Abstract
This report reviews collision monitoring data collected at Vattenfall’s European Offshore Wind Deployment Centre (EOWDC) in Aberdeen Bay between June 2023 and December 2024. The study aimed to quantify bird interactions with offshore wind turbines using a mono-camera system developed by Spoor AI, focusing on turbine AW05 from a camera mounted on AW10. Objectives included summarizing observed bird activity, assessing potential collision events, and comparing findings with pre-construction collision risk estimates.
Over the 19-month monitoring period, the camera system operated reliably, capturing approximately 95% of daylight hours. A total of 2,007 bird tracks were recorded, with activity peaking during the breeding season (June–July) and declining in winter. Birds were typically detected at mid-range distances (250–700 m) and exhibited complex flight paths, with mean speeds of 14.9 m/s and average tracking durations of 26 seconds. These behaviors suggest that individual birds may remain in the wind farm for longer than is assumed in standard collision models, but that exposure at a population level may be lower.
Five potential collision events were flagged but, upon detailed review, none were confirmed. Most involved birds at considerable distances from turbines or exhibiting natural behaviors such as diving. Collision risk modeling based on observed activity predicted fewer than one collision at AW05 over the entire study period (0.002 collisions), aligning with the absence of confirmed events. This contrasts sharply with conservative pre-construction estimates of up to 8.54 collisions per turbine per year under high avoidance assumptions.
Complementary studies at EOWDC, including radar and GPS tracking, corroborate high levels of meso-avoidance behavior, reducing collision risk. Findings emphasize that collisions are rare events and highlight the value of stereoscopic camera systems for improving positional accuracy and reducing false positives.