Posters

Abstract

Bird collisions remain a key ecological constraint for wind-power expansion. Automated detection–shutdown (shutdown-on-demand) systems promise to mitigate risk, yet quantitative multi-species evidence from operational settings is limited. Between 2021 and 2024, three turbines in central Spain were equipped with detection–reaction systems to monitor raptor interactions. Over 10 400 detections of six genera (Milvus, Gyps, Aquila, Falco, Accipiter, Circus) were analysed in R (v4.3.3) using non-parametric tests (Kruskal–Wallis, Wilcoxon, χ²), Spearman rank correlations and Holm adjustments. Metrics included minimum avoidance distance, frequency of STOP-signal activation, and the effect of wind speed on approach behaviour. Avoidance varied significantly among taxa (χ² = 1065, p < 0.001). Median distances were Gyps 131 m (N = 4 340), Milvus 88 m (N = 5 767), Aquila 78 m (N = 230), Falco 75 m (N = 1 189). Gyps fulvus maintained significantly larger buffers (p < 0.01), whereas Milvus milvus entered the ≤61 m rotor zone in 26 % of detections, indicating >2.5-fold higher collision exposure. STOP signals were triggered for Aquila chrysaetos 77.5 %, A. adalberti 73 %, Milvus migrans 65.5 %, M. milvus 57.4 % and Gyps fulvus 52.4 %. Wind strongly modulated approach patterns. Calm conditions (≤2 m s⁻¹) produced frequent but relatively safe approaches (M. milvus median 88 m, N = 1 109; G. fulvus 124 m, N = 641). High winds (>12 m s⁻¹) reduced activity but compressed avoidance margins (M. milvus 72 m, G. fulvus 104 m). Wind–distance correlations were weak (ρ ≤ 0.08) except moderate positive for A. chrysaetos (ρ = 0.45). Findings reveal pronounced species-specific strategies: large soaring vultures sustain wider buffers, while agile kites often approach rotor zones. Results confirm the operational maturity (TRL ≥4) of detection–shutdown systems and underscore the need for adaptive thresholds—by species and wind regime—to optimise mitigation while minimising unnecessary downtime.