Abstract
Bird collisions with anthropogenic objects are well documented in the literature, including those involving wind turbines. The purpose of this study was to evaluate and help improve the effectiveness of an automated detection and deterrent system designed to minimize the risk of raptors colliding with wind turbines. We evaluated the DTBird® system (Liquen Consultoría Ambiental, S.L., Madrid, Spain), which is designed to detect and deter raptors flying near and in the risk zone of wind turbines. The DTBird system includes a camera/video-based detection module that detects and tracks objects based on settings calibrated for birds with specific wingspans, and a collision-avoidance or deterrence module that emits sounds designed to discourage birds from proceeding into the collision risk zone of an operational turbine. The deterrence module first emits an audible warning signal when the surveillance system estimates that a detected flying object (whether a bird or an inanimate object) has crossed a calibrated distance threshold. If the surveillance system estimates that the tracked object crosses a second, closer distance threshold, then it emits a stronger dissuasion signal intended to scare the bird away from the signal noise and turbine.
The ultimate goal of the study was to quantify the effectiveness of the DTBird system as a measure to reduce collision risk for golden eagles (Aquila chrysaetos) and other large raptors. If found to be effective, and if accepted by the U.S. Fish and Wildlife Service (USFWS), the DTBird system or other risk-reduction technologies could be considered for use by commercial wind energy facilities in eagle conservation plans as a Best Management Practice (BMP) under the Eagle Rule, a minimization measure for take permits or habitat conservation plans, or as an adaptive management measure in a Bird and Bat Conservation Strategy. Determining whether DTBird is suitable for use in eagle conservation plans or other minimization measures was beyond the scope of this study.
As the first in-depth study of DTBird with raptors in the United States, this study estimated detection and deterrence of eagles and buteos, and identified several important limitations of the technology and the study design for evaluating detection and deterrence of the target raptor species. Limitations included a large number of false-positive detections (i.e., detections that were not large raptors), unclear deterrent responses from in situ eagles and other raptors, potential bias from use of eagle-like unmanned aerial vehicles (UAVs) as surrogates for live eagles, and detection degradation from sun glare, clouds, and visual clutter. These limitations suggest future studies that could build on our initial findings.
The study was hosted by the Manzana Wind Power Project, which is owned and operated by Avangrid Renewables and located in Kern County, California. Over the course of a 9-month study period from December 2016 through August 2017, we used fixed-wing UAVs as surrogates for live eagles in experimental flight trials to evaluate the performance of the DTBird detection and deterrent-triggering systems installed at this facility. The UAVs used for the study were similar in size and painted to resemble a golden eagle, and carried onboard avionics that provided high temporal and spatial resolution Geographic Positioning System (GPS) tracking data. We also evaluated the effectiveness of the DTBird deterrence module by examining the behavioral responses of in situ raptors evident in videos recorded by the seven DTBird systems installed at the facility. We derived estimates of the probability of detection from the UAV flight trials as a surrogate for live eagles and estimates of the probability of deterrence from classifying the responses of in situ raptors. We then estimated the probability of collision-risk reduction from deploying DTBird as the cross-products of the estimated probabilities of detection and deterrence.