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

Annual browsing intensity in the Hobbs et al. plots was rarely below 0.25 on average, and frequently exceeded 0.50 for individual plants.

“In contrast, annual browsing intensity in the Hobbs et al. plots, defined as the proportion of shoots browsed, was rarely below 0.25 on average, and frequently exceeded 0.50 for individual plants (see Fig.12and 13 inHobbs et al. 2024).”

Hobbs et al. concluded that carnivore recovery produced a relatively weak trophic cascade.

“In contrast, Hobbs et al. concluded that carnivore recovery produced a relatively weak trophic cascade.”

In a landscape where willow recovery is highly variable, omitting context allows isolated successes to stand in for more complex and uneven outcomes.

“In a landscape where willow recovery is highly variable (e.g., Hobbs et al. 2024), omitting such context allows isolated successes to stand in for more complex and uneven outcomes.”

This structural bias, amplified by increasing height but persistent asymmetry, likely contributed to Ripple et al.'s inference of a strong trophic cascade, despite underlying trends in willow height and browsing that align more closely with Hobbs et al.'s conclusion of a relatively weak one.

“This structural bias, amplified by increasing height but persistent asymmetry, likely contributed to Ripple et al.'s inference of a strong trophic cascade, despite underlying trends in willow height and browsing that align more closely with Hobbs et al.'s conclusion of a relatively weak one.”

These shortcomings explain the apparent conflict with Hobbs et al. (2024), who found evidence for a relatively weak trophic cascade based on the same height data and a long-term factorial field experiment.

“These shortcomings explain the apparent conflict with Hobbs et al. (2024), who found evidence for a relatively weak trophic cascade based on the same height data and a long-term factorial field experiment.”

Thirty-eight percent of the 53 willows used to construct the model were Salix lutea, a species absent from the Hobbs et al. dataset, and 11% were sampled outside northern Yellowstone.

“Thirty-eight percent of the 53 willows used to construct the model were Salix lutea, a species absent from the Hobbs et al. dataset, and 11% were sampled outside northern Yellowstone.”

Hobbs et al.'s interpretation was based on the same height data as well as a factorial field experiment that manipulated browsing pressure and stream hydrology.

“Their interpretation was based on the same height data as well as a factorial field experiment that manipulated browsing pressure and stream hydrology.”

Although browsing intensity declined through 2010, it increased again in later years, even as some willows grew taller.

“Although browsing intensity declined through 2010, it increased again in later years, even as some willows grew taller.”

In 2001, willows in the Hobbs plots were short, heavily browsed, and structurally deformed, conditions inconsistent with the half-ellipsoidal crown architecture assumed by the model.

“In 2001, willows in the Hobbs plots were short, heavily browsed, and structurally deformed-conditions clearly inconsistent with the half-ellipsoidal crown architecture assumed by the model.”

The strong cascade inferred by Ripple et al. does not reflect different empirical patterns, but rather a circular reinterpretation of the same height data that underpinned Hobbs et al.'s finding of a relatively weak trophic cascade.

“This helps explain the conflicting conclusions: the strong cascade inferred by Ripple et al. does not reflect different empirical patterns, but rather a circular reinterpretation of the same height data that underpinned Hobbs et al.'s finding of a relatively weak trophic cascade.”