Predicting species invasion success and impacts on individual-based modeled fish communities

Background/Question/Methods

Invasion ecology has long been faced with the challenge of predicting how species invasion success and impacts depend on biological features of introduced species and recipient communities. Modeling can be a useful tool for such a broad question, not being as limited as empirical studies in the amount of necessary data and replication. Here we used an individual-based model of fish community dynamics to evaluate invasion processes. The model includes detailed rules for fish life history, size-structured predation, and several important ecological tradeoffs, such as foraging (and growth) versus safety against predation, and diet generality versus predation efficiency. Communities were assembled by sequential species introductions along gradients of basal resources’ productivity and predators’ maximum attack rates. After 100 years of simulation, the resultant communities were then subjected to experimental introductions from resultant species pool. We analyzed the influence of 35 species and environmental variables on the chance of invasion success and native species extinction. The variables were classified according to three criteria: (i) species versus community features; (ii) fundamental versus realized features; (iii) individualistic versus species-community matching features. We compared the performance of several logistic regression models, defined by combinations of those different kinds of variables, using AIC criteria. 

Results/Conclusions

Invasion success was increased by both propagule pressure and previous experimental deletion of native species. Successful invaders tended to possess life-history features similar to those of invaded community, but low diet overlap. This finding is in accordance with theories for both environmental filtering and niche partitioning. On the other hand, invaders dissimilar from native species are more probable to cause extinctions, highlighting a conflict between establishment and impact. More productive and less rich communities are more susceptible to invasions, while their native species are less vulnerable to extinction. Successful invaders also tended to specialize in smaller (and more abundant) resources and to be less piscivorous. In contrast, invaders with highest impacts tend to be more generalist and piscivorous, preferring large food items. The comparison of statistical models shows that the global model (containing all 35 variables) was by far the best to explain both invasion success and native extinction. The prevalence of the global model supports some empirical findings, suggesting that no specific kind of species or environmental variable may be sufficiently good. Therefore, our modeling approach emphasizes that all different sorts of information are necessary to predict the fate of invasions and their ecological consequences.