AI and Satellites Team Up to Track Africa’s Great Migration

Each year, over a million wildebeest, accompanied by zebras and gazelles, travel in a vast circular journey between Tanzania’s Serengeti and Kenya’s Maasai Mara. Known as the Great Migration, this movement covers roughly 800 to 1,000 kilometers, following the seasonal pattern of rain and grass growth.

It’s more than a spectacle; it’s a vital ecological process. The herds aerate and fertilize the land as they move, shaping plant growth and influencing predator populations. Lions, hyenas, and cheetahs time their breeding and hunting around the migration. At the same time, millions of people in East Africa benefit through tourism, rangeland health, and local economies linked to this yearly cycle.

Because the migration acts as a central “engine” of the Serengeti-Mara ecosystem, knowing how many animals make the journey and whether that number is changing is crucial for conservation and land management.

A New Way to Count the Uncountable

For decades, wildebeest populations were estimated through aerial surveys. Aircraft fly along straight lines a few kilometers apart, and biologists count the herds visible from above. From those sample counts, they extrapolate a total population, usually around 1.3 million wildebeest.

While reliable, this method is costly, time-consuming, and weather- and terrain-dependent. To find a more scalable approach, a group of scientists turned to Artificial Intelligence (AI) and satellite imagery.

How AI Is Being Used

In 2022 and 2023, conservation biologists, data scientists, and remote sensing experts worked together to analyze over 4,000 square kilometers of satellite images covering the Serengeti-Mara region. The imagery was captured at a very high resolution (33–60 centimeters per pixel), enough to show individual animals as small clusters of pixels.

The researchers trained two deep learning models to detect and count wildebeest automatically:

A U-Net model, which analyzes pixels to identify potential animal shapes, and
A YOLO model, which detects and classifies individual objects.

By comparing results from both models, the team reduced bias and increased accuracy. Together, they produced the first satellite-based dry-season census of the Serengeti-Mara migration.

What the Results Showed

The AI models identified fewer than 600,000 wildebeest within the surveyed area. That figure is lower than traditional aerial estimates, but scientists caution against assuming a population decline. Differences could result from timing, coverage area, or detection limits between the two methods.

The next step is to run aerial and satellite surveys in parallel, allowing researchers to calibrate both techniques and build a clearer picture of population dynamics. Over time, combining these approaches could make wildlife monitoring faster, cheaper, and more consistent.

Why It Matters

Satellite monitoring offers unique advantages. Images can cover vast landscapes in a single snapshot, providing consistent data across borders and seasons. Unlike manned flights, satellites can revisit the same region multiple times a day, opening the door to near real-time wildlife tracking.

As technology improves, scientists hope to use similar methods to monitor other large-scale animal movements, from caribou in the Arctic to elephant migrations in southern Africa. It’s a step toward understanding how wildlife responds to climate change, human activity, and shifting ecosystems.

The Broader Impact

For conservationists, this isn’t about replacing traditional methods; it’s about expanding the toolkit. By integrating AI, satellites, and fieldwork, researchers can measure ecological trends with a level of precision once thought impossible.

The data gathered could also help local communities by improving land use planning, guiding sustainable tourism, and protecting the delicate balance between wildlife and people.

In essence, satellites and AI are providing a new vantage point, not just over Africa’s grasslands, but over how humanity can use technology to better understand and protect the natural world.