Digital Tools for Accurate and Low-Cost Biodiversity Monitoring

Affordable and cost-efficient, unmanned aerial vehicles (UAVs) are increasingly being utilized for ecological monitoring and biodiversity conservation. Professionals, scientists, and even concerned citizens are using drones equipped with high-resolution cameras and visible, multispectral, and thermal sensors, to evaluate ecosystems, assess disturbances, and analyze the dynamics and changes of biological communities, among other applications. Increasingly, the images obtained by drones are used in combination with deep learning algorithms.^[1]

There are several examples where this technology was applied with significant results. Scientists of the Forest Global Earth Observatory (ForestGEO) have used drones to monitor tropical forests in three dimensions and to identify forest disturbances. Conservation agencies use drones to count wildlife populations and even as a tool against poaching. Different species—from birds to mammals, from the tropics to the polar regions—can be targeted for surveys.

While drones can be a useful tool, especially for repetitive surveys, it is always necessary to identify local and national regulations governing their use and to obtain the appropriate permits. Several private companies are now proposing “all-in” packages including compliance with administrative requirements, development of monitoring protocol, and data analysis and reporting. For example, some fields of application survey bats in wind farms: drones equipped with infrared cameras and other specialized sensors can be used to detect and identify bats by their unique heat signatures and echolocation calls. Organizations, such as Bat Conservation International, are actively promoting the use of drones for bat surveys and providing training and resources for researchers and conservationists (Figure 1).

​

Figure 1: A drone used to monitor bats inside a cave. Photo credit: U.S. Fish and Wildlife Service Headquarters, public domain, via Wikimedia Commons.

This technology uses satellites (Box 1) and other remote sensors for gathering data on land use, vegetation cover, and other key environmental factors. The data is then processed and analyzed to provide insights into the health and diversity of species and ecosystems. One of the key benefits of remote sensing technology is its ability to cover large areas quickly and efficiently. This allows researchers to gather data on biodiversity in remote or hard-to-reach areas that would be difficult to assess using traditional monitoring methods. Additionally, remote sensing data can be used to create maps and models that can be used to inform conservation planning and management efforts.

Remote sensing technology can be used to monitor biodiversity in a variety of ways, including:

• Vegetation monitoring: Tracking changes in vegetation cover, such as deforestation and forests regrowth, provide information that can be used to assess the health of different ecosystems and identify areas that are most in need of conservation efforts. Tools, such as Global Forest Watch, use high-resolution satellite imagery to identify land use change, including deforestation or forest fires, and can send automatic alerts if a significant change is identified (Figure 2). This kind of tools are extremely helpful in providing baseline data for planned project areas and monitoring potential indirect/induced impacts, such as human settlements encroaching into the forest after a road construction.
  

  ​

  Figure 2: Deforestation alerts near the Gunung Palung National Park in West Kalimantan, Indonesia (powered by Resource Watch ).

• Habitat mapping: Mapping the distribution of different habitats, such as wetlands, forests, and grasslands, helps identify areas that are critical for the survival of different species and to prioritize conservation efforts. LIDAR (Light Detection and Ranging) technology is a useful tool for biodiversity monitoring, especially in the context of habitat mapping and forest inventory. LIDAR works by emitting laser beams from an aircraft or drone and measuring the time it takes for the light to reflect from the ground. LIDAR can create highly detailed 3D maps of the terrain, including the height and structure of vegetation. It can also be used to map coastal and even underwater habitats. While most LIDAR sensors are currently ground-based or mounted on aircrafts, there are also a few examples of the sensors being used in space from satellites.

eDNA is genetic material obtained from environmental samples, such as soil, water, or air, without necessarily capturing or directly observing the organism. Animals (including humans) shed fur, fragments of skin, and other organic materials that contain DNA, which is unique for each species. Scientists are now able to amplify these fragments, sequence the genetic material, and cross-check the results against databases that contain the genetic sequence of several thousands of species.

This tool is particularly useful when the surveyor is interested in detecting the presence of rare (such as endangered species) or cryptic species, that are difficult to detect with traditional surveys based on direct or indirect observations. Another very interesting use is the monitoring of the presence of invasive species and, in case an eradication program is ongoing, it helps in making sure that it is effective. Sampling can be conducted by unskilled operators, who will just collect the sample and send it to the laboratory for analysis, reducing the need to organize complicated field expeditions by international experts (Figure 3).

There are still cons, such as the absence of some species in the DNA database (especially the most rare ones and undescribed species) that requires additional steps to be taken, including capturing the animal and “swabbing” it to have its DNA sequenced. However, the analysis may just confirm the presence of the species but not determine its abundance.

​

Figure 3: eDNA sampling for the proposed Asian Development Bank-funded Alaoa Multipurpose Dam, Samoa. Photo credit: ADB.

Another promising approach to biodiversity monitoring is recording and analyzing sound. By using specialized microphones and software, researchers can identify and track the vocalizations of different species, providing insights into the health and diversity of ecosystems. Acoustic monitoring (or Ecoacoustics) is not a new approach. It has been used for decades to detect the presence of cetaceans and bats, for example. Bird, frogs, and insects have distinctive calls that can be used to monitor their presence (Figure 4). For example, by analyzing recordings of bird songs, researchers can identify different species and track changes in their breeding patterns and habitat use over time.

Similarly, by monitoring bat calls, researchers can identify different species, track their migration patterns, and study the impact of environmental factors, such as habitat loss and climate change, on their populations. By having an “acoustic baseline” or “Acoustic Index” of a specific area, it is possible to monitor the relative biodiversity and to gauge if a specific development or habitat restoration is having the expected results. It is even possible to automatically detect harmful or illegal activities (such as logging and hunting) and send real-time alerts to authorities.

​

Figure 4: A sonogram of a recording of a Hawaiian Petrel in Haleakala National Park, Hawaii. Researchers are now able to identify the presence of multiple species at the same time using deep learning algorithms. Photo credit: National Park Service, Public domain, via Wikimedia Commons.

With the increase of computing power and the development of deep learning algorithms, scanning of sounds for animal identification has become accessible to everyone by just using a simple app for smartphones. With advances in technology and data analysis techniques, acoustic monitoring is likely to become an increasingly important tool for biodiversity monitoring.

Environment practitioners are used to doing mathematical modelling to calculate the impacts of development projects on air and water quality, noise, and hydrology, among others. Impacts on biodiversity are more difficult to model, mostly due to the complexity of biological interactions and the lack of reliable species distribution and habitat composition data.

Biodiversity models can be used to understand the factors that influence species distribution and abundance, predict the impact of environmental changes—such as the construction of a road (Figure 5), and inform conservation and management decisions.

The BILBI tool was developed by the Australian Science Agency (CSIRO) to assess biodiversity changes from fine spatial resolution images of the global land surface, using best-available biological and environmental data, modelling, and high-performance computing. BILBI can be used to monitor the progress toward general targets (such as those established by the Kunming-Montreal Global Biodiversity Framework), provide information to national and sub-national biodiversity reporting, assess project alternatives and evaluate impact of developments on biodiversity at a local scale, and identify the most suitable areas for biodiversity offsetting. This capability allows identification of biodiversity risks at the early stage of project development to avoid potential negative impacts, instead of mitigating them after the final design is established. The possibility to account for multiple developments, including in the assessment of potential cumulative impacts, would be essential to developing sustainable high-level plans informed by strategic environmental assessments.

​

Figure 5: BILBI model outcome of a proposed development of a road inside a tropical forest. The model was able to identify direct and indirect impacts and forecast species persistence in proximity of the development. Source: ADB.

^[1] Deep learning uses neural networks to enable machines to learn from large amounts of data, including images, and improve their performance on specific tasks. In biodiversity monitoring, it has been used to recognize and monitor cryptic wildlife using thermal scanning, identify single animals based on body markings for population-counts analysis, and many other applications.

Convention on Biological Diversity. 2022. COP15: Nations Adopt Four Goals, 23 Targets, for 2030 in Landmark UN Biodiversity Agreement. Kunming–Montreal Global Biodiversity Framework. Press release. 19 December.

Cornell University. Merlin Bird ID.

CSIRO. Macroecological Modelling: BILBI—Assessing the Consequences of Change for Biodiversity at Fine Spatial Resolution Globally.

D. Kadish and K. Stoy. 2021. BioAcoustic Index Tool: Long-Term Biodiversity Monitoring Using On-sensor Acoustic Index Calculations. Bioacoustics. 3. pp. 348–278.

D. Stowell and J. Sueur. 2020. Ecoacoustics: Acoustic Sensing for Biodiversity Monitoring At Scale. Remote Sens Ecol Conserv. 6. pp. 217–219.

eBioAtlas. Global repository for eDNA-based biodiversity data.

E.D. McCarthy et al. 2021. Drone-Based Thermal Remote Sensing Provides an Effective New Tool for Monitoring the Abundance of Roosting Fruit Bats. Remote Sens Ecol Conserv. 7. pp. 461–474.

Global Forest Watch website.

J. Hodgson et al. 2026. Precision Wildlife Monitoring Using Unmanned Aerial Vehicles. Sci Rep. 6. 22574.

J.M. De Vos et al. 2015. Estimating the Normal Background Rate of Species Extinction. Conservation Biology. 29. pp. 452–462.

K.C. Cushman et al. 2022. Soils and Topography Control Natural Disturbance Rates and Thereby Forest Structure in a Lowland Tropical Landscape. Ecology Letters. 25. pp. 1126–1138.

S. Sarab et al. 2023. Combining Machine Learning and a Universal Acoustic Feature-Set Yields Efficient Automated Monitoring of Ecosystems. bioRxiv. 865980.

WWF. 2022. Living Planet Report 2022—Building a Nature-Positive Society. R.E.A. Almond et al, eds. Gland, Switzerland: WWF.