Habitat disturbance and invasive fire ants: testing the passenger/driver hypothesis in a longleaf pine savanna

Background/Question/Methods

Correlations between invasive species dominance and native species decline are often used to assess the impacts of invasive species but are usually confounded by environmental changes associated with invasion, such as habitat disturbance. Are invasive species “drivers” or “passengers” of ecological shifts in communities where they occur (MacDougall and Turkington 2005)? The driver model predicts that native and invasive species interact strongly and that native species are limited or excluded by competition with invasive species. Under this model, experimental removal of invasive species should increase abundance and species richness of native species. In contrast, the passenger model predicts that communities are primarily structured by factors other than interactions with invasive species (e.g., environmental change) and that these factors are less constraining to invasive species. Under this model, experimental removal of invasive species should have minimal impact on native species. An invasive species of particular focus in this debate is the red imported fire ant, Solenopsis invicta. Using a 2x2 factorial design (fire ant reduction/control x soil disturbance/control) in longleaf pine savanna (SC, USA), we experimentally tested which model best explains the interaction between fire ants and native ants. 

Results/Conclusions

We used targeted applications of hydramethylnon to achieve ~50% reduction in fire ant abundance in fire ant reduction plots relative to control plots. In fire ant reduction plots overall species richness was reduced by ~40%, suggesting that non-target species were also affected by the hydramethylnon treatment. Ants in the genera Aphaenogaster, Nylanderia, and Aphaenogaster were less abundant in fire ant reduction plots relative to controls. However predatory Hypoponera ants increased in abundance and estimated trophic position (δ¹⁵N) in fire ant reduction plots relative to controls. Soil disturbance treatments did not significantly affect fire ant abundance, native ant abundance, or species richness. However some ants (e.g., Dorymyrmex) increased in abundance with soil disturbance, while others (e.g., Aphaenogaster, Nylanderia) decreased. Because our fire ant reduction treatment also generated declines in other species, it is difficult determine which model is driving community structure. However, at the species-level, responses to the treatments suggest limitation by both the degree of fire ant invasion or soil disturbance among ant species.