Recent studies assessed the effects of WFs (wind farms) on terrestrial mammals, but the response of roe deer Capreolus capreolus to this infrastructure remains unclear. We aimed to investigate whether the spatial distribution of roe deer differs between operational wind farm area and nearby control area during summer, using a UAV-based approach that combined thermal infrared and RGB imaging. We used a DJI Matrice 300 RTK, equipped with Zenmuse H20T camera, to mark locations of roe deer in real-time during evening sessions. We performed a Nearest Neighbor Analysis and Kernel Density Estimation in QGIS. We also performed a logistic regression model of roe deer presence points and random points to verify landscape elements of roe deer distribution. Both in the WF area and in the control area, roe deer showed clustering; however, roe deer in the WF area showed this effect more strongly. In the control area, the area of KDE 50 was about 50% larger than in the WF area. The spatial distribution of roe deer in the WF area depended on distance to settlements, roads, enclaves, and forests, but for the control area, the logistic model was insignificant. The modified density of roe deer in the WF area may have multiple consequences for both the reproduction process and the overall condition of animals. Increased ambient noise could impair not only predator detection but also vocal communication between individuals, especially relevant during territorial marking. While our study did not demonstrate clear avoidance of wind turbines by roe deer, it did reveal varying distributions of this species, which may indicate an indirect effect of the wind farm.
GMAJ et al. (Sun,) studied this question.