Author: Francis Wamonje, PhD

In Kenya’s vast and fragile drylands, a new kind of science is helping restoration efforts take root. Across the arid rangelands of Laikipia County, researchers have turned to environmental DNA (eDNA) to track the early stages of ecological recovery, offering a glimpse beneath the surface into the unseen biodiversity that shapes soil health, vegetation, and livelihoods.

The Lower Naibunga Community Conservancy, located in one of Kenya’s most degraded arid and semi-arid landscapes (ASALs), is now the site of the country’s first comprehensive application of eDNA metabarcoding for restoration monitoring. By reading the genetic traces left behind in soil, the study reveals how restoration efforts—particularly water bunds and reseeding—are quietly transforming degraded land.

Why Monitoring Matters in Dryland Restoration

Kenya’s ASALs cover more than 80% of the country’s land area and support around 10 million people and 90% of the nation’s wildlife. Yet decades of overgrazing, tree loss, and worsening drought have left many areas deeply degraded. Restoration is a national priority, with Kenya committing to ambitious goals under frameworks such as the UN Decade on Ecosystem Restoration and the Bonn Challenge.

However, monitoring success has long posed a challenge. Traditional field surveys require time, taxonomic expertise, and resources that are often scarce, especially in remote, resource-limited settings. Worse still, such methods often miss below-ground biodiversity, which plays a crucial role in restoring ecosystem function.

This is where eDNA metabarcoding offers a powerful alternative. By analysing DNA fragments shed by organisms into their environment—through skin, roots, excreta, or decaying material—researchers can detect a broad range of species from a small sample of soil or water. It is rapid, non-invasive, and cost-effective at scale.

Restoring the Land: The Role of Water Bunds

The Laikipia study focused on semicircular water bunds, simple earthworks designed to trap rainfall and reduce erosion. Between February and March 2023, around 5,000 bunds were constructed across six sites with support from the Wyss Academy for Nature and active participation from local communities.

These structures slow water runoff, allowing it to soak into the soil. This creates small zones of improved moisture where vegetation can grow, helping to kick-start natural processes. To accelerate this recovery, the restoration team planted two drought-adapted grasses: buffel grass (Cenchrus ciliaris) and Maasai lovegrass (Eragrostis superba). Both are native to East Africa and known for stabilising soil, improving forage quality, and tolerating arid conditions.

The question the researchers set out to answer was simple: Are these interventions working?

How eDNA Metabarcoding Works

To find out, 18 soil samples were collected from georeferenced water bunds across the six study sites. Care was taken to prevent contamination, and the samples were kept cold and transported to the laboratory for analysis.

The team analysed two genetic markers:

• 16S rDNA, which identifies bacteria and archaea in the soil.
• rbcL, a chloroplast gene used to detect and classify plant species.

The sequencing process enabled researchers to match these markers to known species or genera in reference databases, providing a snapshot of the organisms living in or passing through each site.

What the DNA Revealed

Soil Microbes

The bacterial analysis identified over 19,000 unique DNA sequences, representing 40 microbial phyla and 554 genera. Among them were plant growth-promoting rhizobacteria (PGPR)—beneficial soil microbes that enhance plant nutrient uptake, help suppress disease, and support drought tolerance. These included:

• Bacillus circulans and Bacillus koreensis, known for nutrient solubilisation and stress tolerance.
• Bradyrhizobium elkanii and B. yuanmingense, nitrogen-fixing bacteria that help convert atmospheric nitrogen into forms that plants can absorb.

Also detected was Arthrobacter alpinus, a bacterium that breaks down organic sulphur compounds, suggesting early signs of natural soil remediation and ecosystem function.

Vegetation

The plant DNA data revealed 814 unique sequences across 36 genera. Buffel grass (Cenchrus ciliaris) was the most abundant species across all sites, indicating strong establishment success. However, Eragrostis superba, though known to be present, was not detected—most likely due to gaps in the global DNA reference databases rather than its absence on the ground.

The analysis also revealed several native or naturalised plants with ecological or cultural importance:

• Zaleya pentandra, a hardy ground cover that helps stabilise soil.
• Leonotis leonurus (wild dagga), valued for its medicinal properties and pollinator support.
• Food-producing legumes such as Arachis duranensis (wild peanut) and Cicer arietinum (chickpea).

However, the study also flagged emerging threats. DNA from invasive or toxic species was present, including Heliotropium europaeum (European heliotrope), which is poisonous to livestock, and Eriochloa villosa, a highly competitive grass that can displace native species.

Restoration in Action: What the Results Show

This study demonstrates that even within a year, restoration interventions—when supported by water retention structures and seeding—can begin to regenerate biodiversity in degraded drylands.

The detection of microbial communities known to promote soil health and support plant growth points to promising early recovery. The successful establishment of planted grasses, such as buffel grass, alongside the reappearance of ecologically valuable native species, suggests that bunds are not only holding water but also laying the foundations for broader ecological regeneration.

The discovery of forage species, medicinal plants, and nitrogen-fixers also highlights how restoration can strengthen local livelihoods, supporting grazing, food resilience, and traditional knowledge systems.

At the same time, the presence of invasive or harmful plants reminds us that restoration must be actively managed. The ability of eDNA to pick up such species early on adds to its value as a risk detection tool.

Limitations and Future Directions

The study also highlights areas for improvement. Gaps in global DNA reference databases, particularly for African plants, limit taxonomic precision. The absence of E. superba from the dataset is likely due to such limitations. Expanding these databases will be key to improving the accuracy of future monitoring.

In addition, while eDNA can tell us what species are present, it cannot measure structural features such as plant height, canopy cover, or reproductive success. Nor does it directly assess animal presence or behaviour. For these reasons, the researchers recommend combining eDNA with conventional monitoring methods—such as field surveys, photographic plots, and remote sensing—to build a more complete picture over time.

A Model for Scalable, Science-Based Restoration

What makes this study especially valuable is its demonstration of what is possible when local knowledge is combined with modern molecular science. The water bunds were built by communities using locally understood techniques. The biodiversity data were generated using frontier tools. Together, they provide a model that is practical, affordable, and scalable across similar landscapes.

As Kenya and many other countries work towards large-scale restoration goals, especially under the UN Decade on Ecosystem Restoration, this kind of monitoring will be essential. Without credible, high-resolution data, it is not easy to assess impact, attract investment, or adapt restoration strategies to changing conditions.

This research demonstrates that eDNA metabarcoding provides a viable and effective means to bridge that gap, offering new transparency in how we measure recovery and new accountability in how we manage it.

In these landscapes where life has been stripped back to its basics, recovery is possible. And with the right tools, we can now see it happening—not just above the ground, but deep within the soil.

Invasive species remain one of the most urgent ecological and economic threats of our time. They disrupt ecosystems, reduce biodiversity, and cause billions in damage. Among them is the emerald ash borer (Agrilus planipennis, or EAB), a small, metallic-green beetle that has devastated ash tree populations across North America and is now advancing through parts of Europe.

Detecting pests like EAB early, when they are still rare and relatively contained, is critical. But this is easier said than done. Much of the EAB life cycle occurs beneath tree bark, making it almost invisible until the damage is already done. Traditional survey methods struggle to pick up early infestations. Recent research, however, suggests that environmental DNA (eDNA) could offer a much-needed breakthrough.

Why eDNA Offers a New Hope

The emerald ash borer has caused more than $2 billion in damage in the United States alone. Its larvae bore through ash trees, disrupting the transport of water and nutrients and ultimately killing the host. Since its accidental introduction in the 1990s, EAB has left millions of dead trees in its wake. This loss carries broader ecological consequences, as ash trees play a vital role in supporting wildlife and stabilising soil.

The key to limiting the spread of invasive pests is early detection. But conventional tools—visual surveys, baited traps, and girdled “trap trees”—often prove too slow or imprecise. These methods are labour-intensive, seasonally restricted, and frequently miss low-density populations.

Environmental DNA offers an alternative. All living organisms shed genetic material into their surroundings through skin, faeces, saliva, or, in the case of EAB, larval feeding under bark. By collecting samples from tree tissue, soil, or water, researchers can screen for a species’ DNA and confirm its presence without needing to find the actual organism.

Testing the Method: Tree Cores and DNA

In the study, scientists trialled an eDNA-based detection method for EAB at two contrasting field sites: one with high infestation levels in New Jersey and another with low-level presence in New Hampshire.

In New Jersey, green ash trees (Fraxinus pennsylvanica) showing clear signs of decline—such as canopy dieback—were sampled, although visible symptoms like bark splits or beetle exit holes were deliberately avoided. This allowed the method to be tested in trees that were visibly unwell, but not obviously infested.

In New Hampshire, where EAB populations were thought to be low, white ash trees (Fraxinus americana) were categorised into three levels of visible damage: none, light, and moderate. This site provided a tougher test of the method’s sensitivity.

Tree core samples were taken using a Haglöf increment hammer, which extracts small wood cylinders with minimal harm to the tree. Two cores were collected from each tree—one from the north side, one from the south—at chest height. To prevent contamination, the tool was flame-sterilised between uses, and control samples were taken from nearby non-host species such as oak and birch.

Samples were frozen and later analysed in the lab using a highly specific qPCR (quantitative polymerase chain reaction) assay designed to detect EAB DNA.

What the Results Showed

In New Jersey, the eDNA method successfully detected EAB DNA in 64% of sampled trees during peak summer months—an encouraging result for an early-stage technique. It confirmed that eDNA from larvae feeding inside the tree can be recovered and identified from small core samples.

In contrast, no positive detections were made in New Hampshire, despite some trees showing signs of light or moderate damage. This may reflect genuinely lower pest densities, but it also highlights the influence of seasonal timing. The New Hampshire sampling was done earlier in the year, when larvae were less active and DNA concentrations were likely lower.

Despite these limitations, the study represents a significant step towards the use of eDNA for forest pest surveillance.

Why eDNA Matters for Forest Managers

Compared to traditional methods, eDNA offers a number of advantages:

• Non-destructive: Collecting tree cores causes far less damage than methods like girdling, which kill the tree to attract beetles.
• Sensitive: eDNA can detect small traces of DNA, making it suitable for identifying early-stage or low-density infestations.
• Efficient: Sampling is quick and can be done by a single person. The material can be processed later, providing flexibility in field operations.
• Extended detection window: eDNA signal is strongest later in the growing season, offering a longer time frame than some conventional methods.

Together, these advantages could make eDNA an essential tool for large-scale forest monitoring, particularly when surveillance resources are limited.

As promising as eDNA is, there are challenges to overcome before it can be widely adopted in forestry settings:

1. Seasonal variation: DNA concentrations within the tree fluctuate depending on larval activity and sap flow. Sampling outside the optimal window risks false negatives.
2. Low-density detection: As seen in New Hampshire, detecting sparse populations remains difficult. More work is needed to refine sampling protocols and increase sensitivity.
3. Understanding DNA movement: It’s still unclear how far EAB DNA travels within the tree. If it remains localised near feeding sites, sampling strategies will need to account for that.
4. Head-to-head comparisons: Rigorous studies comparing eDNA with traditional survey methods will help determine when, where, and how each should be used.

Final Thoughts: Beyond Ash Trees: A Broader Role for eDNA

While this study focused on EAB, the implications are much broader. The same approach could be adapted to detect other invasive wood-boring insects, such as the Asian longhorn beetle, or even fungal and microbial pathogens.

For biosecurity agencies and forest managers, eDNA could be the difference between containment and costly, long-term control. With global trade and climate change increasing the risk of biological invasions, having fast, sensitive tools like eDNA will be vital for early warning and response.

It’s important to see eDNA not as a replacement, but as a complement to existing tools. No single method will suit all species or scenarios. But used together, eDNA and conventional approaches can offer a more complete, layered surveillance system—capable of both broad screening and targeted follow-up.

The use of environmental DNA to detect hidden tree pests like emerald ash borer marks a powerful shift in how we monitor our forests. It opens new doors for early detection, rapid response, and smarter resource allocation.