How Animal Footprints Can Reveal the Health of an Entire Ecosystem

That question sits at the heart of new work led by Dr. Zoë Jewell’s team at the Duke University Nicholas School of the Environment. For decades, scientists assumed that cryptic small mammals are simply too similar to tell apart without DNA. Visual identification often fails, while DNA testing can be expensive, slow, and invasive for wildlife.

The team challenged that assumption using something far simpler: footprints. By combining high-resolution footprint images with morphometry software, they trained a computer model to tell apart two nearly identical species, the Eastern Rock sengi and the Bushveld sengi. The model analyzed more than 100 measurements and focused on nine key footprint features. In field tests across South Africa, it identified the two species with an impressive 94 to 96 percent accuracy, based on data from 18 Bushveld sengis at Tswalu and 19 Eastern Rock sengis from multiple sites.

As reported by ScienceDaily, this approach offers a practical, non-invasive alternative to DNA testing. The full scientific method and results are detailed in Frontiers in Ecology and Evolution, pointing to a future where footprints alone can reveal who lives in an ecosystem, without capturing or disturbing the animals.

Footprints as Real-Time Biodiversity Data

At stake is not just a clever trick, but a social and practical shift toward ethical, tech-enabled conservation and broader citizen science. The footprint workflow provides a scalable, real-time signal of species presence and range, reducing the need for capture or sampling and enabling routine biodiversity surveillance across landscapes managed by universities, reserves, and local communities.

From Footprints to Future Frontiers

Looking ahead, the team plans to expand the method to additional species by training models on more footprint datasets, potentially turning footprints into a universal, non-invasive monitoring toolkit. The approach aligns with a growing public expectation for non-invasive wildlife monitoring and AI in ecology, with researchers envisioning coastal, savanna, and forest ecosystems becoming readouts of health rather than hard-to-visit study plots.

With the groundwork laid by Jewell and colleagues, the era of the DNA-only toolkit may be ending; footprints are becoming the real-time sensors of our living world.

Key Takeaways

• Non-invasive footprint analysis can distinguish cryptic small mammals with up to 96% accuracy using nine footprint features from over 100 measurements.
• The method offers real-time biodiversity monitoring at scale, reducing reliance on DNA sampling.
• Researchers aim to extend the approach to additional species, potentially creating a universal monitoring toolkit.