Originally published on The Daily Flare.
New biodiversity monitoring tools combine artificial intelligence, remote sensors and field research to help scientists detect biodiversity loss across forests, wetlands and other ecosystems.
From field observations to automated detection
Wildlife surveys have traditionally depended on researchers visiting sites, finding organisms and recording what they observe. That work remains essential, but it takes time and skilled staff, and repeated surveys can vary because observers, locations or methods change.
Digital monitoring is adding a different kind of evidence. Camera traps photograph animals, acoustic recorders capture calls and other sounds, and aircraft-mounted instruments collect images and measurements across broad landscapes. Artificial intelligence can help process the resulting files, identifying likely species in recordings or pictures so researchers do not have to review every item by hand.
Science reported that conservation researchers see this automation as a way to manage large volumes of field data more quickly. Tanya Berger-Wolf, a computer science professor at Ohio State University and co-founder of the nonprofit Wild Me, said AI can cut some analysis tasks from years to minutes. The outlet also cited a 2022 study in which an AI application processed camera-trap data more than five times faster.
The distinction matters. A sensor records a signal, such as an image or a bird call. Software suggests what produced it. Scientists still need to check those identifications, interpret observations in context and connect them to repeated measurements.
The UK Centre for Ecology & Hydrology is working with international partners to test automated monitoring stations in Europe, North America and tropical regions. The solar-powered stations can run on their own for months and include high-resolution cameras for insects and acoustic equipment to record birds and bats.
That work addresses a major information gap. UKCEH says insects account for about four-fifths of animal life, but only around one-fifth of insect species have been identified.
Measuring ecosystems at several scales
Monitoring can also extend beyond individual animals. NASA-funded BioSCape brought more than 160 scientists together to study terrestrial, marine and freshwater ecosystems in South Africa's Greater Cape Floristic Region, a biodiversity hotspot. Researchers combined fieldwork with observations from two NASA aircraft and one South African aircraft.
The project demonstrated how airborne measurements can contribute to biodiversity research across a large area. The team used the combined evidence to study changes in community composition, ecosystem disturbance, resilience and recovery, and ecosystem function.
Different instruments answer different questions. Images can help map vegetation and habitat structure; thermal measurements capture patterns of heat; field teams can verify what is present on the ground. In the Amazon, digital ecologist Thiago Sanna Freire Silva uses light detection and ranging, or LiDAR, to map flooded forests in three dimensions.
In the Amazon, the Mamiraua Sustainable Development Institute's Providence Project uses automated sound and image systems to study wildlife. More than 40 sensors are spread through the forest and operate around the clock. Biologist Emiliano Ramalho said the systems allow researchers to observe species and behaviors that would be difficult to follow through ordinary field observation alone.
Better records, with limits and local needs
Long-running, standardized observations are one potential advantage of automated stations. UKCEH says automated systems may reduce some variation found in traditional surveys. Images and recordings also leave records that can be checked later.
However, the technology is still developing. UKCEH describes camera systems, recording equipment and recognition software as being at an early stage, despite encouraging initial trials. Sensors may not capture every species equally well, and identification software depends on the images and recordings it has been trained to recognize.
Access to the data matters as well. UKCEH and The Alan Turing Institute are developing edge processing that would let devices analyze images where they are collected, rather than sending files elsewhere for later processing.
Researchers working on the Rainforest X competition described another approach to field surveys: combining drones, audio sensors and environmental DNA analysis, with AI used to analyze the collected material. The challenge involved surveying 100 hectares of Brazilian rainforest within 24 hours and producing useful findings within 48 hours.
Monitoring can inform conservation by showing where species occur, how ecosystems are changing and which places may warrant attention. But sensors do not protect habitat on their own, and a faster identification system cannot substitute for decisions about land use, resources or management.
Together, AI-assisted identification, field sensors, aircraft surveys and LiDAR offer ways to observe biodiversity at different scales, from an insect at a monitoring station to vegetation patterns across a forest. The methods are complementary rather than interchangeable. Continued testing, transparent validation and local capacity will shape how useful they become, while fieldwork remains central to checking what the instruments detect and what those observations mean.
Top comments (0)