Does restoring apex predators to food webs restore ecosystems? Large carnivores in Yellowstone as a model system

• Published in: Ecological Monographs
• Main Authors: Hobbs, N. Thompson, Johnston, Danielle B., Marshall, Kristin N., Wolf, Evan C., Cooper, David J.
• Other Authors: National Science Foundation
• Format: Article in Journal/Newspaper
• Language: English
• Published: Wiley 2024
• Subjects: Ursus arctos
• Online Access: http://dx.doi.org/10.1002/ecm.1598; https://esajournals.onlinelibrary.wiley.com/doi/pdf/10.1002/ecm.1598

Abstract Modification of food webs is a frequent cause of shifts in ecosystem states that resist reversal when the food web is restored to its original condition. We used the restoration of the large carnivore guild including gray wolves ( Canis lupis ), cougars ( Felis concolor ), and grizzly bears ( Ursus arctos horribilis ) to the northern range of Yellowstone National Park as a model system to understand how ecosystems might resist reconfiguration after the restoration of apex predators to the food web. The absence of wolves, cougars, and grizzly bears for nearly a century from the northern range was the primary cause of dramatic changes in riparian plant communities. Willows ( Salix spp.) were suppressed in height by intense browsing by the dominant herbivore, elk ( Cervus canadensis ). The loss of activity by beavers ( Castor canadensis ) coincided with the loss of tall willows. We hypothesized that intense elk browsing interrupted the mutualism between willow and beavers: ecosystem engineering by beavers was a critical component of willow habitat and tall willows were a critical component of habitat for beavers. This interruption made riparian communities resilient to the disturbance caused by the restoration of apex predators. We hypothesized further that reductions in elk browsing attributable to reductions in elk population size were not sufficient to prevent the suppression of willow growth. To test these hypotheses, we conducted a 20‐year, factorial experiment that crossed simulated beaver dams with the exclusion of browsing. We found that willows grew to heights expected for restored communities only in the presence of dams and reduced browsing. Willows experiencing ambient conditions remained well below this expectation. We found no difference in heights or growth rates of willows in experimental controls and willows in 21 randomly chosen sites, confirming that the results of the experiment were representative of range‐wide conditions. A reorganized community of large herbivores was implicated in ...