Restoring rainforests is vital for climate and biodiversity crises. Research in Ecuador's Chocó rainforest, a biodiversity hotspot, employed bioacoustics and DNA-based methods to assess post-agriculture forest recovery. Experts discovered that vocalizing vertebrates' community composition, rather than species richness, correlated with the restoration gradient.
The study, titled "Soundscapes and deep learning enable tracking biodiversity recovery in tropical forests" published in Nature Communications, addresses the question of whether planting trees revive animal communities in regenerating forests. Through acoustic monitoring and DNA surveys, researchers observed species returning to these regenerating forests within a few decades.
AI-Powered Bioacoustic Analysis Listens to Animals in Rainforests To Assess Biodiversity in the Area
Bioacoustic Analysis Using Artificial Intelligence
The rainforests teem with the symphony of wildlife sounds, serving not only as a source of delight but also as a valuable resource for ecologists seeking to gauge land biodiversity.
Traditional bioacoustic analysis is time-consuming and reliant on expert interpretation. However, a group of researchers, led by ecologist Jörg Müller from the University of Würzburg, introduced a more efficient approach.
They harnessed the capabilities of computer technology, akin to smartphone apps that identify birds, bats, or mammals based on their vocalizations, and applied this principle to conservation efforts.
The research encompassed 43 sites within the Ecuadorean rainforest, spanning pristine old-growth forests, areas recently cleared for pasture or cacao cultivation, and regrowing forests.
Sound recordings were collected at four-hour intervals over two weeks. These audio recordings were manually analyzed by experts to compile a species list, revealing a positive correlation between the duration of land clearance and increased biodiversity.
Subsequently, artificial intelligence (AI) models trained on Ecuadorian sound samples were deployed to identify 75 bird species from the recordings. Impressively, the AI tools performed on par with human experts.
In addition to sound analysis, the researchers employed light traps to capture night-flying insects and used DNA analysis for their identification, affirming that the diversity of vocalizing animals effectively represented the broader biodiversity within the ecosystem.
Beyond the realm of ecology, the study's findings have broader implications. Corporations like L'Oreal and Shell, invested in forest restoration projects, may benefit from automated methods to monitor and standardize their ecological initiatives, ensuring alignment with their stated goals.
AI Serves as Scientists' Ears in the Forest
Illegal poaching in Africa has resulted in a significant decline in elephant populations, with some countries like Tanzania experiencing losses of over 60%. The Elephant Listening Project, in collaboration with Conservation Metrics, is employing AI and innovative listening devices to help rangers more efficiently track and protect elephants from poachers.
ELP has deployed listening devices in the Nouabalé-Ndok National Park in the Republic of the Congo to monitor elephant activity and detect poachers. This information is relayed to on-ground rangers for more effective patrols, and the sensors can also identify poaching sounds, like gunshots.
Previously, the delay in sharing this data with rangers was significant, sometimes taking three to four months due to manual filtering. Conservation Metrics aims to expedite this process by using AI and machine learning to analyze ELP's recordings, swiftly identifying elephants and poachers among various forest sounds.
The AI algorithm, trained on thousands of elephant call recordings, can potentially differentiate subtle acoustic distinctions with improved listening devices, addressing challenges such as data retrieval and transfer.
While there are concerns about the timeliness of gunshot detection, the project secured funding from Microsoft's AI for Earth Initiative, supporting conservation projects employing AI for tasks like route optimization and animal and poacher tracking.
RELATED ARTICLE: Vacation Photos of Zebras, Whales and Other Animals Can Help Scientists' Conservation Efforts, Fight Wildlife Extinction
Check out more news and information on Artificial Intelligence in Science Times.
![Earth's Quasi-Moon Kamo‘oalewa Could Originate From Lunar Surface Not Asteroid Belt [Study]](https://1721181113.rsc.cdn77.org/data/thumbs/full/53275/89/56/50/40/earths-quasi-moon-kamo-oalewa-could-originate-from-lunar-surface-not-asteroid-belt-study.png)
