Sheep are rarely dangerous to skiers, but otherwise they have a lot in common with avalanches. That’s what physicists say after mathematically modeling the ungulates’ behavior (and staying well out of their path).
Francesco Ginelli, who researches complex systems at the University of Aberdeen in Scotland, had already studied flocks of birds and schools of fish. But he was curious to learn what was different about the movement of sheep or other grazers. Animals like these have a simple goal, Ginelli says: “They need to eat without being eaten.”
Ginelli and his colleagues started their investigation by simply watching some merino sheep. At an experimental farm in the south of France, the researchers led herds of 100 female sheep into square enclosures 80 meters on each side. For up to 3.5 hours at a time, they let the sheep roam around their pastures. Meanwhile a camera snapped high-resolution photos from overhead, one picture per second.
Then researchers digitized this footage, going frame-by-frame and marking each sheep’s position by hand. The herd’s movements looked strikingly like an avalanche, Ginelli says.
Most of the time, a herd of sheep spreads slowly across an open area. The animals eat as they go. But every once in a while, a sheep near the periphery notices that it’s too isolated from the rest of the group. It suddenly sprints back toward the center, where it will be safer from predators. There’s no event that seems to trigger this movement, no “Baa!” of alarm; it comes out of nowhere. As the first sheep runs, others start to follow it, gathering mass like cascading pile of snow. Then, all at once—and again with no discernible signal—the animals stop moving. They continue grazing as before, now in a densely packed herd. (You can watch this behavior here.)
The researchers tried to build a computer model that would make a digital herd unpack and re-pack itself just like a real one does. They succeeded by creating a set of rules for a digital sheep, Ginelli says:
First, graze. You may either walk slowly or stand still while you eat grass. Try to align yourself with your close neighbors.
Next, freak out (maybe). Run, especially if you see a close neighbor running. This will get the whole herd back into its tightly packed state. In the model, the switch to freak-out mode is partly random, and partly influenced by how close a sheep’s neighbors are and what they’re doing.
This model works “fairly well” at creating artificial sheep herds that move like real ones, Ginelli says. It matters to researchers who are trying to understand how groups of animals behave, and how those behaviors evolved. “The origin of cooperative behavior in social groups is a very important question in evolutionary biology,” Ginelli says.
Merino sheep, it turns out, organize themselves pretty much like an avalanche does. This knowledge might help scientists grasp how other communities of grazers operate. Soon they may discover cow tsunamis or goatquakes—but probably not sharknados.
Image: by B4bees (via Flickr)
Ginelli, F., Peruani, F., Pillot, M., Chaté, H., Theraulaz, G., & Bon, R. (2015). Intermittent collective dynamics emerge from conflicting imperatives in sheep herds Proceedings of the National Academy of Sciences DOI: 10.1073/pnas.1503749112