Post provided by Kayla Kuhlmann
Ecologists have started looking into drones as new tools for wildlife surveys, but how can drone disturbance be minimized in order to produce accurate wildlife counts? In this post, Kayla Kuhlmann describes a drone practice to reduce disruption during acoustic bat surveys as featured in the paper “Miniaturization eliminates detectable impacts of drones on bat activity”, recently published in Methods in Ecology and Evolution.
Drones offer an appealing tool for wildlife surveys
Several countries have lifted legal restrictions on drone flight, and now researchers can consider how drones offer advantages over current methods in wildlife surveying. For instance, drones can navigate over inaccessible landscapes quickly and safely compared to surveyors on foot. In some cases, drones could offer a safer, cheaper, and quieter method to collect aerial data that traditionally requires manned aircraft, such as helicopters. Also, drones come in many forms (from rotary quadcopters, to fixed wing “planes”, and boat drones) and can carry a variety of payloads (including cameras, acoustic recorders, GPS devices, and frequency receivers) which make them versatile for wildlife surveys.
Despite the enthusiasm to conduct wildlife counts with drones, some surveyors still hesitate because of the conflicting reactions by wildlife. Although studies focused on monitoring rhinos, bats & birds, and wildlife in general observed little to no obvious behavioral response to UAV flight, others have documented strong behavioral responses (e.g. seabird species, particularly sub-Antarctic seabirds) and physiological stress responses (e.g. bears). Researchers do not want to disturb the wildlife they are studying, especially since doing so can cause animals to abandon the site and change the results of the surveys.
The several types of drones and the differences in their characteristics inspired us to examine how variation in drone size, noise, and frequencies would affect wildlife activity. The goal was to discover whether different drone models had varying effects on wildlife and to determine which drone characteristics make them most disturbing to our study species.
Detecting bats acoustically with drones
Bats are typically surveyed with specialized acoustic recorders that pick-up ultrasound frequencies. Traditionally, bat biologists detect bats on foot and carry the acoustic recorder by hand. Even though acoustic recorders make bat detection significantly easier, bats are still difficult to survey accurately since they occupy aerial habitat. Essentially, bat species that forage at certain heights above the forest canopy or occupy certain frequencies with their echolocation calls are underrepresented in survey counts because of the limited detection range of acoustic recorders. Since many bat species lack reliable population counts, the conservation status of several species is unknown.
Drones would be useful tools to help survey bats, since elevating the acoustic recorder would bring it into the range of detection for bats that are frequently missed. Additionally, some acoustic recorders are small enough to be added as a payload to many drones. Researchers would be more optimistic about conducting bat surveys with drones if not for the large disruption drones may cause bats. Since bats are nocturnal and tend to avoid noisy environments, they are at risk of being significantly disturbed by drone flight.
Using bats as our study subject, we flew three different drones in bat habitat. We conducted acoustic detection before, during, and after drone flights to measure how bat activity differed between periods of drone disturbance compared to periods of no drone disturbance. Acoustic profiles were conducted on the three drones to define their characteristics. The significant characteristics compared in this study were the drones’ dimensions, weight, noise intensity and frequency range. Based on how bat activity responded to flight with each drone, we can see which drone characteristics are most disruptive.
Fewer bat passes are detected during flights with larger drones
Our acoustic profiles demonstrated that drone size and noise intensity are correlated, indicating that the largest drone (in size and dimension) was the loudest and the smallest drone was the quietest. However, when it comes to frequency range, there is no correlation between drone size and frequencies emitted. These profiles were key to deciphering which drone characteristics disturbed bats the most.
The least bat passes were detected when the loudest and largest drone flew, and bat activity was unaffected by the quietest and smallest drone. Indeed, bat activity was directly correlated with the size and noise intensity of the drones. The differing frequencies emitted by the drones did not have any detectable impact on bat activity, as the smallest drone occupied some of the highest frequencies and the medium-sized drone had the lowest frequency range. It was interesting that most of the bats returned after the drone finished flight and powered off, and that different bat species had varying levels of sensitivity to drone flight.
Summary
Some drones have characteristics that are less disturbing to wildlife and are better suited for wildlife surveys compared to others. For bats, and likely many other animals, smaller and quieter drones minimize the impact of drones on wildlife surveys. Despite this finding, wildlife surveys often require sensors that may be too heavy for miniature drones. With this in mind, we encourage drone users to use the smallest and quietest drone possible within the constraints of their study, and make modifications to ensure their drones produce less noise and disturbance to wildlife. This case study demonstrates that drone miniaturization eliminates any measurable impact of drones on wildlife and therefore produces more accurate and less invasive drone-based acoustic wildlife surveys.
To read the full Methods in Ecology and Evolution article, click on the following link: Miniaturization eliminates detectable impacts of drones on bat activity.
Methods Blog 