- One of the potential impacts on birds from offshore wind farms is the mortality associated with collisions with turbine blades.
- There have been considerable advances in the development of statistical techniques to estimate potential collision-related mortality. However, there are still significant gaps in knowledge regarding the flight heights and avoidance rates of seabirds in relation to offshore wind farms, two key parameters in collision risk modelling.
- This report reviews current information on the flight heights and avoidance rates of key seabird species that occur in UK waters and that are thought most susceptible to effects of collisions with offshore wind farms and which typically may need to be considered in EIAs. Recommendations are provided on the use of this information and where further work is needed. Updated guidance on the use of the Band et al. (2007) collision risk model and a revised spread-sheet for use in relation to offshore wind farms is provided separately (Band 2012).
- An extensive literature search was undertaken to investigate the flight heights and avoidance rates of seabirds in relation to offshore wind farms. In total data from 40 surveys of 32 existing, proposed or consented offshore wind farms were identified.
- The mean proportion of birds predicted to fly at the generic collision risk height window of 20 to 150 m above sea-level varied from 0.03 % for the Little Auk to 33.1 % for the Great Black-backed Gull. For some species, notably divers, auks and sea duck, few individuals were predicted to fly at heights which placed them at risk of collision with wind turbines and there was relatively little variability in this finding between sites.
- For collision risk modelling, it is recommended that consideration should be given to results using both the site-specific and the modelled flight height data presented here. Where there is a clear difference between data recorded on a site-specific basis and the modelled data, the reasons for this – for example, that large numbers of migrating birds pass through the site – should be explored and clearly stated. Where there is good reason to have low confidence in the quality of the site-specific data, for example that it is based on low sample sizes or was collected during unrepresentative periods, the modelled flight height data might be considered more representative.
- The updated guidance and a revised spread-sheet for offshore use of the Band collision risk model (Band 2012) that accompanies this review provides the means for estimating collision risk (i) using site-specific data and assuming a uniform bird density in the risk window; (ii) also using the uniform density model, but using a figure for the proportion of birds at risk height derived from generic data; and (iii) using data on flight height distributions, as produced by the modelling presented here.
- Whilst existing survey methodology produces estimates of the proportion of birds within fixed flight height bands, the modelled data provide the opportunity for investigating flight height bands of differed sizes and extents. As a result these models make it possible to consider how alterations to turbine hub height and the size of turbine blades may affect the collision rate. In these circumstances, to ensure comparative values are presented, only the modelled data presented here should be used. Results using both the upper and lower confidence limits from the flight height distribution should also be presented when using the modelled data.
- Bird data were collected in relation to the sea-level at the time of the survey. However, as sea-level will vary in relation to the fixed turbine structure, in the collision risk modelling process, flight heights should be considered in relation to mean sea-level. The models presented here do, however, offer the possibility of modelling the proportion of birds at collision risk height in relation to a range of sea-levels. Consideration of sea-level is also provided in the updated guidance for offshore use of the Band collision risk model (Band 2012).
- Scientific studies of avian interactions with wind farms have tended to focus on collision and mortality rates rather than actual avoidance rates. Whilst collision and mortality rates may provide a surrogate for avoidance rates, they do not necessarily reflect true avoidance rates i.e. the inverse of the ratio of the number of collisions to the number of collisions that would be predicted in the absence of avoidance behaviour (Band et al. 2007).
- Avoidance behaviour varies in response to distance from turbines and it is important to distinguish between macro-avoidance of the whole wind farm, and micro-avoidance of individual turbines within a wind farm. However, studies of avoidance have varied in their approaches, in particular, in the distances at which avoidance is measured and the avoidance rates reported are not strictly comparable. While some notable studies have taken place recently, without replication from additional sites, there is not a robust enough evidence base to suggest that existing guidance should be changed. The current, limited evidence suggests that avoidance rates may be likely to be more than 99% for some species (divers, Northern Gannet, sea ducks and auks). However, a value of 98 %, as recommended by SNH (2010), should be used as a precautionary avoidance rate until further evidence is available to build on that presented in this review. Given that there is potential for species to show higher rates, and also because of the uncertainty surrounding avoidance rates, it is recommended that collisions estimates associated with avoidance rates of 95 %, 99 % and 99.5 % should also be presented. However, these values should not take precedence in situations where strong evidence points to alternative avoidance rates. In the future, these recommendations may be refined as additional information becomes available.
- There is an urgent need for further research into the flight heights and avoidance rates of seabirds in relation to offshore wind farms. Ideally, this would include direct measurements of these variables through the tagging of individual birds and the monitoring of movements at a broader scale through the use of technologies such as radar, as well as through visual observations.