Every year, one of nature’s most spectacular events unfolds across East Africa — the Great Wildebeest Migration. Millions of wildebeest, accompanied by zebras and gazelles, journey between Tanzania’s Serengeti and Kenya’s Maasai Mara in search of fresh pastures after the rains. This epic movement, spanning roughly 800 to 1,000 kilometres, sustains the region’s grasslands, nourishes predators such as lions and crocodiles, fertilises the soil, and supports countless species, as well as local livelihoods and tourism.

Now, for the first time, scientists are using satellites and artificial intelligence to count and track these herds from space — opening a new frontier in wildlife monitoring.

A New Lens on an Ancient Journey

The migration is the beating heart of the Serengeti-Mara ecosystem. Its scale and timing influence everything from predator numbers to vegetation patterns and water availability. Understanding how many animals participate each year is vital for conservation, tourism, and rangeland management. Yet, despite decades of study, precise population estimates have remained challenging.

For many years, aerial surveys have been the main tool for estimating the wildebeest population, which has been pegged at around 1.3 million animals. Aircraft fly over the plains in long, straight transects, counting animals within specific strips to extrapolate overall numbers. While reliable, this approach is labour-intensive, expensive, and limited by visibility and weather.

Recent advances in satellite imaging and artificial intelligence have opened up new possibilities. High-resolution satellites now capture images sharp enough to distinguish individual large mammals from hundreds of kilometres above Earth. Deep learning models — computer systems trained to identify patterns in massive datasets — can then detect and count animals automatically, a task that would take humans weeks or months.

Seeing the Migration from Orbit

A recent study used satellite imagery of the Serengeti-Mara ecosystem from 2022 and 2023, covering more than 4,000 square kilometres of migration routes. The images were collected at extremely high resolutions — just 33 to 60 centimetres per pixel — meaning each wildebeest appeared as only a few pixels on screen.

Two complementary AI models were applied: one focused on analysing individual pixels, while another identified objects based on shape and texture. When used together, these systems cross-validated one another’s detections, helping reduce errors. Images taken at the start and end of August captured different stages of the dry-season migration, with smaller herds earlier in the month and larger gatherings later.

Across both years, fewer than 600,000 wildebeest were detected within the surveyed dry-season range. While this number is lower than past aerial survey estimates, researchers caution that it does not necessarily signal a population decline. Factors such as animals sheltering under trees, or outside the imaged area, could explain the difference. Ground-based observations and GPS tracking data confirmed that the main herds were within the surveyed zone, suggesting the satellite snapshot was broadly representative.

This marks the first satellite-based census of the Serengeti-Mara migration, demonstrating that it is now possible to monitor large herds of mammals directly from space.

Complementing, Not Replacing, Traditional Surveys

Rather than replacing aerial surveys, satellite methods are designed to complement them. Coordinating both approaches can refine population estimates, identify seasonal shifts, and improve understanding of how herds move through the landscape over time.

Satellite monitoring is not without challenges. Cloud cover can obscure images, and acquiring ultra-high-resolution data can be costly. However, the advantages are significant. Satellites can capture vast regions in a single moment, providing a complete snapshot free from the uncertainty of extrapolated counts. Unlike aircraft, they can repeatedly image the same location, enabling near-continuous observation as migration unfolds.

As more high-resolution satellites are launched, it will become possible to monitor wildlife almost in real time, revealing trends in population and movement that were previously invisible.

Beyond Counting: Understanding Collective Behaviour

The potential of satellite technology extends beyond counting animals. By observing entire herds in motion, scientists can now study how collective movement emerges and evolves at a landscape scale.

The wildebeest migration is a classic example of “emergent behaviour,” where order arises without central coordination. Each animal responds to local cues — greener grass, the direction of neighbours, or predator presence — yet together, they create vast, coherent waves of movement across the savannah.

With repeated, high-resolution imaging, it is now possible to explore how these density waves form and travel, how herd spacing is maintained, and how collective dynamics influence ecosystem processes such as grazing and nutrient cycling. This could also shed light on how climate variability, human disturbance, or habitat changes alter the synchrony and scale of migration.

A New Era of Wildlife Monitoring

The integration of satellite imaging and artificial intelligence represents a turning point in how large-scale wildlife movements are studied. It enables scientists to combine ground truth data, aerial surveys, and space-based imagery to build a comprehensive picture of ecosystem health.

While challenges remain, the promise is immense: a world where the pulse of nature — from the thunder of hooves across the Serengeti to the surfacing of whales or the spread of seabird colonies — can be monitored, understood and protected in near real time.

By turning eyes in the sky toward the Earth’s living systems, humanity gains not only a new scientific tool but also a deeper appreciation of how interconnected life truly is.