According to the WWF Living Planet Report 2024, wildlife populations around the world have declined by an average of 73% over the past 50 years, underscoring an urgent crisis for biodiversity and the health of ecosystems worldwide. Biodiversity, the sum of all life from genes to organisms and their habitats on earth, is a rich network of interdependent species across a wide range of ecosystems. Humanity depends on it for clean air, water, soils, foods, medicines, and much more. Declines in biodiversity are recognized as a major economic threat to human prosperity and well-being.

With extinction rates an estimated 100 to 1,000 times pre-human levels, the rapid erosion of biodiversity is a critical issue, closely interlinked with climate change, pollution, the degradation of wildlife habitat, and over-exploitation of other natural resources. Despite a widespread recognition of these issues, we struggle to obtain continuously updated data at the scale we need to assess the progress of our conservation efforts towards international biodiversity targets, such as those set in 2022 by the Kunming-Montreal International Biodiversity Framework.

The paucity of data is striking, particularly with regard to the most hyperdiverse groups, such as invertebrates, plants, and fungi in megadiverse regions, but even for iconic mammalian species such as the cheetah, there are country-sized gaps in our knowledge of numbers and distribution.

Data Inflow Can’t Keep Pace With Species Declines

Obtaining reliable data on the numbers and distribution of species and their status requires significant funding for typically lengthy and challenging field studies, in addition to highly trained experts. As a result, the flow of data is slow — it can’t keep pace with the frightening speed at which species are being lost. Major reports on the status of biodiversity loss typically output every three to five years. Scientific papers may take more than a year to publish, yet the IUCN Species Monitoring Specialist Group has stated: “Biodiversity data are fragmented, challenging to collect or access, difficult to use, and rarely available to decision makers in appropriate formats — highlighting that the current flow of biodiversity data cannot keep pace with the rate of biodiversity decline.”

We need to improve our current monitoring techniques to reverse biodiversity loss using tools that deliver reliable data, frequently, at landscape scale and a sustainable cost level.

What Tools Do We Currently Have?

Decades ago, most wildlife monitoring was done by direct observation, often over long periods, allowing researchers to build a very detailed picture of animal behavior and whereabouts. This provided some incredible insights into the natural world but was less applicable to monitoring large populations over landscape scales.

Over the last few decades, a common starting point for wildlife monitoring has been fitting instrumentation to the body of target species. Depending on what is fitted, and where, this approach can record daily movements, provide video feeds via a fitted camera, and, increasingly, produce data on some physiological parameters. Instrumentation typically takes the form of collars or harnesses carrying VHF or satellite telemetry receivers, RFID tags, and even implant tags — all of which require animals to be trapped and either physically or chemically immobilized.

The benefit of the approaches above is that an immediate flow of data may be delivered from the instruments directly to the operator. This is typically the case for a relatively short period and a small number of animals, as the cost of fitting and maintaining instruments precludes a landscape-scale survey. That may be all that is needed for some studies, such as using a small tag to track and help protect the migration route of a bird species. However, for landscape-scale monitoring to inform on species numbers and distribution, it is not cost effective.

VHF tracking collars for wildlife such as lions and elephants typically cost between $350 and $500 USD, while GPS and satellite tracking collars (including GSM and Iridium collars) can range from $650 to $3,200 USD. When one factors in the cost of capture and immobilization of the target animal for fitting of the instruments, surgery if necessary, the tags/instruments themselves, and employment of specialized follow-up personnel to regularly track the animals and replace any collars that fail, costs escalate considerably. The total cost of fitting a GPS collar to a lion, including follow-up monitoring and replacement, can range from around $4,000 to more than $10,000 annually.

However, cost isn’t the only, or even most important, consideration. A hidden cost, typically not reported in scientific publications of conservation monitoring and yet extremely valuable to science, is the impact of these invasive monitoring procedures on study animals.

Negative impacts of invasive approaches may even go unnoticed, especially when they occur after the monitoring period has ended. These include reduced fertility rates from repeated attempts to keep instrumentation functioning, changes in ranging behavior, reduced litter sizes, changes in dominance behavior, and many more. Most of these changes can be attributed to the stress involved in the capture and fitting of the instrumentation, carrying it, the risk of infection, and/or perceptions of individual weakness by unmarked conspecifics.

Lastly, a rarely acknowledged risk is the potential impact on the validity of collected data if the physiology of behavior of the target species is compromised. Invasive wildlife monitoring techniques — any elements of capture, handling, and tagging — can alter the behavior, physiology, and even survival of animals to such an extent that the data collected may not be representative or reliable. This raises the risk that conservation decisions based on these data could be flawed. The assumption of data reliability is often misplaced, as stress and injury from invasive methods can cause animals to behave atypically, invalidating the scientific validity of monitoring results.

While many conservation practitioners observe these effects, they are rarely investigated. Publication bias is a real challenge: authors and scientific journals favor ‘positive’ results. A 2018 study by Tensen found that research in wildlife and conservation is biased, with a tendency to publish studies with positive results, while negative findings are under-reported. This can not only distort the scientific record, but also hinder progress towards more effective wildlife monitoring.

Non-Invasive Technologies

More recently, a new class of monitoring methods is emerging that could potentially deliver true landscape-scale, ethically sound monitoring of wildlife.

These non-invasive techniques are gaining in popularity as technologies advance, and offer the opportunity to collect data cost effectively at scale, without compromising animal welfare. Each approach has different strengths, and together, often in combinations customized for each local environment, can provide a cohesive toolbox for monitoring most ecosystems.

Camera traps, eDNA, Passive Acoustic Monitoring, and drone and satellite monitoring have all grown in capability over the last decade, and continue to deliver exciting results with far less disturbance of natural populations. Each has strengths and challenges in different areas, but all remain relatively expensive and require specialized equipment and expertise for data collection.

Many of these techniques are based on the identification of species, or individuals, from a specific biometric or biological characteristic that can be measured for identification.

Now imagine if there was a biometric that was so ubiquitous, and easy to collect, that it could transform the monitoring of biodiversity.

Footprints have some unique qualities that make them ideal for the job. They are easy and cheap to collect, and observing them requires no special expertise or expensive equipment. Most significantly, footprints are ubiquitous over almost every ecosystem on earth where wildlife ranges. They’re effectively signals from nature, left on the ground and constantly refreshed by animal movement patterns.

Footprint Identification Technology (FIT)

Imagine you’re a conservation biologist and have been asked to protect rhinos from poachers in Africa. The first question you might ask is ‘where are the rhinos?’ You’d naturally then want to know how many there are in each location, and how far they range. The answer might typically be ‘we don’t really know,’ or ‘we often see this one rhino over there.’ But if you had a means of getting reliable data on numbers and distribution, you might, for example, be able to allocate anti-poaching resources efficiently, or deploy monitoring resources to those areas, or advocate for a protected area.

Footprint identification technology (FIT), developed by WildTrack, is drawn from a unique combination of the ancient and time-honored art and science of traditional trackers with cutting-edge analytics. This data helps answer bigger conservation questions, such as how to mitigate human-wildlife conflict by tracking the movement of specific problem animals. It supports anti-poaching efforts by monitoring at-risk individuals, and underpins fundamental ecological research by revealing patterns of movement and habitat use. FIT has also been used to identify the success of recolonization efforts of the Amur tiger in NE China.

FIT is an add-in to JMP Statistical Discovery software. It can accurately determine the species, individual identity, and sex of the animal. Using a robust cross-validated discriminant analysis and a user-friendly interface for feature extraction and data visualization, it can deliver an identification of the species, individual, and sex of the animal who made the track.

FIT in JMP software allows a user to import a footprint image, place points at key anatomical landmarks, take a series of measurements of lengths, angles, and areas, output them to a robust cross-validated discriminant analysis, and from there identify species, individual, sex, and age-class of the animal who made them — all within the JMP platform. In this example, we see the FIT model for cheetah identification.

Recently, FIT has integrated AI as an additional approach to footprint classification. This award-winning machine learning pipeline has the potential to greatly increase the speed of footprint classification, and process large volumes of data.

One of the challenges WildTrack faced was the difficulty in finding good footprints from tiny mammals like mice, rats, gerbils, shrews, etc. However, this group can often compose 30% of mammals in any ecosystem, and they react rapidly to environmental change, so being able to monitor them gives us an early warning of ecosystem integrity.

A solution came when WildTrack worked with Dr. Jody Tucker, a colleague from the U.S. Forest Service, who was using trackplates to monitor another small mammal, the Pacific Fisher. Trackplates are rectangular metal or wooden plates with a sooted surface that small animals walk over and leave tracks on. They’re particularly useful for animals who are too light to leave clear tracks, or for surfaces that are too hard or soft to hold tracks. The results were very encouraging and led them to think about using trackplates to collect footprints from tiny mammals.

Although FIT has impressive analytical and data visualization technology, thanks to JMP software and a machine learning pipeline, one of the most powerful features of the technology is community engagement. Most of WildTrack’s 25+ field projects over five continents engage the expertise of traditional trackers and communities who hold traditional ecological knowledge. FIT, because it uses tracks, is intuitive to expert trackers who are most qualified to locate tracks and whose interpretations in the field and lab are key to understanding our data.

Anyone who has been in the field with an experienced tracker — whether someone who has learned the art culturally or from secondary training — is immediately struck by the depth and power of their insights gleaned from seemingly useless smudges on the ground. This art and science, honed by evolution over hundreds of thousands of years, has forged our success as a species. Today, these skills can be found in many indigenous societies, and in the resurgence of interest in tracking as an art in the wilderness schools springing up all over the world. Tracks are found in almost every ecosystem, and can be easily, quickly, and cheaply collected by any citizen on the planet, particularly anyone with tracking expertise.

A Path Forward: Five Principles For Biodiversity Monitoring

Experience has taught us that ethical conservation practices lead to better scientific outcomes. As a global conservation community, we must embrace the following principles:

1. Prioritize non-invasive methods – Monitoring techniques should minimize disturbance to wildlife, ensuring that data collection does not negatively impact species survival or reproduction.
2. Consider the individual, not only the species – Conservation strategies must account for the well-being of individual animals because individuals make up the species. This is especially critical for critically endangered species.
3. Require post-study follow-ups – Scientific journals could, and should, mandate follow-up assessments of conservation interventions to ensure their effectiveness and ethical integrity.
4. Engage local communities – Indigenous and local communities possess invaluable ecological knowledge. Their participation should be central to conservation efforts.
5. Embrace a multi-disciplinary approach – The best conservation solutions integrate biology, data science, AI, and traditional knowledge for a holistic approach to biodiversity protection.

Imagine a world where collecting the most ubiquitous biometric from wildlife delivers truly democratized wildlife conservation, and everyone who has a smartphone can participate. There were an estimated 8 billion visits to the world’s protected areas in 2014. Imagine the data we could collect if even 1% (80 million) of those visitors snapped an image of even one animal track they saw.

WildTrack needs data reviewers, labellers, trackers, business leaders, software and hardware engineers, machine-learning experts, graphic designers — and more. Help them scale up their technology to deliver a sustainable, self-supporting model for wildlife monitoring globally, one footprint at a time. You can reach out to [email protected] for more information and details on how to participate.