impacts on seabirds. J Appl Ecol. doi:10.1111/1365-2664.12731
Marine Scotland. (2018) A Stochastic Collision Risk Model for Seabirds in Flight. Marine Scotland Topic Sheet, 2pp. DOI: 10.7489/12101-1
JNCC (2018) Latest population trends: lesser black-backed gull. [redacted]
O'Brien, S. H., Cook, A. S., & Robinson, R. A. (2017). Implicit assumptions underlying simple harvest models of marine bird populations can mislead environmental management decisions. Journal of environmental management, 201, 163-171.
Wade, H, Masden, E, Jackson, A.C. & Furness, R. (2016). Incorporating data uncertainty when estimating potential vulnerability of Scottish seabirds to marine renewable energy developments. Marine Policy. 70. 108-113. 10.1016/j.marpol.2016.04.045.