Abstract
As the wind industry continues to grow exponentially, an increasing number of studies are documenting bat fatalities due to collisions with operating wind turbines. One possible explanation for such high mortality rates is that bats are attracted to wind turbine sites and to the turbines themselves. Recent evidence confirms that some bats approach and alight on turbine towers and blades and also appear to forage aerially for insects within the airspace swept by the turbine rotor. We tested the first experimental ultrasonic bat deterrents designed for commercialscale wind turbines at the Maple Ridge Wind Farm in Lowville, New York, USA where bat fatalities had been reported the previous year. This facility consists of 195 Vestas 1.65 MW turbines, widely dispersed across a landscape of open agricultural lands and scattered woodlots.
The deterrents emit randomized and continuous ultrasound designed to interfere with normal echolocation in insectivorous bats. We mounted deterrents on the towers of two treatment turbines and two control turbines with similar landscape characteristics and historic mortality rates and performed two experiments in succession. For each experiment, we simultaneously observed one treatment and one control turbine nightly for 10 consecutive nights using thermal infrared imaging cameras, which can capture images in complete darkness and do not disturb normal behaviors. We monitored an area within the rotor-swept zone adjacent to the mounted deterrents nightly for 3.6 hours beginning shortly after sunset.
Overall we observed 618 occurrences of bats (and an estimated 566 bat passes) during 288 hours of video observation, yielding a rate of 4–46 passes on a given night (1.9 bats / hour). While most bats observed were engaged in normal flight, 2% avoided collisions (n = 12), 3% investigated the turbines (n = 16), and <1% collided with the turbine blades (n = 2). Twenty eight percent of bats we observed flew within the rotor swept zone (n = 158). In the first 10-night test, we observed a total of 131 bats ( x = 13.1, SD = 5.5) at the deterrent-treated turbine versus 244 bats ( x = 24.4, SD = 12.9) at the control turbine - a statistically significant difference (t = 2.54, p = 0.026). However, during the second test, there was no significant difference in bat activity between the treatment ( x = 9.5 SD = 8.3) and control ( x = 9.6, SD = 4.8) turbines (t = - 0.003, p = 0.97). We also observed 24 separate instances (n = 56, 10%) of small groups of bats (2–5 individuals) flying together around turbines, which suggests that the timing of migration flights may be an important factor in bat fatalities at this and at similar wind facilities. Wind speed was positively related to bat passes observed (R2 = 0.23, p = 0.01) whereas barometric pressure was a negative predictor (R2 = 0.33, p = 0.002). Temperature, humidity, rotor speed, and cloud cover were all non-significant predictors of bat passes. A multivariate regression analysis showed a significant relationship between two wind measurements, barometric pressure, and the presence or absence of the deterrent (F = 3.87, R2 = 0.424, p = 0.02).
Our mixed results suggest that a variety of factors influence the effectiveness of an acoustic deterrent. The acoustic envelope of our deterrent system was probably not large enough to consistently deter the activity of bats within the large volume of the rotor-swept zone. For deterrents to be effective, they must operate at ranges that are large enough to encompass an entire turbine structure. Future studies must also examine the assumptions behind acoustic deterrence. Although bats are known to avoid ultrasound clutter, little is known about the behavioral responses of bats to artificial broadband ultrasound emissions. It must be demonstrated on a full-size scale that bats both can and will avoid large ultrasound fields before acoustic deterrent systems can be expected to function effectively at wind farms.