Our planet’s ecosystems are deeply interwoven. When a thread is tugged somewhere – say, a species becomes locally extinct – the weave changes. For example, the loss of elephants in a tropical rainforest prevents the seed dispersal of some of the ‘forest giant’ tree species, and ultimately leads to reduced carbon sequestration.
Imagine if we could keep our eyes and ears on all of the planet’s landscapes and ecosystems, from wherever we were on the globe. If we could reliably look and listen out for these kinds of ecosystem changes, and respond to them effectively – before it was too late.
A collaboration between biodiversity experts at the Center for International Forestry Research and World Agroforestry (CIFOR-ICRAF) and open-source AI research community Enterprise Neurosystems aims to do just that, by deploying new low-cost, high-spec data loggers across landscapes and pairing them with machine learning processes to enable rapid retrieval and analysis of a wide range of data.
The project’s ultimate aim is ambitious, but plausible: to create a global monitoring network that allows us to connect up information on species and ecosystems and take a holistic ‘planetary pulse’. This can help direct and hone funds and action on some of our most urgent local and global challenges: biodiversity, climate, food systems, and more.
We’re in the midst of a biodiversity crisis. Global wildlife populations, for instance, have dropped by an average of 69 percent since 1970, according to the Living Planet Index (LPI), which states that “[t]he staggering rate of decline is a severe warning that the rich biodiversity that sustains all life on our planet is in crisis, putting every species at risk – including us.”
To help conserve the species that remain – and our lives as we know them – we need to be able to monitor and evaluate ecosystem health, including the size of key wildlife species populations, in ways that are accurate, cost-effective, and quick and easy to analyze.
At the moment, the main way that data on species abundance is gathered is through camera trapping. Machine learning techniques are increasingly being used to identify species, which saves researchers a lot of time compared to manual identification.
But these methods are still costly and time-consuming. Camera traps are expensive, at around USD 400 each. They also require new batteries every few months, and researchers must also return to the site to pick up the SIM cards containing their data at the end of each survey period for upload and analysis. This hampers the ability of researchers and landscape managers to monitor effectively at scale, and to perceive and respond to threats in a timely way.
The new data logger and programming system designed by Enterprise Neurosystems has the potential to transcend many of the above challenges.
The loggers are low-cost, at about USD 25 apiece, meaning they can feasibly be deployed at scale throughout a landscape, with little maintenance required because their batteries last about ten years. They collect a wide range of data, including images, soundscapes, levels of key gases such as oxygen, carbon dioxide, and methane.
They also network between themselves and transmit data to the cloud through mobile or satellite links, where it’s collated and analyzed through machine learning. This can help individual projects get better results, and aid decision-making at wider scales by syncing data from a large number of projects and landscapes.
Where we’ll start
The next step for this system is rigorous field testing to enable further development of the hardware, technology, and ML approaches. CIFOR-ICRAF will trial the capabilities of this new approach in several sites where we are already actively supporting sustainable wildlife management approaches: Rupununi in Guyana, Northern Ghana, and Yangambi in the Democratic Republic of the Congo.
We plan to install data loggers at these sites in early 2024 and test them over the course of the year, with the aim to expand our testing across these landscapes and into other CIFOR-ICRAF sites if early results show potential.
What this data can do
There are lots of good reasons for gathering better biodiversity data.
As a global community engaged in environmental action, we will need it to know whether or not we’re on track to achieving the 23 targets in the Kunming-Montreal Global Biodiversity Framework (GBF), such as:
Target 3: 30 percent of areas are effectively conserved
Target 4: Threatened species are recovering, genetic diversity is being maintained and human-wildlife conflict is being managed
Target 5: Use, harvesting and trade of wild species is sustainable, safe and legal
Clear and comparable biodiversity data will also be key to the development of Voluntary Biodiversity Credits, in which there is increasing global interest. Metrics to define VBCs, and methods for populating them, are currently being discussed by many of the proponents of the approach, but scalable and field-tested technology and techniques are now needed to advance the concept further.
For CIFOR-ICRAF’s scientists, getting good data on wildlife abundance will be especially key to advancing the work of the FAO-led Sustainable Wildlife Management (SWM) programme, which seeks to tackle the challenge of overhunting by addressing both wildlife conservation and food security. If researchers and land managers know the status of wildlife populations, they can manage hunting activity – through mechanisms like quotas and rotation of hunting areas – accordingly to ensure those species’ survival, and local communities’ food security and livelihoods, into the future.
To find out more, please contact CIFOR-ICRAF senior scientist Lauren Coad: email@example.com
We want you to share Forests News content, which is licensed under Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0). This means you are free to redistribute our material for non-commercial purposes. All we ask is that you give Forests News appropriate credit and link to the original Forests News content, indicate if changes were made, and distribute your contributions under the same Creative Commons license. You must notify Forests News if you repost, reprint or reuse our materials by contacting firstname.lastname@example.org.