Abstract
Northern gannet (Morus bassanus) (hereafter ‘gannet’) are susceptible to numerous anthropogenic pressures introduced by the construction of offshore wind farms (OWFs). Impacts include mortality from collision with turbine blades and habitat loss resulting from displacement from the OWFs’ footprint and out to some distance beyond it (the ‘buffer’). Both impacts have direct implications for mortality of individual birds and could have population level consequences. Stakeholders aim to predict these consequences by way of modelling exercises during Environmental Impact Assessment (EIA) and Habitats Regulations Assessment (HRA) processes. However, there is a need to fully understand the most appropriate ways of integrating gannet avoidance behaviour into the assessment process to support more robust assessment, decision making and management strategies.
In 2022, Natural England (NE) commissioned HiDef Aerial Surveying Ltd. (‘HiDef’) to explore specific issues of macro-avoidance and displacement associated with gannets and OWFs. On consultation with NE, the work was re-focused on macroavoidance only. The aim was to deliver an evidence-based method to ensure macro-avoidance behaviour is appropriately accounted for in collision risk models of gannet at OWFs. This should provide more realistic predictions of the number of birds at risk of collision for EIA and HRA.
The literature review identified that the two types of macro-avoidance (barrier effects and displacement) could not be disentangled from existing studies. For the purposes of this report, macro-avoidance is defined as ‘the fraction of birds in flight that are unlikely to enter the turbine array following construction, where there is a risk of collision with rotating blades’.
In the available literature for gannet, nine studies report macro-avoidance rates for at least ten OWFs that could be used in collision risk modelling. These values ranged from 0.617 to 1.000 and were determined using a mixture of survey methodologies (e.g., horizontal and/or vertical radar, GPS tagging studies, visual, boat-based, aerial surveys and before/after comparisons of densities), and for several very different wind farm sites. Calculating a robust, overall gannet macro avoidance rate is therefore, challenging.
Using a quality scoring system for studies, an overall macro-avoidance rate for gannet was calculated using a weighted mean (0.8330 [95% CI 0.4410 – 0.9959]) and unweighted mean approach (0.8564 [95% CI 0.5349 – 0.9736]). In the weighted mean approach, quality scores and the reported macro-avoidance rates themselves were utilised as weights which incorporated study quality as well as some level of precaution. However, upon discussion with the project steering group, consideration of available approaches, consultation of published literature and expert opinion, it was concluded that a macro-avoidance rate for gannet should be calculated based on a simple mean approach. Nevertheless, the role of individual-based models needs to be fully investigated as an alternative for deriving macro-avoidance rates.
To incorporate macro-avoidance into collision risk modelling, it was recommended that the input densities are corrected by the pre-determined or calculated macroavoidance rates, and a ‘within wind farm’ avoidance rate is then applied in the collision risk model. This would involve very little effort in terms of the tools currently available (e.g., stochastic collision risk model). In this way, temporal effects (i.e., differences in macro-avoidance throughout a year) could be incorporated as well.