In a leap for bat monitoring in India, bat biologist Kadambari Deshpande and technologist Vedant Barje have developed an automated bat detection and classification system. Created in October 2024 and introduced in February 2025, BatEchoMon is an innovative, solar-powered technology that makes data processing faster and more efficient, with near real-time insights.

The researchers have now installed two units of BatEchoMon at a field site on private land in Maharashtra to monitor and understand the importance of bats in agro ecosystems. “For instance, insectivorous bats control agricultural insect pests of several crops, and thus contribute to the protection of yields from pest-driven losses,” says Deshpande, post-doctoral fellow, Long-Term Urban Ecological Observatory (LTUEO) and School of Environment and Sustainability (SES) at the Indian Institute for Human Settlements (IIHS), Bengaluru.

Around the world, bats face threats from habitat loss, human activities, climate change, hunting, and disease. Over a third of bat species are considered threatened or data deficient according to the International Union for Conservation of Nature (IUCN). Globally, 18% of all bat species are data deficient – categorised as such when there is not enough information available to make a reliable assessment of its conservation status or risk of extinction – which is much higher than reported for either other mammals (13%) or birds (1%).

Conservation of bat species requires sustained efforts to monitor them, identify stressors, assess population trends, and develop a comprehensive catalogue for reference. Moreover, quantification of bat activity is important to answer research questions related to bat ecology, habitat use by species, and ecosystem services, Deshpande explains. “The need for a system that could provide reasonably accurate quantification of bat activity, based on acoustic recordings, was the main motivation for developing BatEchoMon,” she says.

The need for an automated system

Bats use sound in a unique way. They navigate using echolocation, which refers to emitting high-frequency ultrasonics and listening to the echo. Using the echo, they determine the size and shape of objects in the environment.

The frequencies of echolocation of insectivorous bats are usually between 15 kHz and 200 kHz, and some go beyond that, the Indian Institute for Human Settlements (IIHS) explained in an online post. For a long time, scientists have used bat ultrasounds to monitor bats in their natural habitat by recording the calls using bat detectors — devices that pick up ultrasounds emitted by bats and convert them into sounds that are audible to humans.

Using acoustics to study bats gained popularity since the early 2000s and has witnessed a rise in India in the last decade, Deshpande says. In India, acoustic methods are mostly used for species identification and distribution.

The number of bat echolocation calls is a useful index of activity in insectivorous bat monitoring. “Questions about how insectivorous bats use different habitats and at what magnitude they control agricultural insect pests as well as disease vectors such as mosquitoes can be answered using acoustic data on feeding events,” Deshpande explains.

Researchers currently use bat detectors, but they come with significant challenges. For instance, there is a lack of a comprehensive reference resource for Indian bats as many species’ calls are still unrecorded and sorting out bat calls from background noise is an exhaustive process. In the post, IIHS also points out that most of the current software is made for European and American bats, which means Indian species might be misclassified or missed.

Bat detectors also require painstaking manual work to review and correct software-sorted calls. Collecting acoustic data over long periods involves manual work to screen calls, distinguish species, and estimate bat activity, Deshpande explains. “This is especially difficult and time-consuming when the aim is to get total levels of bat activity at large time scales such as overnight or across seasons to measure bat feeding in agricultural habitats. For such purposes, manual work for timely results and inferences is not possible in a reasonable time frame. This issue limited us from asking interesting questions at scale,” she adds.

Listen:

BatEchoMon as a solution

BatEchoMon is an autonomous system that runs on a Raspberry Pi microcomputer and a microcontroller. It is also connected to AudioMoth, a popular low-cost ultrasonic detector, programmed to work as an ultrasonic microphone for the system. “Raspberry Pi is the brain of the whole system. We collect bat acoustic data from the Audiomoth and process or make sense of it using Raspberry Pi,” explains Barje. He works as a consultant at IIHS LTUEO and as lead at WildTech Project, Wildlife Conservation Trust.

The system is configured to activate automatically at sunset when insectivorous bats begin flying for food, and records their ultrasounds using AudioMoth. The recordings are then transferred to the microcomputer, where the data is sorted and stored for real-time analysis.

The bat calls are first separated from other overlapping sounds such as of insects, or other anthropogenic or environmental ultrasounds such as the ultrasonic components of truck horns or electric wire noise, that may get recorded occasionally. Then, the peak frequency and the structure of a call are analysed to match with a pre-trained model to identify the species. “This is done using a Convolutional Neural Network (CNN) based algorithm. The output of this process is a spectrogram containing bat calls. Only files with bat calls are further stored as raw audio files,” Barje says. This saves storage space and analysis time as all the non-bat call parts are removed.

The system can detect ultrasounds up to a frequency of 192 kHz and can currently identify six to seven common Indian bat species. BatEchoMon is customised to study insectivorous bats, as fruit bats (except for the fulvous fruit bat, Rousettus leschenaultii) do not use echolocation to navigate, communicate and feed, Barje says.

Since it uses solar panels to charge its batteries, they last longer than usual. The total production cost of BatEchoMon is about one-third the cost of advanced bat detectors and similar systems available in other countries, the researchers reveal.

While BatEchoMon is not a bat detector, “It is a system that includes bat detectors and processes data as it is being recorded, and analyses species-wise bat activity on-the-fly so that scientists can directly get bat activity summaries in real time,” Barje explains.

Generally, bat detectors are trigger-based, which means they only record if a bat or bat-like frequency sound is present around the recorder. BatEchoMon listens continuously and stores only those files that include bat calls. “It also generates different statistics such as which species has been most active through the night, which species was active when, and so on, providing the total number of species present in the area throughout the recording period,” Barje says. Whereas with existing detectors, these have to be manually interpreted using hours of data for one night.

The current challenge is collating training datasets to expand the number of species that BatEchoMon can identify. The training process needs to be done manually which takes a lot of time and effort, but once done, it can last forever. “BatEchoMon has been tested in one of our field sites in Maharashtra,” Barje says.

The innovators now want to engage more bat researchers and institutions with BatEchoMon’s application. According to Deshpande, “other bat researchers can help make the system more robust by contributing training data,” and then it can be made available to citizens to better understand the bats around them. Since it is more affordable than bat detectors and the processing and analysis time after recording is lesser, they could also be made more accessible to citizen scientists in the future.

When asked about the significance of automated monitoring systems in the future, Deshpande says that in the context of long-term monitoring, they are important. “Automated monitoring using BatEchoMon may be one of the important ways forward to increase the uptake of non-invasive bat research, which is still recent in India. This can further our understanding of bats, their ecology, and ecosystem services,” she adds.

Read more: Darkness-loving bats are being nudged out of the shadows

Banner image: Bat biologist Kadambari Deshpande and technologist Vedant Barje with the BatEchoMon system installed by them. Image by Vivek Hasyagar.

Credits

Priyanka Shankar Editor

Topics

EcologyTechnologyTechnology and ConservationWildlifeMaharashtra

See Topics