Using variation partitioning techniques to quantify the effects of invasive alien species on native urban bird assemblages

Quantifying the impacts that invasive alien species (IAS) cause on affected systems is not an easy task. Here, we explore the application of variation partitioning techniques to measure and control for the effects of possible confounding factors when studying the impact that feral pigeons, European starlings, and house sparrows cause on native urban bird communities in Mexico. We argue that these IAS are provoking a severe impact on whole assemblages of native passerines only if (a) their marginal effect is statistically significant, (b) it remains so after partialling out other explanatory variables, and (c) is larger than (or similar to) the conditional effect of these covariates. We censused passerine bird assemblages and measured habitat variables in a number of greenspaces in three replicate study areas. Then, by means of partial redundancy analyses, we decomposed the total variability in bird data as a function of IAS, physical variables and vegetation data. In one of the study areas the marginal effect of IAS on native assemblages was significant, but the conditional effect was not. We conclude that this IAS effect was confounded with other explanatory variables. In the other study areas, no (marginal or partial) significant effect was found. Without invoking interspecific competition, our results support the opportunistic hypothesis, according to which IAS can exploit ecological conditions in modern cities that native species cannot even tolerate. Finally, apart from the Precautionary Principle, we found no scientific justification to control the abundance of the three IAS in our study areas.