Alien predators are more dangerous than native predators to prey populations

Abstract

Alien predators are widely considered to be more harmful to prey populations than native predators. To evaluate this expectation, we conducted a meta-analysis of the responses of vertebrate prey in 45 replicated and 35 unreplicated field experiments in which the population densities of mammalian and avian predators had been manipulated. Our results showed that predator origin (native versus alien) had a highly significant effect on prey responses, with alien predators having an impact double that of native predators. Also the interaction between location (mainland versus island) and predator origin was significant, revealing the strongest effects with alien predators in mainland areas. Although both these results were mainly influenced by the huge impact of alien predators on the Australian mainland compared with their impact elsewhere, the results demonstrate that introduced predators can impose more intense suppression on remnant populations of native species and hold them further from their predator-free densities than do native predators preying upon coexisting prey.

Figure 1

Mean effect sizes of prey in replicated predator manipulation experiments. (a) Effects of native and introduced predators and (b) effects of native and introduced predators, with the effects of introduced predators divided between Australian experiments and experiments conducted elsewhere. Effect size is calculated as Hedges' d. Bars represent 95% bias-corrected confidence intervals.

Figure 2

Mean effect sizes of prey in unreplicated predator manipulation experiments. (a) Effects of native and introduced predators and (b) effects of native and introduced predators, with the effects of introduced predators divided between Australian experiments and experiments conducted elsewhere. Effect size presented as back-transformed ln(X_e/X_c), where X_e and X_c are the treatment and control prey responses, respectively. Bars represent 95% confidence intervals.