- UK-based researchers who have developed a low-power, open-source acoustic monitoring device say it shows promise for monitoring wildlife and illicit incursions by mankind into remote habitats.
- The researchers say that the device, which is about the size of a matchbox, can be made for as little as $43 per unit — a price-point that could be key to ensuring coverage across large landscapes, where numerous monitoring devices are required.
- The AudioMoth can be programmed to monitor wildlife populations by recording the calls of specific target species while at the same time serving as an alert system when the sounds of human exploitation, such as the blast of a shotgun or the roar of a chainsaw, are detected.
Researchers in the UK who have developed a low-power, open-source acoustic monitoring device say it shows promise for monitoring wildlife and illicit incursions by mankind into remote habitats.
The team behind the AudioMoth wrote in a study published in the journal Methods in Ecology and Evolution last month detailing the results of proof-of-concept trials for the device that it was designed to overcome some of the barriers to effective remote monitoring of biodiversity: “The cost, usability and power efficiency of available wildlife monitoring equipment currently inhibits full ground-level coverage of many natural systems.”
The AudioMoth was created by two computer science PhD students at the University of Southampton, Andrew Hill and Peter Prince, together with Alex Rogers, a professor at the University of Oxford. The researchers say that the device, which is about the size of a matchbox, can be made for as little as $43 per unit — a price-point that could be key to ensuring coverage across large landscapes, where numerous monitoring devices are required.
What’s more, each AudioMoth unit can be programmed to monitor wildlife populations by recording the calls of specific target species while at the same time serving as an alert system when the sounds of human exploitation, such as the blast of a shotgun or the roar of a chainsaw, are detected.
Evelyn Piña Covarrubias, a PhD student in biology at the University of Southampton and a co-author of the Methods in Ecology and Evolution study, told Mongabay that the device was named AudioMoth “because moths have the most sensitive hearing of any animal.” (A 2013 study found that the greater wax moth, Galleria mellonella, can hear ultrasonic frequencies up to 300 kHz).
Simply hearing exceptionally well is not all that an acoustic recorder in, say, a remote rainforest has to be capable of, however. For one thing, it must withstand rather harsh conditions. After an initial field test of the devices, nearly one-quarter of the AudioMoths that had been deployed had some water damage, so the team subsequently tried out a combination of an off-the-shelf waterproof electronics enclosure with a waterproof, permeable acoustic membrane and silica gel sachets to absorb moisture. This solution proved effective at protecting the devices while allowing clear acoustic recordings to be made, and only cost about $8 per unit.
To overcome the energy hurdles posed by the need to continuously monitor large areas of remote wilderness, the AudioMoth is designed to trigger “event logging” only when its classification algorithms identify a “desired acoustic event” — i.e. the call of a target species or the sound of a potentially illegal human activity. Essentially, the AudioMoth wakes up periodically and takes a brief audio sample for analysis. If the algorithms detect a noise the AudioMoth has been configured to capture, it will start recording. If not, the device goes back to sleep, thus saving energy.
“AudioMoth is more energy efficient than currently available passive acoustic monitoring (PAM) devices, giving it considerably greater portability and longevity in the field with smaller batteries,” the researchers write in the study.
Real-time alerts and the conservation of big cats in the Yucatán Peninsula
Covarrubias is part of a team that also includes researchers from Universidad Autónoma Metropolitana in Mexico that has been studying populations of jaguars and pumas using AudioMoth devices in three contiguous regions of protected and unprotected forest in Mexico’s Yucatán Peninsula.
The big cats’ preferred prey is peccaries, deer, and coati, all species for which the cats have considerable human competition, given that they are regularly hunted by local communities.
“Rural communities living near these nature reserves manage the natural resources in their communal lands, known as ejidos,” Covarrubias said in a statement. “Subsistence hunting of game meat and logging by residents is permitted in the ejidos, but there are no effective measures in place to regulate the hunting pressure.”
Reduced availability of their primary prey and habitat fragmentation driven by livestock production pose a major threat to the cats. It’s estimated that there are just 6,000 jaguars left on the Yucatán Peninsula, inhabiting less than 40 percent of their historical range. There isn’t as much information available on puma populations because, unlike jaguars, pumas do not have unique spot patterns by which individuals can be identified and studied via camera traps.
Conservation efforts for both species are often hampered by lack of data and insufficient resources for detecting and disrupting illegal activities. But Covarrubia says that the AudioMoth can help change that by transmitting real-time alerts to park rangers when it detects a gunshot or a truck engine where those sounds should not be heard. The alert would include not just information about the type of sound detected but also its location.
“Protected areas across the Yucatán Peninsula are far too under-resourced to afford effective and safe patrols of the vast tracts of natural forests,” Covarrubia noted. “Most acoustic loggers on the market are too expensive for network deployments or have a short battery life. With AudioMoth, local rangers and managers will have access to a monitoring system covering potentially large areas and can act immediately on alerts.”
AudioMoth case studies
While the full results of the study in the Yucatán Peninsula have not yet been released, the Methods in Ecology and Evolution study describes two real-world case studies the AudioMoth team carried out to demonstrate the device’s capabilities.
The first case study was designed to test whether or not there was an active population of Cicadetta montana, the only cicada species native to the UK, in the country’s New Forest National Park. Because New Forest cicadas spend most of their lives as nymphs underground, only emerging as adults in roughly seven-year cycles, they are particularly difficult to detect through traditional survey methods.
The AudioMoth team deployed 87 of the acoustic monitoring devices in four locations in New Forest National Park between 2016 and 2017. The researchers used recordings of the cicadas made in Slovenia to determine that an “extended buzz lasting 30 [seconds], with a dominant 14 kHz frequency band” characterized the call made by males of the species. Because that frequency is rare in the calls of other insect species of the New Forest, it was used to inform the detection algorithm used by the AudioMoths.
The devices ultimately did not log a single call of the New Forest cicada over the course of the two-year study period, but the algorithm the team developed for identifying the insects was tested by playing the cicada recordings from Slovenia inside an anechoic chamber together with recordings of background noise captured in the New Forest, and “the algorithm achieved a true positive rate of 0.98 and a false positive rate of 0.01.”
And even the failure to detect the New Forest cicada was instructive, in its own way. Over the course of the study, those 87 AudioMoths captured 129 hours of audio triggered by false positive algorithm responses from sources like dog whistles, leaf noises in strong winds, and bird songs. Sorting through that data and verifying that they were indeed false positives was a far simpler task than it would have been had acoustic monitors that were recording continuously been deployed: Over the study period, continuous recorders would have captured about 156,000 hours of audio data that would then require analysis.
“The capacity to perform real-time detection with a programmable algorithm vastly reduces on-board memory requirements, and post-deployment data analyses,” the researchers write in the study.
The second case study tested the AudioMoth’s ability to recognize and capture gunshots in the tropical rainforests of Pook’s Hill Reserve, a private nature reserve in Belize. The researchers deployed 36 AudioMoths in pairs, one configured to trigger a recording only when gunshots were detected, the other to record continuously. They then fired 65 test shots using two guns that are commonly used for hunting in the area: a 12 gauge and a 16 gauge shotgun.
They used data generated by these experimental gunshots to develop an algorithm that looks for “the characteristic rate at which select frequencies peak and then decay from the initial muzzle blast.” When the team ran the data from their test shots through the AudioMoth’s detection algorithm, it was able to identify the gunshots that were as much as 500 meters away “with a success rate of 66%,” they write in the study.
While the AudioMoth comes with configuration software that allows for basic customization of the device, some low-level knowledge of the programming language C “is required to achieve full use of AudioMoth’s flexibility and produce new detection algorithm implementations,” they add.
But ultimately, they don’t see that as being an insurmountable barrier to wide-scale deployment of the device: “AudioMoth provides opportunities for groups with limited budgets to perform systematic bioacoustics research, for example by benefiting from economies of scale in group purchases. With further developments in the new technologies described here, we are getting closer to achieving a basic requirement of sustainable development, that local communities can afford to monitor their own natural resources.”
• Hill, A. P., Prince, P., Covarrubias, E. P., Doncaster, C. P., Snaddon, J. L., & Rogers, A. C. (2017). AudioMoth: Evaluation of a smart open acoustic device for monitoring biodiversity and the environment. doi:10.1111/2041-210X.12955
• Moir, H. M., Jackson, J. C., & Windmill, J. F. (2013). Extremely high frequency sensitivity in a ‘simple’ear. Biology letters, 9(4), 20130241. doi:10.1098/rsbl.2013.0241
Follow Mike Gaworecki on Twitter: @mikeg2001
FEEDBACK: Use this form to send a message to the author of this post. If you want to post a public comment, you can do that at the bottom of the page.