Search

The diversity-invasibility hypothesis and ecological theory predict that high-diversity communities should be less easily invaded than species-poor communities, but empirical evidence does not consistently support this prediction. While fine-scale experiments tend to yield the predicted negative association between diversity and invasibility, broad-scale observational surveys generally report a positive correlation. This conflicting pattern between experiments and observational studies is referred to as the invasion paradox and is thought to arise because different processes control species composition at different spatial scales. Here, we test empirically the extent to which the strength and direction of published diversity-invasibility relationships depend on spatial scale and on the metrics used to measure invasibility. Using a meta-analytic framework, we explicitly separate the two components of spatial scale: grain and extent, by focusing on fine-grain studies that vary in extent. We find evidence of multiple drivers of the paradox. When we consider only fine-grain studies, we still observe conflicting patterns between experiments and observational studies. In contrast, when we examine studies that are conducted at both a fine grain and fine extent, there is broad overlap in effect sizes between experiments and observation, suggesting that comparing studies with similar extents resolves the paradox at local scales. However, we uncover systematic differences in the metrics used to measure invasibility between experiments, which use predominantly invader performance, and observational studies, which use mainly invader richness. When we consider studies with the same metric (i.e., invader performance), the contrasting associations between study types also disappear. It is not possible, at present, to fully disentangle the effect of spatial extent and metric on the paradox because both variables are systematically associated in different directions with study type. There is therefore an urgent need to conduct experiments and observational studies that incorporate the full range of variability in spatial extent and invasibility metric., (© 2019 by the Ecological Society of America.)

Species invasions have a range of negative effects on recipient ecosystems, and many occur at a scale and magnitude that preclude complete eradication. When complete extirpation is unlikely with available management resources, an effective strategy may be to suppress invasive populations below levels predicted to cause undesirable ecological change. We illustrated this approach by developing and testing targets for the control of invasive Indo-Pacific lionfish (Pterois volitans and P. miles) on Western Atlantic coral reefs. We first developed a size-structured simulation model of predation by lionfish on native fish communities, which we used to predict threshold densities of lionfish beyond which native fish biomass should decline. We then tested our predictions by experimentally manipulating lionfish densities above or below reef-specific thresholds, and monitoring the consequences for native fish populations on 24 Bahamian patch reefs over 18 months. We found that reducing lionfish below predicted threshold densities effectively protected native fish community biomass from predation-induced declines. Reductions in density of 25–92%, depending on the reef, were required to suppress lionfish below levels predicted to overconsume prey. On reefs where lionfish were kept below threshold densities, native prey fish biomass increased by 50–70%. Gains in small (<6 cm) size classes of native fishes translated into lagged increases in larger size classes over time. The biomass of larger individuals (>15 cm total length), including ecologically important grazers and economically important fisheries species, had increased by 10–65% by the end of the experiment. Crucially, similar gains in prey fish biomass were realized on reefs subjected to partial and full removal of lionfish, but partial removals took 30% less time to implement. By contrast, the biomass of small native fishes declined by >50% on all reefs with lionfish densities exceeding reef-specific thresholds. Large inter-reef variation in the biomass of prey fishes at the outset of the study, which influences the threshold density of lionfish, means that we could not identify a single rule of thumb for guiding control efforts. However, our model provides a method for setting reef-specific targets for population control using local monitoring data. Our work is the first to demonstrate that for ongoing invasions, suppressing invaders below densities that cause environmental harm can have a similar effect, in terms of protecting the native ecosystem on a local scale, to achieving complete eradication.