Large‐scale responses of herbivore prey to canid predators and primary productivity

AIM: The primacy of top‐down (consumption) and bottom‐up effects (primary productivity) as forces structuring ecological communities is a controversial topic. The exploitation ecosystems hypothesis (EEH) was invoked to explain biogeographical trends in plant and consumer biomass, and differs from the top‐down/bottom‐up dichotomy by predicting that the relative strength of these processes will vary along gradients of primary productivity. Here we test the prediction of the EEH that herbivore biomass should increase with increasing primary productivity where predators are rare, but show a negligible response to primary productivity where predators are common due to population regulation by predators. LOCATION: Boreal and temperate regions of North America and Eurasia, and deserts of Australia. TIME PERIOD: 1970–2016. MAJOR TAXA STUDIED: Cervids and kangaroos. METHODS: We obtained abundance indices of cervids at 42 locations from the literature and conducted spotlight surveys at 27 locations to derive estimates of kangaroo abundance. For analyses, herbivore abundances were converted to biomass per km². We tested our prediction using linear mixed effects models. RESULTS: Herbivore biomass showed divergent responses to increasing primary productivity and the abundance of canid predators (grey wolves, Canis lupus/dingoes, Canis dingo). The slope of the relationship between herbivore biomass and net primary productivity did not differ between Australia and the northern boreal and temperate regions. Herbivore biomass increased in response to primary productivity where canid predators were rare, but showed muted responses to increasing productivity where canid predators were common. MAIN CONCLUSIONS: Canid predators have strong suppressive effects on herbivore biomass that scale with primary productivity. Our study shows that the EEH has wide application to canid‐predator–herbivore dynamics and may be relevant to the management of herbivores because it can provide an indication of how herbivore biomass and densities may vary in relation to ecosystem productivity and the presence and absence of canid predators.