Bat-ter Stay Away: An Innovative Approach for Deterring Bats Near a Simulated Wind Turbine

[Transcript]

Hi, I’m Arsiema! Did you know that in 2023 over 1 million bats died in North America alone due to wind turbines? You see, bats are so crucial to our ecosystems as they eat crop-damaging insects, reducing the need for pesticides. Current deterrent methods that keep these bats away from turbines, such as ultrasonic noise, often aren't a permanent solution.

I tried a new approach by testing a robotic bat-motion deterrent, which has never been tested individually to deter bats from turbines. I compared it to broadband-like ultrasonic noise (which is a version of ultrasonic noise) by observing how each affected a sensor mimicking bat echolocation in a simulated wind turbine environment. My data showed that motion was more effective between the two deterrents at disrupting the sensor’s readings, suggesting that it could disrupt a bat's echolocation and deter them from turbines. These results indicate that motion-based deterrents are a promising direction for future research. Thank you for listening! See you at the fair!

Did you know that over 1 million bats were killed by wind turbines in North America in 2023? Bats are vital to ecosystems; they pollinate plants, disperse seeds, fertilize plants with their nutrient-rich droppings, and eat insects that damage crops (reducing the need for pesticides). Over the past decade or so, some provinces have enacted rules and guidelines regarding bat deaths caused by wind turbines. These recommend a threshold number of bat deaths per turbine, with some provinces advising curtailing turbines (shutting them down or slowing them) when bat activity is present. Despite these guidelines, too many bats are still dying, as there are now plenty more turbines than a few years ago. This project can raise awareness of this topic, potentially enabling more up-to-date rules on turbine-caused bat mortalities and more research on motion-based deterrents.

The purpose of this project is to determine how motion vs. ultrasonic noise affects the clarity of echolocation signals, and to potentially help develop motion-based bat deterrents to reduce bat mortality from wind turbines. If these deterrent methods are effective in real-world scenarios, farmers and society will benefit. Bats save North American farmers $3.7 billion per year in pesticide costs. It is known that pesticides are considered bad because they are inherently toxic, pollute the environment, and can cause serious long-term health issues in humans and animals. Fewer pesticides would be much better for society and health, lowering the risk of chronic diseases.

Background research for this project was conducted primarily using PubMed, ResearchGate, and other websites that showcase previous studies and experiments related to this topic. The websites used for research are well-known, highly reputable scientific websites.

To design the system, Arduino-compatible components were used to simulate a bat approaching a wind turbine with a deterrent in place. The model featured two sensors: one mimicking bat echolocation and another producing broadband-like ultrasonic noise with rapid, irregular pulses. It is important to note that broadband noise is much more complex than the noise this sensor produces, with it spanning multiple frequencies, whereas the sensor in this experiment operates at a fixed frequency. The first prototype used a robotic bat powered by a single servo motor, and the turbine was coated with tinfoil to increase reflectivity. Later, the bat was upgraded to two servo motors (one per wing) for more realistic motion, and the turbine was wrapped with painter’s tape to reduce reflectivity and improve sensor accuracy. The components were programmed with mainly AI-generated code, which was reviewed by a teacher. Since the sensor operates like a focused beam, alignment of all components was necessary. In real-world scenarios, alignment would be less important, as most sound propagates in three dimensions.

In the first experiment, 5 trials were run per condition (control, motion, noise, and combined—noise and motion). In the second experiment, it was increased to 15. Trials lasted 20 seconds, with sensor readings taken every second. Using the readings, the mean and sample standard deviation (for error bars) were calculated for every condition. Controlling variables included running trials on the same day, at around the same time, for the same length of time, using brand-new components, in a medium-sized space away from walls (to reduce echoes), and keeping the room fairly quiet.

Both models work by the sensor sending out a short burst of high-frequency ultrasonic sound waves to measure distance. These sound waves are emitted from the transmitter (figure 1), reflect off an object, and return to the receiver (figure 1), functioning much like echolocation used by bats. It then displays these measurements in the Serial Monitor of the Arduino app. An optimal deterrent method would exhibit substantial variability in the readings. This just means that the measurements would be much more different than the actual distance. Then—like previously stated—a mean for each condition is calculated using Excel. Then, error bars were calculated using sample standard deviation. The mean was selected to capture the full magnitude of the sensor's response, ensuring that each deterrent's impact is accounted for in the data. To compare the different deterrents fairly, sample standard deviation was used to assess the consistency, strength, and statistical reliability of these responses relative to one another.

The findings (figure 2) of the first prototype (figure 3) indicate that all three deterrent methods—motion, noise, and combined—were effective in increasing variability in simulated echolocation readings, as their error bars don’t overlap with any other conditions. The data suggests the combined deterrent was the most effective, with its error bar also not overlapping. Robotic bat motion and broadband-like ultrasonic noise were statistically the same, despite the bat motion causing more deviation.

Findings (figure 4) from the second prototype (figure 5) suggest, once more, that all deterrent methods were effective in increasing variability in the sensors' readings. The combined method had the highest variability, followed by the robotic bat motion. The main difference between the prototypes' data is that all deterrents were significantly different from each other, as none of the error bars overlap across conditions.

Initial trials showed no statistically significant difference between motion and noise as deterrents. However, after refining the design, all treatments were statistically different, with the combined treatment being most effective, followed by motion, then noise. This change was likely due to improved experimental control. The tin foil used earlier likely introduced uncontrolled variables by generating unintended noise and reflecting ultrasonic waves, blurring the distinction between motion and noise treatments. It also increased variability through inconsistent vibration, making results less reliable. After removing the tin foil and improving the system, the variation became less pronounced, more consistent, and more clearly defined. This may have allowed true differences to emerge, showing that motion is more effective than noise alone and that combining both produces the strongest deterrent effect.

This experiment suggests that combining robotic bat motion with ultrasonic noise is a more effective deterrent than using either method alone. Such a combination could revolutionize current bat deterrent technology, as most existing practices rely solely on ultrasonic deterrents that have shown mixed results. The data demonstrates that robotic bat motion was superior to broadband-like ultrasonic noise in increasing variability in ultrasonic sensor readings within a model wind turbine environment. This increased variability indicates that this deterrent could potentially interfere with bat echolocation, keeping bats away from turbines, thereby reducing fatalities. This not only opens the door to further research into motion-based deterrent technologies but also raises awareness of the topic, enabling the development of more reliable technologies to address the growing issue of bat mortality due to wind turbines. While research has shown that ultrasonic deterrents can reduce deaths for some bat species, their effectiveness is inconsistent, partly because bats can adapt to the noise. In contrast, it is not yet known if bats would adapt to robotic motion-based deterrents, making this an important area for continued investigation and future field trials.

If this experiment were conducted differently, it would be reinforced to allow it to be tested outside. This would allow for results while accounting for external factors such as rain, heat, and environmental noise. It can then be compared to ultrasonic noise, as it’s known that ultrasonic acoustic deterents are less effective in warm environments. Future research should investigate bats' responses to combined and/or motion-deterrent methods applied to a turbine over time, and whether these methods remain effective in real-world conditions.

I am extremely grateful to everyone who made this year's CWSF and the Northern BC regional fair possible. Special appreciation to ARC Resources for funding my trip to Edmonton. I would like to thank my sponsor teacher, Mr. O’Brien, for all the support he provided with the forms and registration. I’m so appreciative of Jenny Copeland, our regional fair director, for the help she provided with the fair this year and for her many years of helping with science fairs (this being her last). A ginormous thank you to my dad, as he made sure I didn't drill a hole in my finger! Another huge thank you to my mom, as I know it's hard to see me occupy your basement for weeks with power tools. I’m so grateful I have all these people supporting my curiosity and being there when I need them!

References (APA)

Websites

Arnett, E. B., Hein, C. D., Schirmacher, M. R., Huso, M. M., & Szewczak, J. M. (2013). Evaluating the effectiveness of an ultrasonic acoustic deterrent for reducing bat fatalities at wind turbines. PLoS ONE, 8(6), Article e65794. https://doi.org/10.1371/journal.pone.0065794

Baerwald, E. F., D’Amours, G. H., Hadwin, B. J., & Barclay, R. M. (2008). Barotrauma is a significant cause of bat fatalities at wind turbines. Current Biology, 18(16), R695–R696. https://doi.org/10.1016/j.cub.2008.06.026

Brokaw, A. (2025, August 27). Acoustic overload: How noise pollution impacts bats. Bat Conservation International. https://www.batcon.org/acoustic-overload-how-noise-pollution-impacts-bats/

Bats and Wind Energy Cooperative. (2021). Curtailment and deterrence. https://www.batsandwind.org/research/curtailment-deterrence

Drost, P. (2025, October 30). Wind turbines keep killing bats in Canada. Advocates say this needs to change. CBC. https://www.cbc.ca/radio/whatonearth/bats-dying-wind-turbines-9.6952188

Eco Wind. (2024, December 3). Ultrasonic deterrents to reduce bat mortality at wind turbines—Short science summary. Tethys. https://tethys.pnnl.gov/summaries/ultrasonic-deterrents-reduce-bat-mortality-wind-turbines-short-science-summary

Gilmour, M. E., Richardson, M. S., & Haussler, M. U. (2020). Comparing acoustic and radar deterrence methods as mitigation measures to reduce human–bat impacts and conservation conflicts. Global Ecology and Conservation, 22, Article e00914. https://doi.org/10.1016/j.gecco.2020.e00914

Guest, E., Pearce, H., & Lucas, J. (2022). An updated review of hypotheses regarding bat attraction to wind turbines. Mammal Review, 52(2), 187–200. https://doi.org/10.1111/mam.12273

Zwilling, J. (2025, September 9). Bats use both sight and sound to hunt more efficiently in light, miniature sensors show. Phys.org. https://phys.org/news/2025-09-sight-efficiently-miniature-sensors.html

Lawson, M., Jenne, D. S., Thresher, R., Houck, A., & Straw, B. (2020). An investigation into the potential for wind turbines to cause barotrauma in bats. Scientific Reports, 10, Article 22394. https://doi.org/10.1038/s41598-020-77748-w

North American Society for Bat Research. (2024). NASBR statement on wind energy impacts on bat populations. https://www.nasbr.org/wind

van Hoof, P., Bittel, J., & Holderied, M. (2015, March 26). How do swarming bats avoid crashing into each other? National Geographic. https://www.nationalgeographic.com/animals/article/150326-bats-animals-science-echolocation-traffic-navigation

Warner, K. (2016, April 15). Echolocation - Bats. U.S. National Park Service. https://www.nps.gov/subjects/bats/echolocation.htm

Weaver, S. P., Hein, C. D., Simpson, T. R., Evans, J. W., & Castro-Arellano, I. (2020). Ultrasonic acoustic deterrents significantly reduce bat fatalities at wind turbines. Global Ecology and Conservation, 24, Article e01099. https://doi.org/10.1016/j.gecco.2020.e01099

Werber, Y., Levin, E., & Yovel, Y. (2022). Drone-mounted audio-visual deterrence of bats: Implications for reducing aerial wildlife mortality by wind turbines. Remote Sensing in Ecology and Conservation, 9(2), 251–262. https://doi.org/10.1002/rse2.308

AI

OpenAI. (2026). ChatGPT (April 11 version) [Large language model]. https://chatgpt.com

Pictures

Northern American Society for Bat Research. (2024). Bat deaths due to wind turbines in Canada [Table]. NASBR. https://www.nasbr.org/wind

Northern American Society for Bat Research. (2024). Bat deaths due to wind turbines in the United States [Table]. NASBR. https://www.nasbr.org/wind

Wainscoat, J. (2018). Brown bat [Photograph]. Unsplash. https://unsplash.com/photos/brown-bat-_f8IZ0gGS6E

Bat near a wind turbine [Photograph]. (2026, January 9). Energies Media.  https://energiesmedia.com/wind-turbines-triggered-a-reaction-save-bats/

ATL Equipment Manual. (n.d.). Ultrasonic sensor module HC-SR-04 [Photograph]. https://atl.aim.gov.in/ATL-Equipment-Manual/ultrasonic-sensor-module-hc-sr-04/