In a landscape as vast and dynamic as Tsavo, where wildlife and human communities share space, this question isn’t just academic—it’s critical. As human populations grow and land use changes, incidents of human–elephant conflict (HEC) have become increasingly frequent and severe. Preventing human–wildlife conflict is one of Tsavo Trust’s key focuses, and we work daily to ensure that both people and elephants can coexist across this shared landscape. A new study from Taita Taveta County offers a powerful tool to get ahead of the problem: risk maps built on advanced species distribution modelling.
At Tsavo Trust, we know that data-driven insights like these can be game-changers in our ongoing efforts to protect elephants, especially Super Tuskers, and to support the people who live alongside them.
Understanding the conflict
Over 70% of Kenya’s wildlife lives outside National Parks, often in human-dominated landscapes. In these areas, particularly around the borders of Tsavo East and West National Parks, elephants frequently encounter farms, homes, and infrastructure. These encounters can lead to:
-
Damage to crops and property
-
Threats to human safety
-
Retaliatory killings of elephants
Between 2004 and 2018 in Taita Taveta County alone, 97% of reported human–elephant conflict cases involved crop raiding, threats to human safety, or property destruction. These incidents take a toll on both livelihoods and wildlife conservation efforts.
How Science Can Help
In response to this escalating challenge, researchers from the University of Helsinki and Kenya’s Wildlife Research and Training Institute developed a new approach: using ensemble species distribution models to predict where human–elephant conflict is most likely to occur.
Put simply, this method layers multiple predictive algorithms to create the most accurate, balanced “risk maps” possible. Instead of just asking where elephants are, the models ask: where are elephants likely to interact with people—and under what conditions?
Key findings from the study include:
-
Proximity to houses and croplands is the strongest predictor of conflict.
-
Risk is highest near human settlements, particularly between 3–7 km outside protected area boundaries.
-
Low-risk areas are primarily within national parks, while high-risk zones cluster near farmland, settlements, and water sources.
But beyond these expected results, the study also uncovered less obvious—and highly useful—patterns:
-
Water availability matters more than rivers: Conflict risk is more strongly associated with proximity to waterholes than to rivers. In Taita Taveta, many rivers are seasonal and dry for much of the year, meaning waterholes—especially borehole-fed and solar-pumped ones—play a greater role in attracting elephants.
-
Vegetation density is a double-edged sword: Areas with higher vegetation cover (as measured by Enhanced Vegetation Index) showed a curvilinear relationship with conflict—suggesting that moderately green zones (like farms and woodland edges) attract elephants more than dense forest or open scrub.
-
Different models yield different risks: Algorithms like Maxent may overpredict conflict zones, while others underpredict them. By combining them into an ensemble model, the study offers a more realistic, balanced view—one that’s especially valuable for conservation practitioners trying to allocate limited resources.
By visualising this data spatially, the resulting maps don’t just show where conflicts are likely to happen—they help explain why. And that’s the kind of insight we can use on the ground.
What This Means for Tsavo Trust
For those of us working in the field every day, this research is more than an academic exercise—it has real, practical applications.
At Tsavo Trust, we’re already integrating this kind of spatial thinking into our work. Understanding where human–elephant conflict is most likely to occur helps us act before it happens:
-
Aerial reconnaissance patrols allow us to monitor elephant movement across the landscape, identify potential conflict zones early, and intervene when necessary.
-
Community conservancies, like the Kamungi Conservancy, create buffer zones that protect both wildlife and farmland by maintaining safe corridors for elephant movement.
-
Water provision projects reduce pressure during dry seasons, when elephants are more likely to raid farms in search of water.
-
Community engagement and education ensure that local people are part of the solution—because long-term conservation depends on local ownership.
In addition, our 10% Fence Plan offers a targeted, cost-effective solution: installing elephant exclusion fences around smallholder farms to protect high-value crops. These living fences not only prevent crop loss but also transform household economies.
Take Mary, a farmer in Kamungi:
“For years, most of our crops were destroyed by elephants. Since Tsavo Trust installed the elephant exclusion fence, we can finally plant and harvest without fear,” she shares with relief.
This new modelling approach aligns closely with our on-the-ground experience. It reinforces what we’ve long observed: preventing conflict requires a deep understanding of both elephant behaviour and how people use the land. And it confirms that well-targeted, community-driven interventions—whether guided by data or shared knowledge—can make a measurable difference.
Looking ahead: Smarter conservation for all
The value of this study lies not just in its maps, but in its message: we can use data to protect both elephants and people. By identifying high-risk zones, we can direct resources where they’re needed most—be it deploying ranger teams, installing early-warning systems, or working with farmers on land-use strategies.
Most importantly, it highlights the importance of working together—scientists, conservationists, government agencies, and communities—to build a future where elephants and people can thrive side by side.
At Tsavo Trust, we remain committed to that goal. Learn more about how you can help support our efforts.