The Great Wildebeest Migration one of nature grandest spectacles, sees millions of wildebeest, zebras and gazelles travel between Tanzania’s Serengeti and Kenya’s Maasai Mara each year in search of fresh grazing lands. This 800-1,000 km journey fuels an entire ecosystem feeding predators, fertilising soil, sustaining grasslands and supporting local communities that depend on rangelands and tourism.
Understanding how many animals take part is crucial, as any change in their numbers can have ripple effects on predators, vegetation and livelihoods. Traditionally, scientists have relied on aerial surveys to estimate the wildebeest population, which has been pegged at around 1.3 million. Aircraft fly over the region in straight transects to count herds and extrapolate totals a method that, while effective, is time-consuming and costly.
Now, scientists are turning to space. In a recent study, researchers used high-resolution satellite imagery combined with artificial intelligence (AI) to detect and count wildebeest across the Serengeti-Mara ecosystem. The team, including biologists, machine-learning experts and remote sensing specialists, analysed satellite images from 2022 and 2023 covering more than 4,000 square kilometres.
Using deep learning models known as U-Net and YOLO, the researchers were able to identify wildebeest from above each animal appearing as just a handful of pixels. The models detected fewer than 600,000 wildebeest within the dry-season range. While this figure is lower than earlier aerial estimates, scientists caution that this doesn’t necessarily indicate a population drop. Some animals could have been under tree cover or outside the imaged areas, and more comparative studies are needed to understand differences between counting methods.
To ensure the data’s accuracy, the survey area was cross-checked using GPS tracking from collared wildebeest and on-ground monitoring. The findings mark the world’s first satellite-based dry-season census of the Serengeti-Mara migration. Rather than replacing aerial surveys, researchers say satellite-based monitoring will complement them, helping refine both approaches for more precise population estimates.
Satellites offer key advantages they can capture vast areas in a single snapshot and eliminate the guesswork that comes with extrapolating smaller samples. Although satellite imagery is expensive and sometimes hampered by cloud cover, advancements in high-resolution imaging now allow scientists to observe the same area multiple times a day, bringing real-time wildlife monitoring closer to reality.
Beyond counting animals, this technology opens a new chapter in understanding how herds move together. The migration is a phenomenon of collective behaviour there no leader, yet thousands move in synchrony based on simple cues like vegetation and neighbour movement. With satellite data, scientists can now study how these movements spread like waves across landscapes and how they influence the broader ecosystem.
The study shows that satellites and AI are not just tools for population tracking but keys to unlocking the mysteries of coordination and survival in one of Earth’s most extraordinary natural events.