What if wolves returned to a park?

Wolves are hunters. In Yellowstone park, wolves mostly hunt elk.

For a long time there were no wolf packs in the park. What if the wolves came back?

A top predator is a hunter that no other animal hunts. In Yellowstone National Park, in the United States, that hunter is the gray wolf, and its main food in winter is elk.

Early park managers killed wolves, and the last pack in Yellowstone was killed in 1926. In 1995, wolves were brought back. This lab asks: when a top predator returns to a park, does it change only the animals it eats, or could the change spread further, all the way down to the plants?

Predator control in the late 1800s and early 1900s targeted wolves in Yellowstone; at least 136 were killed in the park between 1914 and 1926, the year the last pack was killed. After the Endangered Species Act and years of planning, 31 wolves from western Canada were released in 1995 and 1996.

When a top predator is removed or restored, effects can pass down a food web: predator to prey, prey to plants. Ecologists call these indirect effects of predators on plants a trophic cascade. This lab starts from a model park without wolves, restores them, and then asks how well that simple picture matches what has been measured in the real park.

In Olly's model park, the wolves are gone.

Nobody hunts the elk now. There could be more elk. More elk eat more grass and willow leaves.

Our model park is a simple forest food web based on Yellowstone, with 13 kinds of living things. In this lab it starts with the gray wolf taken out, like the park before 1995.

In the model, the wolf's only food is elk, so with no wolves the elk could become more common. More elk eat more grass and more willow and aspen, so those plants could become fewer. Then insects, deer mice, snowshoe hares, beavers and songbirds could find less food, or fewer willows to nest in. That is 8 species affected by losing one hunter.

The lab's starting world is the model's forest web with the gray wolf node removed. The model links wolves only to elk, their main prey in Yellowstone, so the top-down rule marks elk as possibly more common. Rule 4 marks the elk's foods, grasses and willow and aspen, as possibly fewer. Rule 5 then marks insects, deer mouse, snowshoe hare, beaver and songbird as finding less food (the songbird through its willow nesting places).

No consumer loses all its food, so none could disappear: 1 more common, 2 fewer and 5 with less food, 8 species affected in all.

Now Olly puts the wolves back into the food web.

The wolves hunt the elk again. In the model, nobody is short of food now. The web is back the way it was.

Put the gray wolf back into the model's web and every one of those 8 changes disappears: the readout says 0 species affected. The elk have a hunter again, so the model no longer marks them as more common, and the willows are no longer marked as fewer.

Try the comparison view to see both parks side by side. Then try taking out other animals. Removing the elk makes the model wolf lose all its food, and removing the coyote affects 10 species.

Restoring the wolf node returns the model to its intact web: gone 0, more 0, fewer 0, less food 0, affected 0. In this model, then, restoration simply reverses the removal cascade, link for link.

Other single removals show how the web is wired. Removing the elk affects 4 species: the wolf could disappear, grasses and willow could become more common, and the grizzly bear could find less food. Removing the coyote affects 10 and the grizzly bear 9. The counts are of species that could change, never of animals.

Wolves eat elk. Elk eat willows. So wolves and willows are linked!

Our model is very simple. It shows who eats whom, not how many animals there are.

Each link in a food web is a meal, so a change at the top can pass down, link by link: wolves eat elk, elk eat willows, and many small animals use willows. That chain is called a trophic cascade.

The model is a simplification. It counts only who eats whom, so it never says how many elk or willows there are. Its wolves eat only elk, but real wolves also eat deer and bison. Its only other elk hunter is the grizzly bear, and it leaves out the park's black bears and cougars, people who hunt elk outside the park, and weather such as drought and hard winters.

The cascade comes from two of the model's five rules: removing a hunter releases its prey (rule 3), and released prey eat their foods down (rule 4), which then leaves less food for others (rule 5).

The model's limits matter more here than in most labs. It is qualitative and gives no population sizes or rates. It follows top-down effects only one level, ignores diet switching, competition such as wolves killing coyotes, disease and weather, and gives elk only two predators, wolves and grizzly bears. The National Park Service lists grizzly and black bears, coyotes and cougars among the park's large predators, adds human hunting outside the park, and says this complexity makes long-term elk trends hard to project. It shows the textbook cascade; it cannot test whether that cascade is the main thing happening in the real park.

Wolves came back to Yellowstone in 1995.

Now the park's northern elk herd is smaller than it was then, and some willows have grown taller. Scientists are still finding out why.

In 1995 and 1996, 31 wolves from western Canada were released in Yellowstone. The northern elk herd's winter count, about 17,000 when the wolves arrived, fell to 3,915 in early 2013, and it was 6,673 in 2022. The Park Service links the fall to wolves, cougars, bears, hunting and drought.

In some places, fewer willow, aspen and cottonwood shoots are being nibbled, and willows and cottonwoods have grown taller. The Park Service says there are "some indications" that wolves and other large carnivores are helping, but others argue that fewer elk for any reason changed the browsing. A 2022 study found that measuring only the tallest young aspen had made their regrowth look 4 to 7 times stronger than random sampling showed; random sampling still showed an effect, but a weaker one. Research is still under way.

NPS records 31 wolves released in 1995–1996 (14 and 17), plus 10 relocated from northwestern Montana in 1997. Wolves have contributed to lower survival of elk calves and older females and influenced elk habitat use. The northern herd's winter count was about 17,000 in 1995, below 10,000 in 2003, 3,915 in early 2013 and 6,673 in 2022; NPS attributes the decline to carnivore recovery, hunter harvest and drought-related effects on pregnancy and survival.

On plants, the NPS wording is cautious: there are "some indications" that elk–carnivore interactions contribute to a release of willows and other woody plants on the northern range, with fewer browsed leaders in some areas and significantly taller cottonwood and willow. Others argue that lower elk densities from predators, hunters and weather altered browsing. Brice, Larsen and MacNulty (2022) showed that sampling only the tallest young aspen overestimated aspen regeneration 4–7 times compared with random sampling; random sampling still described a trophic cascade, but a weaker one. Research continues on the relative effects of climate, hydrology, predation and herbivory.