This report presents an estimate of the risk of collision between harbour seals (Phoca vitulina) and tidal turbines on the basis of observed behaviour patterns derived from targeted telemetry tracking studies and recent population survey data. The collision risk associated with a proposed turbine array development in the Pentland Firth was used as a worked example of the method.
A brief summary of the movement data collected during transmitter deployments on harbour seals in the Pentland Firth study area in 2011 is presented, with an emphasis on information that is most likely to be of use in assessing the potential impacts of tidal turbine deployments.
A process to estimate the number of times that telemetry tagged seals could pass through the area of a proposed turbine array is described. The method incorporates dive depth data to estimate the number of times tagged seals would have passed through the swept area of individual turbines in a hypothetical turbine array within the site.
Telemetry data from seals tagged at various sites in the Pentland Firth and Orkney were used to support an estimate of the number of seals likely to be at risk of interacting with such an array. The expected number of collisions between harbour seals and turbines in the array was based on details of their movements relative to the locations and movements of hypothetical tidal turbine blades, assuming that seals are oblivious to the presence of devices, i.e. show no avoidance and take no evasive action.
- It was estimated that 1.2 seals per year (approx. 95% confidence interval (C.I.) 0.8- 2.0) would collide with individual turbines.
- Scaling that estimate up to the full 86 turbine array suggested that 103 (approx. 95% C.I. 73 – 152) collisions would occur per year.
- This estimate was approximately 15% of the ra te derived, previously, from collision risk models for the same site.
The preliminary information from collision ex periments was then used to give estimated mortality rates from such collisions. The effects of interpolation error due to timing of GPS fixes and the small, but significant, GPS position error on the estimates of interaction rates are discussed and an attempt is made to incorporate estimates of variability from other sources such as population estimates. Areas where such error distributions are poorly understood and require additional research are highlighted.